Information Technology Governance and Innovation

Adoption in Varying Organizational Contexts

Mobile Government and Software as a Service

D I S S E R TAT I O N

zur Erlangung des akademischen Grades

Dr. rer. pol.

im Fach Wirtschaftsinformatik

eingereicht an der
Wirtschaftswissenschaftlichen Fakultt
Humboldt-Universitt zu Berlin

von
Dipl.-[Link] Till J. Winkler

Prsident der Humboldt-Universitt zu Berlin:

Prof. Dr. Jan-Hendrik Olbertz

Dekan der Wirtschaftswissenschaftlichen Fakultt:

Prof. Dr. Ulrich Kamecke

Gutachter:
[Link]. Oliver Gnther, Ph.D.
[Link]. Carol V. Brown, Ph.D.
eingereicht am: 7.6.2012
Tag der mndlichen Prfung: 12.7.2012
 Abstract

This cumulative dissertation contributes to the question of the theoretical relation-

ship between information technology (IT) governance and the adoption of IT-based
innovations. IT governance has been described specifically as the locus of responsi-
bility for IT functions within organizations. Innovation adoption in this context
refers to the decision of an organization to make use of a technological innovation.
Two principal research questions (RQ) guide this dissertation: (1) how does the mode
of IT governance influence adoption of new technologies, and conversely (2) how does
the adoption of new technologies affect organizational IT governance?
In order to address RQ1, I conducted four studies in a public sector context
regarding innovations in Mobile Government (M-Government) referring to the use of
mobile technology to improve government services and internal processes. In a survey
with 50 German municipalities, I investigated the strategic motivations for adopt-
ing a broad range of emerging M-Government services. The results indicate that
municipal governments take a different pace in IT-based innovation adoption and
therefore can be described by clusters of Innovators, IT experienced, Efficiency-
oriented and Laggards (Chapter4.1). By an in-depth analysis of interview data
from 12 municipalities, I derive a well-grounded framework of drivers and inhibitors
of M-Government adoption. Furthermore, based on cross-case analysis, I provide
empirical evidence that the mode of IT governancemore precisely, the question of
whether responsibilities for IT and organization are effectively alignedis a crucial
prerequisite to foster innovation adoption in public sector organizations. The findings
also show why most municipalities focus on internal M-Government innovations
(Chapter4.2). For this reason, I examined M-Government adoption on the citizen
level in a survey with more than 200 participants. The model tests indicate that
external M-Government services, such as urban sensing, are also effective means
to enable more citizen participation, while perceived privacy risks are not major in-
hibitors (Chapter4.3). Finally, applying a simulation approach and a case validation,
I demonstrate that such services can improve a municipalitys level of environmental
information at comparable cost to internal information acquisition procedures and
in this sensesimultaneously allow for implementing service and process innovations
(Chapter4.4).
Regarding RQ2, I consider the adoption of enterprise Software as a Service (SaaS).
In this context, it is hypothesized that for some applications SaaS-based provision
leads to a governance shift of IT responsibilities from IT towards business units.
Based on an in-depth analysis of four cases of SaaS adoption, I take a multiple-
contingency perspective to isolate the factors that potentially influence the allocation
of application governance (Chapter5.1). An operationalization and test of the pro-
posed contingency model in a survey with 207 large firms reveals, that responsibility
for SaaS-based applications is indeed allocated more frequently to business units.
Drawing on multiple theoretical perspectives, this can be (partly) explained by a
smaller scope of the use of SaaS-based applications and the changing competency
requirements for SaaS-based delivery. However, the locus of the initiative emerges
as the most determining factor for explaining application governance (Chapter5.3).
Recognizing the inherent limitations of a factor-based approach, two cases of SaaS
adoption are compared in detail by applying a process-theoretic paradigm. Here the

iii
 locus of initiative emerges as an intermediate variable that links the mode of over-
all IT governance with the specific application governance outcome (Chapter5.3).
Such process view is taken as a premise to analyze the role of information system
specificity for SaaS governance. In a subsample test for SaaS applications, I am able
to unveil that the functional, human and technological specificity of a SaaS have a
dual influence on the locus of application governance (Chapter5.4).
In summary, this dissertation sheds light on the question of how IT governance and
its mechanisms can foster innovativeness in certain contexts (e.g., through aligning
IT responsibilities in public sector organizations), and conversely how the mode of
IT governance itself can be shaped by the emergence of new technological innova-
tions (e.g., external delivery models such as SaaS). These findings enhance classic
IT governance theory by providing new insights on the mutual relationship of IT
governance and IT innovation and thus corroborate the complementarity of organi-
zational and technological architecture. Methodologically, this work demonstrates
the richness provided by alternating between qualitative and quantitative empirical
approaches. Finally, a number of relevant practical implications for IT decision
makers in governmental and entrepreneurial contexts are outlined.

Keywords: Information Systems, IT Governance, IT Innovation, IT Adoption, Mobile Gov-

ernment, E-Government, Software as a Service, Empirical studies, Multimethod research.

iv
 Zusammenfassung

Diese kumulative Dissertation leistet einen Erklrungsbeitrag zu der Frage der

theoretischen Beziehung zwischen der IT-Governance und der Adoption von IT-
basierten Innovationen auf Organisationsebene.1 IT-Governance kann in diesem Zu-
sammenhang als der Ort der Verantwortungshoheit fr IT-Entscheidungen verstan-
den werden. Adoption bezieht sich auf die Aneignung einer technologischen Inno-
vation durch eine Organisation. Zwei bergeordnete Forschungsfragen leiten diese
Dissertation: (1) Wie beeinflusst die Form der IT-Governance die Aneignung neuer
Technologien, und umgekehrt (2) wie beeinflussen neue Technologien die Form der
IT-Governance?
Hinsichtlich Forschungsfrage (1) wurden vier Studien zu Innovationen im Mobile
Government (M-Government) durchgefhrt, d.h. zu der Nutzung von mobilen Tech-
nologien im ffentlichen Sektor mit dem Ziel Verwaltungsdienstleistungen und interne
Prozesse zu verbessern. In einer Studie mit 50 deutschen Stadtverwaltungen wurden
die strategischen Motivationen untersucht, die zur Annahme (oder Ablehnung) ei-
nes breiten Spektrums an neuartigen M-Government-Diensten fhren knnen. Die
Ergebnisse deuten darauf hin, dass ffentliche Verwaltungen einen unterschiedli-
chen Grad der Aneignung IT-basierter Innovationen aufweisen und sich somit in
Cluster von Innovatoren, IT-Erfahrenen, Effizienz-orientierten und Laggards
einteilen lassen (Kapitel4.1). Aus der detallierten Analyse von Interviews mit 12
stdtischen IT-Entscheidern wird darauf ein gegenstandsverankertes Rahmenenwerk
von Treibern und Hindernissen fr das M-Government entwickelt. Im Rahmen von
Fallvergleichen zeigt sich zudem eine empirische Evidenz dafr, dass die Form der IT-
Governance genauer, die Frage ob Verantwortlichkeiten fr IT sowie Organisation
und Personal effektiv miteinander verbunden sind eine wesentliche Voraussetzung
fr die Umsetzung von IT-Innovationen darstellt. Die Ergebnisse zeigen auch auf,
warum sich viele Stdte bisher auf interne M-Government Anwendungen konzentrie-
ren (Kapitel4.2). Aus diesem Grund wird in einer Studie mit ber 200 Teilnehmern
die Akzeptanz fr einen M-Government-Dienst auf Ebene des Brgers analysiert.
Modelltests zeigen, dass externe M-Government-Dienste, wie z.B. solche der urbanen
Datenerfassung (Urban Sensing), einen probaten Weg zu mehr Brgerbeteiligung
ermglichen, wohingegen Datenschutzbedenken auf Nutzerseite kein wesenliches Hin-
dernis darstellen (Kapitel4.3). Schlielich wird durch einen Simulationsansatz und
der Validierung in einer Fallstudie demonstriert, dass externe M-Government-Dienste
den Informationsgrad von Verwaltungen erhhen knnen bei vergleichbaren Kosten
zu der internen Informationsgewinnung und somit gleichzeitig Dienstleistungs- und
Prozess-Innovationen erzielt werden knnen (Kapitel4.4).
In Bezug auf Forschungsfrage (2) wurde die Aneignung von Unternehmenssoft-
ware as a Service (SaaS), d.h. die Nutzung von Geschftsanwendungen als webba-
sierte Dienste, untersucht. In diesem Zusammenhang wird hypothetisiert, dass die
SaaS-basierte Bereitstellung fr einige Anwendungen zu einer Verschiebung der IT-
Verantwortlichkeiten von IT-Abteilungen zu Fachbereichen fhrt. Basierend auf vier
Fallstudien wird zunchst ein kontingenzbasierter Ansatz gewhlt, um solche Fakto-
ren zu isolieren, die einen potenziellen Einfluss auf die Verteilung der Anwendungs-
1
IT: Informationstechnologie

v
 hoheit haben (Kapitel5.1). Eine Operationalisierung und Tests des vorgeschlagenen
Kontingenzmodells in einer Studie mit 207 Grounternehmen zeigen auf, dass die
Verantwortung fr SaaS-baasierte Anwendungen tatschlich hufiger Fachabteilun-
gen zugeordnet ist. Bezug nehmend auf mehrere theoretische Perspektiven kann dies
(zum Teil) durch einen geringeren Nutzungsumfang von SaaS-basierten Anwendun-
gen in Unternehmen sowie durch sich verndernde Kompetenzanforderungen fr das
SaaS-basierte Anwendungsmanagement erklrt werden. Als am strksten ausschlag-
gebender Faktor tritt jedoch der Ursprung der Initiative der SaaS-Einfhrung hervor
(Kapitel5.2). In Anerkennung der methodeninhrenten Einschrnkungen eines fak-
torbasierten Vorgehens werden zwei Flle von SaaS-Einfhrungen unter Verwendung
eines prozesstheoretischen Ansatzes analysiert. Der Ursprung der Initiative zeigt sich
hierbei als intermedire Variable, die den Modus der bergreifenden IT-Governance
mit dem konkreten Resultat auf Anwendungsebene kausallogisch verbindet (Ka-
pitel5.3). Eine solche Prozesssicht dient ebenfalls als Prmisse um die Rolle der
Informationssystem-Spezifitt auf SaaS-Governance (d.h. die Anwendungshoheit) zu
untersuchen. Ein Test der Stichprobe fr SaaS-Anwendungen deckt auf, dass die
funktionale, personelle und technologische Spezifitt eines SaaS-Informationssystems
einen dualen Einfluss auf den Ort der Verantwortungshoheit ausbt (Kapitel5.4).
Zusammenfassend gibt diese Dissertation Aufschluss darber, wie IT-Governance
und entsprechende Mechanismen die Innovativitt in bestimmten organisationalen
Kontexten begnstigen knnen (in ffentlichen Verwaltungen z.B. durch die Ver-
knpfung von bestimmten IT-Verantwortlichkeiten) und umgekehrt wie die Form
der IT-Governance selbst durch das Aufkommen von technologischen Neuerungen
(z.B. durch externe Bereitstellungsmodelle wie SaaS) umgestaltet wird bzw. wer-
den muss. Diese Ergebnisse erweitern die klassische IT-Governance-Theorie durch
neue Erkenntnisse bezglich des wechselseitigen Verhltnisses von IT-Governance
und IT-Innovation, wodurch die Komplementaritt zwischen der organisatorischen
und der technologischen Unternehmensarchitektur untermauert wird. Methodisch
demonstriert diese Arbeit den Reichtum, der durch den wechselnden Einsatz von
qualitativen und quantitativen Anstzen erzielt werden kann. Abschlieend werden
eine Reihe von Implikationen fr IT-Entscheider in ffentlichen und privatwirtschaft-
lichen Kontexten aufgezeigt.

Schlsselwrter: Wirtschaftsinformatik, IT-Governance, IT-Innovation, IT-Adoption, Mobile

Government, E-Government, Software-as-a-Service, Empirische Studien, Multimethodaler An-
satz.

vi
 I am wholeheartedly grateful to my academic advisors Oliver Gnther, for providing
a great framework for conducting this research and supporting my studies throughout
all phases of this dissertation, and Carol V. Brown, for mentoring and sharing her
longstanding experience in IT organization and governance research.

I also would like to express my gratitude to Francis Bidault and our colleagues at the
European School of Management and Technology (ESMT), who fruitfully collaborated
with us during some of the studies that this research is based on.

In particular I want to thank all my colleagues at Humboldt-Universitt zu Berlin

including visiting guests for inspiring me and never hesitating to discuss and further
my research, especially (in alphabetical order) Alexander Benlian, Franziska Brecht,
Christoph Goebel, Ksenia Koroleva, Hanna Krasnova, Tyge Kummer, Steffen Kunz,
Henrik Leopold, Jan Mendling, Luis Ortigueira, Marco Sarstedt, Kerstin Schfer, Peter
Trkman, and Holger Ziekow.

Furhermore, it is a great pleasure to thank the faculty members and the participants of
the 2011 Doctoral Consortium at ECIS, first and foremost Maung Sein, Cathy Urquart,
Sirkka Jarvenpaa and, again, Carol Brown, for providing a wonderful platform for aca-
demic growth and excellent feedback to my dissertation and, after all, for a fun atmo-
sphere.

At this point, I would also like to make reference to my friends and colleagues at my
former employer, who provided ideas for this research even though they (and possibly I)
were not even aware of it at the time, in addition to helping validate some of the instru-
ments used for my studies.

I am also much obliged to the secretaries and student assistants at our department for
coordinating and supporting some of the operative tasks connected to this work. Moreover,
a number of students wrote their master, bachelor or seminar theses in the context of
this research, whom I want to thank collectively at this time. Although the last, not least,
I want to give credit to my cousin Andreas Witzel, who made sure that the supply with
fresh papers never stopped.

Writing this thesis would not have been possible without the support of my family and
friends surrounding me. I particularly want to thank Luna for her enduring support.
This thesis is dedicated to my parents.
Contents

1 Introduction1
1.1 Mobile Government .................................................................................................. 5
1.2 Software as a Service .............................................................................................. 7
1.3 Thesis Structure......................................................................................................9

2 Foundations of IT Governance and Organization Design13

2.1 Preamble................................................................................................................. 13
2.2 Introduction .............................................................................................................. 13
2.3 Six Dimensions of IT Organization Design ......................................................... 14
2.3.1 Allocation of IT Decision Rights ............................................................ 14
2.3.2 Allocation of IT Resources ..................................................................... 15
2.3.3 Coordination Mechanisms ........................................................................ 16
2.3.4 Financial Autonomy...................................................................................... 17
2.3.5 Sourcing Arrangements .......................................................................... 18
2.3.6 Capabilities and Skills ............................................................................... 19
2.4 IT Organization Archetypes .................................................................................. 20
2.4.1 Centralized Model .................................................................................... 20
2.4.2 Decentralized Model .................................................................................21
2.4.3 Shared Services Model .............................................................................. 22
2.4.4 Corporate Coordinator Model .................................................................. 23
2.5 Motivation for Further Research ......................................................................... 24
2.6 Summary ................................................................................................................. 27

3 Methodological Foundations29
3.1 Epistemological Preamble ..................................................................................... 29
3.2 Qualitative Methods .............................................................................................. 31
3.2.1 Interviews ................................................................................................... 31
3.2.2 Content Analysis ........................................................................................ 31
3.2.3 Grounded Theory .................................................................................... 32
3.2.4 Case Studies.............................................................................................. 33
3.3 Quantitative Methods............................................................................................ 35
3.3.1 Survey...........................................................................................................35
3.3.2 Structural Equation Modeling................................................................... 36
3.3.3 Clustering and Subgroup Analysis ......................................................... 38
3.3.4 Simulation ................................................................................................. 39

ix
Contents

4 IT Governance and Innovation Adoption in E-Government41

4.1 Innovations in Mobile Government .........................................................................41
4.1.1 Preamble.....................................................................................................41
4.1.2 Introduction .............................................................................................. 41
4.1.3 Theoretical Foundations and Hypotheses Development ...................... 42
4.1.4 Qualitative Pre-Study ............................................................................... 46
4.1.5 Empirical Study........................................................................................ 48
4.1.6 Model Tests and Results Discussion...................................................... 50
4.1.7 Conclusion ................................................................................................ 57
4.1.8 Summary .................................................................................................... 59
4.2 Towards Transformational IT Governance .......................................................... 60
4.2.1 Preamble.....................................................................................................60
4.2.2 Introduction .............................................................................................. 60
4.2.3 Related Work ....................................................................................................... 61
4.2.4 Research Method........................................................................................ 62
4.2.5 Framework for Mobile Government Adoption ...................................... 64
4.2.6 Mobile Government Cases ........................................................................ 68
4.2.7 Conclusion ................................................................................................ 73
4.2.8 Summary .................................................................................................... 75
4.3 Citizen Acceptance in Urban Sensing................................................................... 76
4.3.1 Preamble.....................................................................................................76
4.3.2 Introduction .............................................................................................. 76
4.3.3 Theoretical Foundations and Model Development................................ 77
4.3.4 Methodology ............................................................................................. 82
4.3.5 Model Analysis and Discussion............................................................... 83
4.3.6 Conclusion ................................................................................................ 87
4.3.7 Summary .................................................................................................... 89
4.4 Municipal Benefits of Urban Sensing .................................................................. 90
4.4.1 Preamble.....................................................................................................90
4.4.2 Introduction .............................................................................................. 90
4.4.3 Related Work ....................................................................................................... 91
4.4.4 Research Method........................................................................................ 96
4.4.5 Simulation Model ........................................................................................ 99
4.4.6 Case Validation .......................................................................................... 103
4.4.7 Conclusion ................................................................................................ 112
4.4.8 Summary .................................................................................................... 115

5 Innovation Adoption and IT Governance in Enterprise Information Systems117

5.1 The Impact of Software as a Service on Information Systems Authority.117
5.1.1 Preamble.....................................................................................................117
5.1.2 Introduction .............................................................................................. 117
5.1.3 Theoretical Foundations ........................................................................... 118
5.1.4 Research Methodology .............................................................................. 121

x
 Contents

5.1.5 Results A Contingency Model for Application Governance ..............123

5.1.6 Comparative Case Studies ........................................................................ 130
5.1.7 Conclusion ................................................................................................ 135
5.1.8 Summary .................................................................................................... 137
5.2 Comparing Authority for On-Premises Applications and SaaS........................ 138
5.2.1 Preamble .................................................................................................... 138
5.2.2 Introduction .............................................................................................. 138
5.2.3 Three Theoretical Lenses on Application Authority .............................. 140
5.2.4 Research Model and Hypotheses ............................................................ 142
5.2.5 Methodology ............................................................................................. 151
5.2.6 Analysis ....................................................................................................... 156
5.2.7 Discussion .................................................................................................... 160
5.2.8 Summary .................................................................................................... 164
5.3 A Process Model for Explaining Governance of Software as a Service. . .165
5.3.1 Preamble .................................................................................................... 165
5.3.2 Introduction .............................................................................................. 165
5.3.3 Related Work ........................................................................................................166
5.3.4 A Process Model for SaaS Adoption and Governance .......................... 168
5.3.5 Empirical Illustration of the Process Approach: Two Cases ..............170
5.3.6 Conclusion ................................................................................................ 177
5.3.7 Summary .................................................................................................... 178
5.4 The Dual Role of Information Systems Specificity for Governing SaaS. .180
5.4.1 Preamble .................................................................................................... 180
5.4.2 Introduction .............................................................................................. 180
5.4.3 Theoretical Lens ........................................................................................ 182
5.4.4 Research Model ........................................................................................... 183
5.4.5 Methodology ............................................................................................. 190
5.4.6 Analysis ....................................................................................................... 191
5.4.7 Discussion .................................................................................................... 197
5.4.8 Summary .................................................................................................... 202

6 Conclusion and Contributions203

6.1 Theoretical Contribution: On the Relationship of IT Governance and
Innovation .................................................................................................................204
6.2 Practical Contribution: Reflections on IT Governance in Public and Pri-
vate Sector .............................................................................................................. 207
6.3 Methodological Contribution: On the Use of Multimethod Research. . .210
6.4 Limitations and Future Research ......................................................................... 214

Appendices217
1 M-Government Survey with German Municipalities (Online) ........................... 217
2 Citizen Acceptance Survey (Online)................................................................... 229
3 Urban Sensing Simulation Model (Scripts) ......................................................... 237
4 Workshop Documentation (Excerpt) ................................................................... 239

xi
Contents

5 SaaS Interview Guideline (Shortened)................................................................ 243

6 Business and Information Technology Survey (Supplements) .......................... 245
6.1 Cover Letter .............................................................................................. 245
6.2 Questionnaire ............................................................................................. 247
6.3 Sample Description ..................................................................................... 257
7 SaaS Adoption Processes (Enlarged) ................................................................... 265
8 Information Systems Specificity (Items and Cross-loadings) ...........................269

Bibliography271

xii
List of Figures

1.1 Principal research questions and definitions ...................................................... 4

2.1 IT organization archetypes .................................................................................. 16

3.1 Grounded theory elements .................................................................................... 33

4.1 Research model ....................................................................................................... 44

4.2 Sample description..................................................................................................50
4.3 Structural model results .........................................................................................55
4.4 Mean factor scores per municipality cluster .......................................................56
4.5 Interview sample description .............................................................................. 62
4.6 Adapted coding paradigm fromStrauss and Corbin(1998) .............................. 64
4.7 Framework for m-government adoption ............................................................... 65
4.8 Research Model ....................................................................................................... 78
4.9 Respondent age and occupation ........................................................................... 83
4.10 Structural model results .........................................................................................86
4.11 Field of research .................................................................................................... 93
4.12 System Dynamics syntax ........................................................................................ 95
4.13 Build, evaluation, intervention cycles ................................................................... 97
4.14 System dynamics representation......................................................................... 99
4.15 Complaint and defect management process and selected subprocesses. . .104
4.16 Survey results ........................................................................................................... 106
4.17 Simulation results ................................................................................................. 109
4.18 Estimated and actual adoption curves ................................................................ 111

5.1 Adapted coding paradigm from Glaser (1992).................................................... 123

5.2 Contingency model for application governance ................................................... 124
5.3 Research model ....................................................................................................... 142
5.4 Application governance patterns ........................................................................ 157
5.5 Model tests and subgroup analysis...................................................................... 158
5.6 SaaS adoption processes (overview) ..................................................................... 175
5.7 Interrelationship of application governance factors.......................................... 176
5.8 Research model ....................................................................................................... 184
5.9 Multidimensional visualization of application types .......................................... 197

1 Survey reminders and response distribution ...................................................... 259

2 Respondent work years and position .................................................................. 261

xiii
List of Figures

3 Manufacturing vs. service industries ................................................................... 262

4 Physical vs. informational products ...................................................................... 262
5 Business to consumer vs. business to business ................................................... 262
6 Employee firm size ................................................................................................. 263
7 Financial firm size ................................................................................................. 263
8 Histogram of IT employees and IT budget.......................................................... 264
9 Position of the CIO .................................................................................................. 264
10 SaaS Adoption Process Case B ......................................................................... 266
11 SaaS Adoption Process Case C ......................................................................... 267

xiv
List of Tables

1.1 Thesis structure ........................................................................................................ 9

1.2 Overview of publications ........................................................................................11

2.1 Organizational design dimensions ............................................................................. 20

2.2 Key characteristics of four IT organization archetypes .................................... 25

3.1 Overview of research methods .............................................................................. 35

4.1 Interviewees and city sizes..................................................................................... 46

4.2 Operationalization of antecedent factors ............................................................ 48
4.3 Descriptive results of service attractiveness ...................................................... 51
4.4 Tolerance values of formative indicators ............................................................. 52
4.5 Convergent validity criteria .................................................................................. 53
4.6 Discriminant validity criteria............................................................................... 54
4.7 Municipality clusters and group differences ...................................................... 57
4.8 M-government service outcomes ...........................................................................67
4.9 M-government cases overview ................................................................................... 70
4.10 Contingencies for m-government adoption and target groups...........................73
4.11 Measurement instrument, descriptive statistics and reliabilites ........................84
4.12 Convergent and discriminant validity criteria....................................................85
4.13 Simulation model parameters............................................................................... 108
4.14 Qualitative aspects of introducing mobile reporting .......................................... 110

5.1 Case companies and key figures ........................................................................... 122

5.2 Contingency factors and proposed influence on application governance. .129
5.3 Case comparison and contingent forces ............................................................... 132
5.4 Comparison of induced and present governance mode....................................134
5.5 Hypotheses overview................................................................................................... 150
5.6 Application types and sample distribution ......................................................... 153
5.7 Validity criteria ........................................................................................................ 155
5.8 Results of the hypotheses tests ............................................................................160
5.9 Case key figures........................................................................................................ 171
5.10 Measurement model validity ..................................................................................... 192
5.11 Model tests .............................................................................................................. 193
5.12 Application types ................................................................................................. 196

6.1 Contributions to IT governance theory................................................................ 208

xv
List of Tables

6.2 Multimethod research overview ........................................................................... 211

3 Industry classification ........................................................................................... 260

4 Respondent industries ........................................................................................... 260
5 Items and cross-loadings ........................................................................................ 269

xvi
1 Introduction
Innovation is the primary source of competitive advantage for companies and the basis
of economic development (Schumpeter1926;Burns and Stalker1966;Acemoglu2012).
Most organizations, both in the public and private sector, constantly face the challenge
to innovate, i.e. to bring out novel products or services as well as to improve internal
processes in order to compete on the external market and increase productivity (Utter-
back and Abernathy1975;von Hippel1988). Information technology (IT) today plays a
pivotal role in organizational innovation adoption (Acemoglu2012). Hardly any product,
service or process innovation can succeed without being supported, if not enabled, by
IT (Davenport1993). For example, in the last decade public sector agencies worldwide
have dedicated much effort to bringing government services online to the Internet, a
development that has been widely termed as Electronic Government (UN2012). In
addition, in the private sector most enterprises have implemented centralized repositories
for customer data to facilitate customer relationship management (CRM) processes and
exploit market opportunities (Chen and Popovich2003;Kumar et al.2011).
Obviously, what is an IT-based product or service innovation for one party (i.e., the
vendor or provider) may represent a process innovation for the other (i.e., the client
or user organization). In this sense, public agencies that bring out new E-Government
services enable their customers to innovate in the process of interaction with their govern-
ment. Or regarding the second example, companies that use vendor solutions to support
their CRM processes, benefit from the product innovation brought out (earlier) by this
vendor. Therefore, it is not the mere investment and the adoption of IT innovations that
creates valueas the early literature on IT value suspected when trying to resolve the
IT paradoxon (e.g.,Brynjolfsson1993;Triplett1999). Rather, IT-based innovations create
value only when the technology itself also fits to the needs of the client and is embedded
in the processes of the user organization (Brynjolfsson and Hitt2000;Soh et al.2000).
This also motivates why adopting IT-based product/service innovations and
implementing new technologies in organizations often leads to major organizational change
(Keen1981;Lyytinen and Newman2008)because exploiting the opportuni- ties
brought about by these innovations often requires a change of the companys (or
governments) practices, processes, and culture likewise (Sia and Soh2007;Strong and
Volkoff2010). For this reason, IT-based innovation adoption may sometimes even entail
a transformational impact on organizations and organization structure (Venkatraman
2005;Irani et al.2008;Winkler et al.2008).
In a structural view, organizations typically bundle functions that are specialized on
planning, designing and operating IT resources for the rest of the organization (i.e., the
business) in ahowever naturedIT function (Agarwal and Sambamurthy2002). The

xvii
1 Introduction

question on how to align the IT function with the business organization, especially on a
structural, procedural and relational level, is commonly viewed as the central concern of
IT governance (e.g.,Brown and Magill1994;Sambamurthy and Zmud1999;Schwarz
and Hirschheim2003;Weill and Ross2004a;Van Grembergen2004). A crucial, if
not the most fundamental, dimension of IT governance refers to the allocation of IT
decisions rights in an organization. That is, this dimension focuses on which are the
major decisions regarding IT management and use and who should make them. Given
there is a multitude of stakeholders, such decision rights can be both shared horizontally
(i.e., between business and IT stakeholders) and vertically (i.e., between C-level, senior
level, mid level and staff level) within an organization (Weill and Ross2004a). Or, in
a simplified view (i.e., combining the horizontal and vertical dimensions), IT decision
rights are shared between centralized and decentralized groups (Brown and Magill1994;
Sambamurthy and Zmud1999). In line with the broader organization science literature
(Daft2009), the IT governance literature emphasizes that there is no universal way for
designing IT governance. Rather the best way of governing IT functions depends on
certain, foremost business-related, contingencies (seeBrown and Grant2005, p. 703, for
an overview). For example, it has been confirmed that smaller companies tend
to centralize IT governance, while larger companies create more complex federal and
decentralized structures. However, asBrown and Grant(2005, p. 704) note, absent
from the list of [contingency] variables is [still] a discussion on technology and technology
adoption, where surprisingly, little to no research was found.
In practice, companies that struggle with a lack of innovativeness often ask who
should be responsible for IT-based innovations, business or IT units? (e.g.,Power2012).
Having argued that IT-based innovations create value only when they become part of
the organizations work routines, it becomes apparent that the adoption of IT-based
innovations is a key governance issue, which requires the integration of both business
and IT stakeholders. However, we may still ask to which degree of involvement this
should happen. Regarding the relationship of IT innovation and IT governance, the
literature provides the rationale of a strategy-structure fit (Brown and Grant2005,
p. 204). That is, firms that seek competitive advantage primarily through differentiation
(i.e., by product and service innovations) tend to decentralize IT governance structures
in order to sustain technological responsiveness to the needs of internal (and external)
customers. Conversely, firms that follow a cost leadership strategy tend to centralize IT
governance in order to leverage internal economies of scale (Weill and Ross2004b).
Nevertheless, the rationale of balancing scale versus responsiveness possesses some
inherent limitations. First, it largely focuses on a companys product and service inno-
vations and thus does not inform on how to allocate decision rights for increased process
innovationwhich is often the primary goal of organizational IT use (Davenport and
Short2003). Second, it merely focuses on the business drivers and thus neglects the
potential technology contingencies. Given the past pendulum swings between centralized
and decentralized forms of IT use (Peak and Azadmanesh1997;Brynjolfsson and Hitt
1998;Evaristo et al.2005), the mode of governance may also clearly depend on the
type of technology that is prevailing (howsoever this technology can be characterized,

2
Orlikowski and Iacono2001). Third, it largely regards the IT as a bulk function for
which companies allocate decision rights to ensure overall innovativeness. However, the
singularity in the occurrence of IT innovations implies that it may be appropriate to
also take a more modular view on IT governance, i.e. depending on the IT artifact that
is subject to the IT-based innovation of interest. Fourth, it takes a unidirectional view
assuming that IT innovation is always the result of strategy and governance structure.
However, many IT innovations that enter the enterprise from IT market, i.e. from the
vendor side, occur without a defined business demand. Such innovations may conversely
also impact the mode of IT governance. Fifth, past research has been largely directed
at enterprise IT governance, i.e. at the private sector. The rationales for defining
appropriate governance arrangement in public sector organizationsin their nature to
be non-profit drivenmay clearly deviate from this (Weill and Ross2004a, pp. 185-214;
Sethibe et al.2007). In sum, despite more than two decades of IT governance research
(Brown and Grant2005), we know few about the mutual relationship of IT governance
and IT innovations.
This thesis investigates the role of IT governance arrangements in various IT innovation
and adoption contexts. The approach taken in this thesis aims to enhance our under-
standing by building on the extant literature. In particular, it (1) explicitly considers
service and process-based IT innovations, (2) explores the technology contingencies of IT
governance pertaining to the IT artifact, (3) conceptualizes governance arrangements for
different IT sub-functions, and (4) takes a bidirectional view where the emergence of an IT
innovation itself may impact the mode of governance and vice versa, (5) across different
public and private sector contexts. In the framework of this work, we understand an IT-
based innovation as the benefits that result from adopting and using a new technology in
an organization. Adoption in turn refers to the decision of (an individual or) an
organization to make use of an organization (Rogers1962). Overall, this thesis is guided
by two principal research questions:
RQ1: How does the mode of IT governance influence the adoption of new technologies,
and conversely
RQ2: How does the adoption of new technologies affect organizational IT governance?
To address these research questions, I consider two distinct IT-based innovations that
have recently attracted much attention both in theory and in practice. The first inno-
vation refers to the implementation of Mobile Government (M-Government) services by
public agencies, the second to the adoption of enterprise Software as a Service (SaaS).
For each of these different contexts of innovation, four separate studies are conducted
that combine qualitative, quantitative and design-oriented research methods. Regarding
innovations in M-Government, I demonstrate how the strategic framework as well as
the mode of IT governance in municipalities has a bearing on their innovativeness, more
precisely on the extent and focus to which emerging M-Government solutions are adopted.
In an enterprise context I study, how SaaS adoption impacts IT governance and under
which circumstances this can lead to a shift of decision rights towards business units.
Both research questions are depicted in Figure1.1

3
1 Introduction

RQ1: How does the mode of IT governance

influence the adoption of new technologies?

IT Innovation
IT Governance
Adoption

IT governance describes the locus Adoption refers to the decision of an

of responsibility for IT functions. RQ2: How does the adoption of new individual or organization to make
(Brown and Magill 1994) technologies affect organizational IT use of an innovation (Rogers 1962)
governance?

Figure 1.1:Principal research questions and definitions

Furthermore, this thesis also aims to provide concrete practical guidance to foster
decision making in diverse innovation contexts. Regarding M-Government adoption, I
provide insights into the factors that are important (and those that are not important) to
achieve citizen acceptance of M-Government services. Finally, taking an action research
approach, I describe the case of a municipality where I actively observed the introduction
of an M-Government service (i.e., an urban sensing service). In the context of enterprise
SaaS, I first propose and validate a contingency model that may inform practitioners
when shifts in the governance of an enterprise application may occur. However, acknowl-
edging the limitations of such factor-based approach, I propose a process model to better
understand IT governance phenomena in SaaS adoption contexts. Following from this,
I finally revisit the empirical data and unveil a new dualism specifically related to the
technological and artifact-level contingencies for IT governance arrangements.
Besides the theoretical and practical contributions provided in each of these studies,
two important findings emerge from this compound research that extend the classical
view of IT governance. First, based on the case evidence on M-Government adoption, it
shows that those public agencies which effectively connect (and thus largely centralize)
decision rights for IT and organization succeed in implementing process and service
innovations. To some extent, this contradicts the rationale in enterprise IT governance
that organizations will be more innovative when decentralizing IT governance. I introduce the
concept of transformational IT governance to account for this proposition and provide a
broader discussion of this issue in the conclusion. Second, in the course of the presented
studies, I develop a transaction cost theoretic framework to explain (SaaS) application
governance phenomena. It becomes apparent that the classic strategy-structure fit and
the rationale to centralize IT governance for greater efficiency does not necessarily hold,
or may even need to be reverted, for SaaS-based solutions. This finding and the trans-
action cost theoretic framework are also discussed in the conclusion.
In the following I will briefly motivate the choice of Mobile Government and Software
as a Service as two current IT-based innovations, before I explain the thesis structure.

4
 1.1 Mobile Government

1.1 Mobile Government

Since the late 1990s, Electronic Government (E-Government, also sometimes called
Digital Government or Electronic Governance1) has emerged as an independent field
of research (Grnlund and Horan2004). E-Government can be defined as the use of
information and communication technology (ICT) in public administrations combined
with organizational change and new skills in order to improve public services and demo-
cratic processes (EU2003). Although the use of IT (or ICT) by governments has been the
subject of research of earlier studies (seeAnderson and Henriksen2005), the term E-
Government itself has been born in course of the Internet boom (parallel with E-
Commerce) primarily by the idea to bring government services online (Grnlund and
Horan2004). It thus represents a comparably interdisciplinary field of research that
draws on different related areas such as political science, social science and information
systems research (Heeks and Bailur2007).
Not too long after the emergence of E-Government, the term mobile government (M-
Government) was coined to describe such E-Government efforts that include the use of
mobile and wireless technologies (Kushchu and Kuscu2003). The wide recognition of M-
government is driven by the penetration of mobile devices and the emergence of the
mobile Internet (i.e., mobile broadband networks) (ITU2010). Mobile government can
be defined as a strategy and its implementation involving the utilization of all kinds of
wireless and mobile technology, services, applications and devices for improving benefits
to the parties involved in e-government including citizens, businesses and all govern-
ment units (Kushchu and Kuscu2003). Akin to E-Government, different foci of M-
Government are usually differentiated depending on the target group of M-Government
efforts, i.e. Government-to-Citizen (G2G), Government-to-Business (G2B), Government-
to-Government (G2G), Government-to-employee (G2E), and vice versa (i.e., C2G, B2G,
E2G).
External (i.e., G2C and G2B) M-government applications may be further classified by
whether they provide informational or transactional services. Similar to E-Government,
services for information dissemination are generally less problematic, since they en-
able only unidirectional information flow and thus pose less requirements regarding
identification and authentication of the recipient. Early examples for informational
M-Government services include disaster notifications, traffic news, or even voting via
SMS2 (Al-khamayseh et al.2006;Rossel et al.2006;Trimi and Sheng2008). Today, an
increasing number of cities offer mobile applications (i.e., smartphone apps) that provide
a variety of information related to living in that city, e.g. public transport schedules,
touristic information, refuse collection information, etc. (seeVitako2011, pp. 10-14,

1
The terms government and governance should not be confounded in course of this thesis. While the
former (government) is used to refer to the organizational entity of a public agency, the latter is
largely used in the context of Information Technology (IT) governance, which is concerned with the
set of mechanisms that determine how the IT function is managed and aligned within the wider
organizational context.
2
Short message service

5
1 Introduction

for some examples from Germany). However, transactional mobile services, such as
online payment services, tax declaration, car registration known from the E-Government
domain, are still rare in Mobile Government, as they pose greater integration and security
needs (UN2012, p. 41). The question might even be, whether there is a demand for such
services on a mobile channel. Since government transactions (e.g., a tax declaration)
typically require longer interaction with an information system, some parties may argue
that transactional government services are generally not an appropriate use case for a
mobile device.
Internal M-Government applications (i.e., G2G and G2E) arein the light of this
thesislargely viewed as process innovations (Davenport1993). That is, by the use
of mobile and wireless technology, public agencies can handle internal processes more
effectively and efficiently. (For this reason, other authors have also termed this seg-
ment of M-Government as internal efficiency and effectiveness IEE,Trimi and Sheng
2008). Examples include the equipping of government staff (especially field workers)
such as police, firefighters, and field inspectors with mobile devices to provide them with
appropriate information and allow for on-the-spot data processing (Kushchu and Kuscu
2003). This can save valuable time for back-office work, improve data quality, and enable
better dispatching, amongst other benefits (Vitako2011, p. 11). Obviously, the idea of
internal M-Government (or IEE) is not entirely new. For example, public safety depart-
ments have used wireless communication systems ever since the existence of these tools
(Desourdis2002). However, the proliferation of commercial broadband networks and off-
the-shell mobile devices (i.e., mobile phones and tablet PCs) undoubtedly also leads to a
new momentum for M-Government in public authorities which are not concerned with
public safety. Furthermore, in a wider sense internal M-Government applications may also
affect other government workers than those in the field, e.g. when equipping a city hall
with local wireless networks and/or enabling teleworking (Trimi and Sheng 2008). The
increasing consumerization of IT may also lead government employees to expect those
mobile tools in their workplace, that they are used to from their home environments
(Bernnat et al.2010).
Altogether, emerging M-Government solutions represent a broad range of potential
innovations in E-Government that may entail benefits for government customers (i.e.,
citizens and businesses) as well as employees. Although global adoption, diffusion and use
of E-Government itself is still far from reaching a final stage (Grnlund and Horan2004;
UN2012), I regard the subset of M-Government as a particularly interesting research
objective to study IT-based innovation adoption. This is mainly for three reasons: First,
at the time of writing this thesis M-Government is still a relatively new phenomenon that
has been driven by the recent popularity of the mobile Internet and according devices.
Second, it exhibits a very cross-disciplinary character involving technical, social and po-
litical aspects. And finally, this innovation is situated in the public sector, whichdespite
the acknowledged goal to create public valuehas traditionally been less researched in the
IS field (Scholl2006).

6
 1.2 Software as a Service

1.2 Software as a Service

The second IT-based innovation considered in this thesis refers to the use of enterprise
software as a Service (SaaS). In contrast to M-Government, SaaS represents a delivery
model innovation in the way how enterprise software is provided, rather than a concrete
bundle of new applications.3 SaaS refers to the provision of standard software via the
Internet from an external provider who serves multiple customers (tenants) by the same
instance (Cusumano2010). Thus, SaaS can be regarded as a special form of application
outsourcing (Lee et al.2003). Compared to traditional enterprise software, which is
either hosted on dedicated instances at provider side or installed on the companys
own infrastructure (i.e., on-premises), SaaS generally allows for greater economies of
scale due to a better utilization of infrastructure resources. Economically, it is often
emphasized that with SaaS customers rent software (and the underlying infrastructure
resources) instead of buying perpetual-use licenses (Choudhary2007b;Susarla et al.
2009;Lehmann et al.2010).
The SaaS model is not an entirely new phenomenon, rather it has evolved from earlier
forms of web-based delivery which have been termed as application service providing
(ASP) (e.g.,Gnther et al.2001;Susarla et al.2003) or sometimes also netsourcing
(Loebbecke and Huyskens2006). While the borders between ASP and SaaS certainly
have been fluent, it is often argued that the distinguishing criterion for SaaS is the
multitenancy characteristic, i.e. the capability to serve multiple tenants from a single
set of resources (Benlian and Hess2010a). However, the more determining reason for
becoming the accepted term may lie in the commercial breakthrough of the SaaS model,
rather than any definitional distinction. In a recent forecast, market researchers predict
that by 2015, 13 percent of worldwide software spending will be on SaaS delivery and that
24 percent of all new enterprise software purchases will be of a service-enabled software
(Mahowald et al.2011). The main drivers of this commercial successcompared to prior
modelshave obviously been increasing bandwidths, increasing computing power as well
as specific advancements in distributed computing and web development techniques (e.g.,
rich user interfaces, asynchronous web applications and web service standards) (Sun et al.
2007;Fraternali et al.2010). Thus, we can say that SaaS and related delivery models
today represent the commercial realization of the long-held dream of computing as a
utility (Parkhill1966;Carr2004).
Extending beyond that, SaaS is now also considered a part of Cloud computing, more
precisely as the highest layer of the Cloud computing stack (Armbrust et al.2010).
Cloud computing refers more broadly to the use of any kind of computing resource as
a service (aaS) over the Internet (Hayes2008). 4 Three main layers of Cloud services
are distinguished: infrastructure services (IaaS) that provide computational resources,
basic storage and network functionality, platform services (PaaS) that typically provide
3
Although one may argue that M-Government as well represents a delivery model innovation, in a
sense that government services are now delivered via a mobile channel.
4
Some authors even extend the notion of a Cloud to distributed computing within local area networks.
However, an extensive discussion of public versus private clouds shall be omitted here.

7
1 Introduction

a development and execution environment to build software applications from compo-

nents (e.g., including database, web and application server components), and application
services (i.e. SaaS) that comprise web-based applications for enterprise use (Lenk et al.
2009). In this logic, a SaaS may be built on a PaaS and use IaaS, so that the SaaS
provider in turn can (but not necessarily has to) become a Cloud user (Armbrust et al.
2010). Altogether, it stands to reason that the increasing servitization of applications
and application components leads to entire ecosystems and supply chains of IT service
provision that may span an increasing number of providers and different types of services.
Today, SaaS is (with more than 50 percent) still by far the largest segment of the entire
market of Cloud-based services and is expected to remain at this position in the future
(Gens2008).
The market for SaaS generally covers most of the applications that are also available as
traditional enterprise software, e.g. Enterprise Resource Planning (ERP), Supply Chain
Management (SCM), Customer Relationship Management (CRM) as well as Content,
Communications and Collaboration (CCC) application types (Gartner2009). However,
some applications that are web-native, e.g. email, teleconferencing and web-hosting, are
obviously more likely to be procured via SaaS than those that require local hardware and
integration (e.g., engineering and design, production planning and automation systems). For
example, from the four mentioned application types, CCC and CRM applications are much
more frequently procured via SaaS than ERP and SCM systems (Gartner2009). The
literature largely explains this by stating that the main drivers for adopting SaaS
applications are lower application specificity, lower strategic value, lower uncertainty, and
higher imitability (Benlian et al.2009). Companies foremost expect cost advantages from
using SaaS, i.e. a variabilization of fix IT investments by renting software (Choudhary
2007b;Benlian and Hess2011). On the other hand, greatest inhibitors of SaaS adoption
are frequently the security risks from giving data control to an external party, e.g. caused
by data theft and data corruption (Xin and Levina2008;Benlian and Hess2011).
Altogether, since SaaS has entered the enterprise landscape and companies make ex-
periences in the use of SaaS, the IS literature has provided significant insights about the
factors of SaaS adoption. More recently, some authors have also begun to address the
management challenges imposed by the use of SaaS (e.g.,Khajeh-Hosseini et al.2010;
Bento and Bento2011). Given the eminent market expectations and the significance of
SaaS for client organizations, I consider SaaS as a vital subject to study the impacts of
IT-based innovation on IT governance. That is, since SaaS has passed the initial
stadium adoption, it should be feasible to observe potential differences in the way IT
artifacts are governed between SaaS using and non-using organizations. Furthermore,
enterprise applications and their organizational embedding are often seen at the core of
the IS discipline (Orlikowski and Iacono2001;King and Lyytinen2006). Thus, the selection
of SaaS as an IT-based innovation is expected to provide relevant insights for a broad
range of companies. Finally, I also consider this selection appropriate to allow for cross-
sectoral considerations in comparison with the M-Government scenario.5

5
Obviously, SaaS is not limited to the enterprise field (seeJanssen and Joha2011), and neither is the

8
 1.3 Thesis Structure

1.3 Thesis Structure

The purpose of this chapter was to introduce the research presented in this thesis and to
motivate why M-Government and SaaS are two examples of IT-based innovations that
are particularly suitable for studying the mutual relationship between IT governance and
innovation adoption. In the following, I provide an outline of the thesis structure. The
overall flow of the chapters is summarized in Table1.1.

Table 1.1:Thesis structure

Foundations of IT Governance and Organization Design
Dimensions of IT organization design (2.3)
Four IT organization archetypes (2.4)
Motivation for further research (2.5)

Methodological Foundations
Epistemological framing (3.1)
Qualitative methods (3.2)
Quantitative methods (3.3)

IT Governance and Innovation Innovation Adoption and IT

Adoption in E-Government Governance in Enterprise IS
Innovations in Mobile Government The impact of SaaS on IS authority
(4.1) (5.1)
Towards Transformational IT Comparing authority for on-premises
Governance (4.2) and SaaS (5.2)
Citizen Acceptance in Urban Sensing A process model for explaining
(4.3) governance of SaaS (5.3)
Municipal Benefits of Urban Sensing The dual role of IS specificity for
(4.4) governing SaaS (5.4)

Conclusion and Contributions

Theoretical contribution (6.1)
Practical contribution (6.2)
Methodological contribution (6.3)
Limitations and further research (6.4)

In order to better explain the context of this thesis (i.e., IT organizations and their
structural alignment within the wider organization), the following chapter (Chapter2)
provides an introduction to contemporary IT governance and organization design. The
chapter reviews the broader Information Systems (IS) and Management literature and
mobile channel reserved for the government sector (see M-Commerce,Siau et al.2001). However, I
argue that these two innovations are currently just in the right phase of diffusion to study the effects
of (and effects on) IT governance in public (and private) sector.

9
1 Introduction

proposes a conceptual framework consisting of six dimensions that are crucial in IT or-
ganization design. Since these dimensions are inherently correlated, we further integrate
them in a 22 framework that puts the allocation of IT decisions and IT resources
into the focus and explains the emergence of distinct organizational archetypes. The
understanding of these archetypes is an important basis for this thesis, inasmuch as the
following chapters will make reference to some of the underlying design dimensions. In
addition, this chapter points out further research opportunities regarding the contingen-
cies that determine the emergence of different organizational archetypes. It therefore
also serves as a broader motivation for the research conducted in this dissertation.
In Chapter3, I provide an overview of the different methodological foundations that
are required to conduct the research presented in dissertation. These include the use
of qualitative and quantitative empirical methods as well as design-oriented approaches.
The chapter is headed by a preamble, which briefly discusses the embedding of this work
in the philosophy of science and explains the epistemological view that is adopted in this
dissertation. Since the individual research contributions only briefly explain the method-
ological backgrounds of each study (i.e., the readers knowledge of the methodologies is
generally presumed), this chapter can be understood as a reference for the methodologies
used in this thesis.
The main part of this dissertation consists of two chapters, each containing four sub-
chapters that report on the studies conducted. Chapter4deals with governance and IT-
based innovations in E-Government, in particular Mobile Government (M-Government).
In the first subchapter (4.1), I investigate the adoption of a broad spectrum of M-
Government services among a sample of German municipalities. Based on the findings,
in subchapter4.2I explore four cases of M-Government adoption in detail and analyze
the role of IT governance in this context. To address some of the inhibitors prevailing in
municipalities, I shed more light on the citizen side of M-Government adoption in
subchapter4.3. More precisely I focus on the adoption of urban sensing, which rep-
resents an emerging class of external M-Government services. Finally, subchapter4.4
makes the proposition that more municipalities should consider M-Government services
in their municipal E-Government strategies by describing a concrete case of urban sensing
adoption and providing an in-depth investigation of the benefits achieved.
The second half of the main part (Chapter5) is devoted to IT innovations and gov-
ernance of enterprise information systems (IS), particularly for Software as a Service
(SaaS). The first subchapter (5.1) explores the potential impact of SaaS on IS authority
by analyzing four cases of SaaS adoption and proposes a contingency model to explain
application-level governance phenomena. This model is then refined and evaluated in
a large-sample survey where I also compare governance arrangements for SaaS and on-
premises software (Chapter5.2). Given the limitations imposed by such factor-based
approach, I revisit some of the cases from Chapter5.1and demonstrate a process-
theoretic approach to analyze application-level governance phenomena (5.3). Based on
this procedural conceptualization of SaaS adoption, I am able to resolve some of the
inconsistencies that emerged from the purely factor-based contingency perspective on
SaaS adoption in Chapter5.4. These findings unveil a dual influence of the specificity

1
0
 1.3 Thesis Structure

characteristics of SaaS application for allocating application governance and thus poten-
tially also IT governance.
The final chapter (Chapter6) summarizes the results of this research and discusses
its contributions. In particular, I outline the contributions regarding the key theoretical
constructs of this research, i.e. the mutual relationship of IT governance and innovation
adoption (6.1). Furthermore, I discuss practical contributions reflecting the role of IT
governance in public versus private sector organizations (6.2). Finally I outline method-
ological contributions that can be gathered from this work regarding the use of mixed
methods in Information Systems research (6.3). The dissertation concludes by outlining
the overall limitations and providing directions for further research (6.4).
In the remainder of this thesis, the subchapters will be simply referred to as chap-
ters. Forming parts of a cumulative dissertation, these chapters represent self-contained
research papers with separate introductions, theoretical foundations and conclusions.
Therefore, they can also be read independently from each other according to the interests
of the reader. At the time of publishing this thesis, the introductory chapter and the
eight research papers have been published or are still in the process of publication. For
clarification, each chapter will be introduced by a short preamble stating the authors and
status of publication, as well as making specific acknowledgments, if applicable. Table1.2
provides an overview of the published chapters of this thesis.6

Table 1.2:Overview of publications

Ch. Outlet Reference
2 Computer Science Handbook, Third Edition - Information Systems and (Winkler and
Information Technology, Taylor & Francis Brown2013b)
4.1 Internationale Tagung Wirtschaftsinformatik (WI) 2011 Proceedings (Winkler and
Ernst2011)
4.2 European Conference on Information Systems 2011 (ECIS) Proceedings (Winkler et al.
2011b)
4.3 European Conference on Information Systems 2012 (ECIS) Proceedings (Winkler et al.
2012a)
4.4 Journal of Theoretical and Applied Electronic Commerce Research (JTAER) (Winkler et al.
Special Issue on Smart Applications for Smart Cities: New Approaches to 2012b)
Innovation
5.1 International Conference on Information Systems (ICIS) 2011 Proceedings (Winkler et al.
2011a)
5.2 Working paper (under review) (Winkler and
Brown2013a)
5.3 Multikonferenz der Wirtschaftsinformatik (MKWI) 2012 Proceedings (Winkler and
Gnther2012)
5.4 International Conference on Information Systems (ICIS) 2012 Proceedings (Winkler and
Benlian2012)

6
The author of this dissertation is also the first author of all of the constituting papers. Nevertheless,
the narrative perspective will switch to plural (we) to express the joint opinion of all authors.

1
1
2 Foundations of IT Governance and
Organization Design

2.1 Preamble
This chapter presents a slightly shortened version of a book chapter byWinkler and
Brown(2013b) to appear at Taylor and Francis in the third edition of the Computing
Handbook Set Information Systems and Information Technology (Volume 2), edited by
Heikki Topi and Allen Tucker. I gratefully acknowledge the permission of my coauthor
for including this work as an introductory chapter of this dissertation as well as the
suggestions of the editors during the review process.

2.2 Introduction
How to organize and configure the internal Information Technology (IT) function1 has
been a critical issue since the beginning of enterprise computing. One of the most im-
portant challenges in IT organization design is selecting the extent to which IT decision-
making and IT resources (including the IT workforce) are centralized (Brown and Magill
1994). The key rationale for centralization is to leverage economies of scale; the underly-
ing rationale for decentralization is to ensure local responsiveness to internal and external
customers, including innovative solutions (Sambamurthy and Zmud1999;Agarwal and
Sambamurthy2002;Weill and Ross2004a).
Over the past decades, IT organizations have oscillated between centralized and de-
centralized forms (Peak and Azadmanesh1997;Evaristo et al.2005). In the beginning
of enterprise data processing, mainframe computers and magnetic tape devices were
commonly organized in central data centers. After the late 1980s and the vast growth of
distributed computing (Von Simson1990), client-server and firm-wide enterprise resource
planning applications led to IT re-centralizations (Brown2003;McAdam and Galloway
2005). Many firms further consolidated large parts of their IT infrastructure and appli-
cation operations into independent shared services organizations (Evaristo et al.2005).
These serve several lines of business to gain further economies of scale advantages as well
as to improve the quality of overall IT service delivery through introducing standard
IT practices (Schulz et al.2009). While recent IT reference frameworkssuch as ITIL,
1
The terms information systems (IS) and information technology (IT) are both used in the literature to
describe the IS/IT organization and IS/IT function. In this chapter we will use the term IT when
referring to an organizational unit performing all or some of the IT functions within an enterprise.

13
2 Foundations of IT Governance and Organization Design

ISO/IEC20000, CMMI and COBITprovide some guidance for designing the IT function
and internal processes (Pardo et al.2011;Marrone and Kolbe2011), this chapter takes
an enterprise-level perspective.
In this chapter we present four IT organization archetypes that differ based on the
centralization versus decentralization of both (1) IT decision rights and (2) allocated
IT resources. We describe these archetypes based on four additional design dimensions:
(3) coordination mechanisms, (4) financial autonomy, (5) sourcing arrangements, and
(6) IT-related capabilities and skills. Being mindful that in the past the form of organiz- ing
the IT function has been heavily dependent on technological development, we predict that
recent technology trends, such as cloud computing and the consumerization of IT, are
likely to affect IT organization designs of the near future.

2.3 Six Dimensions of IT Organization Design

Organizations (for-profit as well as non-profit) typically consist of multiple units that may
represent different functions or departments, lines of business, markets or geographies
(Daft2009). We use the term IT organization to refer to the collectivity of human
resources that perform IT-related tasks, such as planning, building and operating in-
formation technology applications and their underlying computer and communications
infrastructures, as well as the relationships, practices, norms, and capabilities of these
resources. This definition does not restrict the notion of an IT organization to the
existence of a single organizational unit (i.e., the IT department). Rather, it offers
the possibility to assume different design options for different IT units, depending on
the needs and capabilities of the business unit(s) supported. We also propose that six
important dimensions distinguish an IT organizational design, as described below.2

2.3.1 Allocation of IT Decision Rights

According to IT governance theory, decisions on information technology can be made

in a more centralized or decentralized fashion (Brown and Grant2005). In a corporate
setting, centralization typically refers to allocating decision making at the corporate level,
while decentralization refers to decision authority at the divisional level or even lower
organizational levels (Brown and Magill1994). A simple scheme includes two primary de-
cision areas: IT applications and IT infrastructure operations. A widely adopted pattern
in which infrastructure decisions are centralized, but business application decisions are
primarily made by the divisions, has been commonly termed a federated or federal model
(Sambamurthy and Zmud1999). More recently,Weill and Ross(2004a, p. 6) proposed
a five-part classification scheme that distinguishes decisions about business application
needs, IT investment and prioritization, IT architecture, IT infrastructure strategies, as
2
As the focus of this chapter is on explaining varying organization structures, we refrain from an in-
depth discussion of IT processes. However, we will make reference to process-based IT reference
frameworks and core IT processes where suitable.

14
 2.3 Six Dimensions of IT Organization Design

well as overall IT principles, with different patterns associated with different business
priorities. Defining accountability and the sharing of decision rights between the two
extreme poles of centralization and decentralization is commonly seen as a key challenge.
However, some studies have demonstrated that companies with well balanced IT decision
rights exhibit better business-IT alignment and thus ultimately achieve superior firm
performance (Weill and Ross2004a, p. 202). An IT reference framework such as COBIT
can be used to apply overarching accountability schemes to the design of decision rights
on the activity and role level.

2.3.2 Allocation of IT Resources

The second dimension captures the structural aspect of the IT organization, i.e., the po-
sition and location of the IT human and technology resources within the wider enterprise.
Although some prior literature has implied that IT decision rights and IT resources reside
together in an organizationwe argue that these two dimensions should be considered
separately ([Link] et al.1992;Brown and Grant2005). For example, IT decisions
may be made in a decentralized manner by business units, while IT resources operate
under either divisional or corporate IT authority. Similarly, IT staff may be allocated to a
line organization, but these IT resources implement services under centralized authority.
IT resource allocations have also been categorized as either IT demand or IT supply
resources (Thiadens2005;Mark and Rau2006). That is, divisional IT units may plan
for and formulate the IT resource demand for IT services at a division or business unit
level, although a central IT unit (or an external supplier) may have responsibility for
actually supplying the IT services to meet the specific business demand. Demand
activities for IT operations, for example, include monitoring the delivery of IT services
and issuing requests for minor changes to the infrastructure. Demand activities for IT
application development include business process analysis, requirements definition, and
user acceptance testing, as well as overall IT project management and steering. Although
the focus of reference frameworks such as ITIL and COBIT is standardizable IT processes
for IT supply units, they can provide some guidance also for designing demand-sided IT
activities. For example, ITIL defines a dedicated demand management process as a
responsibility of a demand manager (reporting to an IT unit).
In practice, the degree of centralization of IT resources differs widely under different
IT organization archetypes (Brynjolfsson and Hitt1998). In highly decentralized IT
organizations, divisional IT units also accomplish IT supply tasks, while in very central-
ized IT organizations, corporate IT groups also manage much of the IT demand. The
distribution of resources has overall been found to reflect the extent to which companies
pursue economies of scale, versus enabling local responsiveness through the allocation of
resources (Brown and Magill1994).
The first two dimensions of our frameworkallocation of IT decision rights and alloca-
tion of IT resourcesform the axes for the 22 matrix in Figure2.1. In addition to the
Centralized and Decentralized polar extremes, two other IT organization archetypes are

15
2 Foundations of IT Governance and Organization Design

defined. In the Shared Services model, IT decision rights are highly decentralized, but the
IT resources that perform IT tasks are highly centralized. In the Corporate Coordinator
model, the IT resources are highly decentralized or outsourced, but a central office holds a
higher degree of IT decision rights. Four additional design dimensions for characterizing
these archetypes are described below.

Resource
Corporate allocation Division

Centralized Model Corporate Coordinator

CEO CEO
CIO
office
Corporate CIO BU 1 BU 2 BU 1 BU 2

IT IT
dept. dept.

ESPs ESPs

Decision
allocation
Shared Services Decentralized Model
CEO CEO
Legend:
BU Business units and subunits
BU 1 BU 2 BU 1 BU 2
CIO BU 1 CIO BU 2 CIO Corporate IT units
(demand) (demand)
Division (supply) BU 1 BU 2 Divisional IT units
CIO CIO
ESPs External service providers
ESPs ESPs ESPs Reporting lines

Request / fulfillment flows

Figure 2.1:IT organization archetypes

2.3.3 Coordination Mechanisms

The mechanisms for coordinating IT tasks across multiple organizational unitse.g.,

corporate functions, business units or divisions, and/or corporate and divisional IT
groupsare an important complementary design dimension to the formal allocation of
decision rights and resources (Brown1999). They can be viewed as an overlay of the
structural organization, which enables horizontal, not just vertical, information sharing
(Daft2009, p. 95). In general, the more complex and dispersed allocations of decisions
and resources are, the more sophisticated coordination mechanisms need to be to effec-
tively coordinate and integrate across the different parties involved in decision making and
execution (Peterson et al.2000). Three categories of coordination mechanisms have been
emphasized in the literature: structural mechanisms, procedural mechanisms, and
relational mechanisms (Van Grembergen2004, p. 20).
Structural mechanisms include standing groups or committees (in contrast to tempo-
rary teams or task forces), and formal roles that link across different organizational units.
Widely used standing groups for IT governance decisions are, for example, IT steering

16
 2.3 Six Dimensions of IT Organization Design

committees with key business representation and IT management councils (Brown1999).

Formal liaison roles for IT demand management have also been implemented in many
organizations within both business and IT units, e.g., account managers and business
analysts reporting to IT units, as well as divisional information officers, business process
owners and key users residing in business units. Specific examples of tasks for such
committees and liaison roles are now also part of common IT reference frameworks.
Procedural mechanisms are the specified rules and standard practices for decision-
making and alignment between business and IT units (Peterson et al.2000). Processes
that span business and IT units include the IT strategy process, the IT budgeting and
investment review process, project controlling processes, system change request and
service level management procedures, etc. Naturally, both formal roles and standing
groups are highly involved in effectuating procedural mechanisms. Common reference
frameworks typically define a number of processes that involve these roles and groups
e.g., ITILs demand management, service level management, change management and
incident management processes.
Relational mechanisms characterize those practices that aim to link stakeholders in
different organizational entities informally (i.e., outside of their role description or formal
responsibility). Common approaches are communities of practice, key user networks,
physical co-location, temporary job rotations or simply interdepartmental events. While
IT reference frameworks largely neglect the less formalizable relational mechanisms,
academic researchers have emphasized the importance of informal mechanisms as a
necessary complement to formal mechanisms (Brown1999;Chan2002). For example,
relational mechanisms are apt to facilitate knowledge sharing and mutual understanding
among different stakeholder groups (Peterson et al.2000).

2.3.4 Financial Autonomy

The strategic management and accounting literature differentiates between different

forms of financial autonomy for divisional units, such as cost, break-even, profit and
investment center types (Anthony and Govindarajan2007, p. 247). Applied to (corpo-
rate and divisional) IT units, the type of center not only has important implications
for internal chargeback arrangements between business and IT, but also determines
the degree of financial and managerial autonomy of an IT unit (Venkatraman1997).
Reference frameworks such as ITIL and COBIT generally acknowledge the importance
of this organizational design dimension, but provide minimal design guidance.
In a cost center type, the IT unit is led by budget goals and is thus exclusively
accounting for the costs of delivering internal IT services. Chargeback mechanisms
are typically not in place (thus creating a possible incentive for business managers to
underfund their units).
The break-even center defines service-based chargebacks based on the actual costs for
delivering IT services. Thus, being a mixture between cost and profit centers, the goal
of this center type is to close break-even. Since IT costs (e.g., for personnel, hardware,

17
2 Foundations of IT Governance and Organization Design

software) are often not directly accrued to an IT service, more complex cost and activity
accounting schemes need to be established than in a cost center type. Such cost models
often approximate the actually incurred IT cost, combining direct and indirect costs
(Ryan and Raducha-Grace2009).
Profit centers have greater financial autonomy inasmuch as their management carries
responsibility for costs and (internal or external) revenues for IT services. Costs are
charged to the customers on a more competitive basis, often oriented towards market-
based transfer prices. However, in practice, business units are often obligated to contract
with an internal IT profit center, so the degree of market competitiveness with external
IT service providers is limited.
Investment centers extend profit center responsibilities to include accountability for the
investment of accrued capital, so that this type of IT unit can be viewed as an independent
company within the company. In large corporations, both profit and investment centers
are commonly constituted in separate legal entities, subsidiary to the parent company.

2.3.5 Sourcing Arrangements

IT decision makers continuously face the question about which tasks can be better and
more efficiently performed by an external party. The IS literature provides a large body
of knowledge with relevant considerations related to IT outsourcing (seeLacity et al.2009 for
an overview). Outsourcing arrangements can be differentiated regarding the coordina- tion
mode with an external providere.g., selected contractual obligations (arms-length
relationships) for cost efficiency, versus long-term relational partnerships (embedded)
for strengthening IT resources and technological flexibility (Lee et al.2004). Notably,
in recent years, the focus has shifted from long-term, comprehensive IT outsourcing
arrangements and purely economic considerations to contracts that also target quality,
flexibility and innovation goals (e.g.,Whitley and Willcocks2011). Recent literature
also emphasizes the need for in-house capabilities for governing the different kinds of
outsourcing relationships effectively (e.g.,Willcocks and Griffiths2010). One model of
nine IS capabilities for modern IT organizations, for example, includes four capabilities
that are directly related to managing outsourcing providers: informed buying, contract
facilitation, monitoring, and vendor development (Feeny and Willcocks1998).
While both ITIL and COBIT describe some processes and activities related to manag-
ing third-party services, from an enterprise design standpoint, the crucial concern is the
locus of outsourcing governance, i.e., whether sourcing capabilities are allocated at the
business level, the central, or the divisional IT side (Agarwal and Sambamurthy2002).
IT outsourcing decisions can also result in a change in decision rights for that particular
IT function, including decentralizing more such rights to business units (Brown2003).
For example, in situations where resources for IT demand already reside in business
units or divisional IT groups, this organizational configuration increases the outsourcing
readiness of these units and thus the likelihood that an outsourcing relationship will be
governed directly by the division. This may as well create more pressure on central IT

18
 2.3 Six Dimensions of IT Organization Design

units to compete with external providersespecially when business units are not obliged to
contract internally. Financially autonomous IT supply units that are organized as
subsidiary to their parent corporations (i.e., captive IT centers), can therefore also be
viewed as transitional structural arrangements prior to outsourcing IT supply to an
external party (Kreutter and Stadtmann2009). In such situations, building appropriate
demand-side IT capabilities may become a strategic priority (Feeny and Willcocks1998).

2.3.6 Capabilities and Skills

We define a capability as the application of knowledge, competencies and skills residing

in human resources, to accomplish given organizational goals (Peppard and Ward2004).
Our second dimension, allocation of IT resources, refers to the structural assignment of
human resources within the organization, whereas this dimension focuses on the
aggregate proficiencies that IT human resources within an enterprise need to have. The
IS literature has proposed different categories of capabilities in IT organizations. In
addition to the nine-capability framework ofFeeny and Willcocks(1998), a common
typology derived from marketing research distinguishes between inside-out, outside-in,
and spanning capabilities (Wade and Hulland2004). 3
Inside-out refers to capabilities that are internally focused, such as IT infrastructure,
IT development and (more generally) cost effective IT operationshere referred to earlier
as IT supply capabilities. Outside-in and spanning capabilities are externally oriented,
placing an emphasis on requirements and customer relationships, including IT planning
and change management, IT/business partnerships, market responsiveness, and external
relationship management. These capabilities are likely to be aligned closely with business
units and here we characterize them as IT demand capabilities.
Some more fine grained competency and skill categories can be found in both the
academic and practitioner literature, including a framework of 36 skills in five categories
(Zwieg et al.2006), skills related to roles in ITIL and CMM capabilities, as well as in
frameworks such as the Skills Framework for the Information Age promoted by industry
groups within the U.K. (SFIA2012). 4 With the increasing pressure of IT organizations to
compete on the product and labor markets, the development of appropriate IT demand
and IT supply competencies becomes a more important imperative. A wide range of
IT human resource practices, such as recruitment, training and retention, and proactive
career development can guide IT organizations to achieve this goal (Luftman2011).
Table2.1summarizes the seminal literature that has motivated our inclusion of each
3
AlthoughWade and Hulland(2004) refer to these as categories for resources, their definition of
resources as assets and capabilities that are available and useful in detecting and responding to
market opportunities or threats is congruent with the notion of capabilities used in this chapter.
4
The SFIA Foundation is a not-for-profit organization that exists to own, promote, develop and maintain
the Skills Framework for the Information Age. The members of The Foundation are UK Industry
bodies in the field of IT: BCS (The British Computer Society), e-skills UK (e-skills UK Sector Skills
Council Ltd.), The IET (The Institution of Engineering and Technology), IMIS (The Institute for the
Management of Information Systems), and itSMF (IT Service Management Forum).

19
2 Foundations of IT Governance and Organization Design

of the six design dimensions.

Table 2.1:Organizational design dimensions

Dimension Key design questions Selected literature
1)Allocation of Which decision rights are allocated to Brown and Magill1994;
decision business units, corporate, and IT Sambamurthy and Zmud1999;
rights stakeholders? Weill and Ross2004a;Brown
and Grant2005
2)Allocation of Which degree of centralization is Boynton et al.1992;
IT resources appropriate? Brynjolfsson and Hitt1998;
Where is the split between IT demand and Mark and Rau2006;Thiadens
supply resources? 2005;Daft2009
3)Coordination Which integration mechanisms (structural, Brown1999;Peterson et al.
mechanisms procedural, relational) are implemented? 2000;Chan2002;
Van Grembergen2004
4)Financial Which degree of autonomy is appropriate for Venkatraman1997;Anthony
autonomy IT units? and Govindarajan2007;Ryan
Which center type is implemented (cost, and Raducha-Grace2009
break-even, profit, investment center)?
5)Sourcing Which degree of external sourcing is Agarwal and Sambamurthy
arrangements appropriate? 2002;Lee et al.2004;Lacity
Which services are sourced externally? et al.2009;Willcocks and
Which organizational units govern sourcing Griffiths2010;Whitley and
relationships? Willcocks2011
6)Capabilities Which capabilities are needed for IT demand Feeny and Willcocks1998;
and skills and IT supply? Peppard and Ward2004;Wade
How are these developed within the and Hulland2004;Zwieg et al.
organization? 2006;Luftman2011

2.4 IT Organization Archetypes

In Figure2.1we presented the four basic archetypes of IT organization configurations that
are based on the first two dimensions described above: the distribution of IT decision rights
and IT resources. In the following, we describe these archetypes in more detail, including
their characteristics on the other four dimensions, their occurrence in practice, their
strengths, as well as some common challenges.

2.4.1 Centralized Model

In a centralized model, most IT decision rights are allocated to the corporate level
and IT resources are reporting to a central IT unit subordinate to corporate control
while serving multiple business units. An IT steering or advisory committee has been
recognized as an important coordination mechanism for ensuring business leader input
into IT decision-making (Brown2003;Huang et al.2009). Under this model, the IT
function is typically operated as a cost- or break-even center with simple chargeback

20
 2.4 IT Organization Archetypes

arrangements. For example, in a corporate setting, a combination of global and business

unit-related IT budgets may be managed together with project-level and person-day
based internal pricing. External contractors are typically governed by the corporate
IT unit. Therefore, central IT resources not only need to be equipped with IT supply
capabilities, but also with sufficient IT demand capabilities to identify business needs
and translate these into successful delivery by internal resources and external partners
(as applicable).
Centralized models were the primary type of IT organization during the early era
of mainframe computing and into the late 1980s when relational databases had arisen,
however, networking was still limited (Peak and Azadmanesh1997). A second wave
of centralization also occurred the mid-1990s as large firms initially implemented com-
plex enterprise system packages (Brown2003;McAdam and Galloway2005). Today,
centralized IT functions are also still the predominant model for small and medium
sized businesses (Huang et al.2009). Strengths of this model relate to an inherently
high degree of standardization and corresponding efficiency through the sharing of IT
resources and an underlying IT architecture across all divisions. Common challenges are
business responsiveness and often a (perceived) lack of business contribution, that is, the IT
organization may appear to act as a black box from a divisional perspective. Many
centralized models have experienced improvements in IT responsiveness by enhancing
both formal and informal coordination mechanisms, e.g. by introducing dedicated liaison
roles and cross-functional IT meetings (Brown1999;Huang et al.2009).

2.4.2 Decentralized Model

In a decentralized model, business units make IT decisions (divisional or lower level) and
are also responsible for managing IT resources. In the pure decentralized model, a central
IT unit does not exist, which means that today it can be viewed as an almost anarchic
configuration, with no or little coordination on a corporate level (Weill and Ross2004a, p.
58). In small divisions, coordination can even be achieved via informal, relational
mechanisms, costs may not be accrued as a separate IT budget and chargeback arrangements
may not be implemented. If decentralized models make use of external
suppliers/contractors, potentially for selected IT sourcing or project resources, these are
typically governed outside of corporate control.
The decentralized model became more common after the expansion of mini-computers in
the late 1970s, when most of the information processing took place in closed (propri- etary)
systems managed by local IT experts (Peak and Azadmanesh1997). The rapid growth of
desktop computing and more modern distributed computing architectures also facilitated
more decentralization (Von Simson1990). The disadvantages of this model as a pure
model lie in the cultivation of silo structures and a lack of IT cost transparency. Similar
downsides relate to the commonly undesirable phenomenon of shadow IT, i.e., the
existence of ad-hoc IT solutions built, used, and managed by the business without central
involvement or approval (Raden2005). However, decentralized configurations can still be
appropriate in cases where a strategic independence of a certain business division

21
2 Foundations of IT Governance and Organization Design

is desired, which may even include divestment-readiness (Leimeister et al.2012). This

model can also be appropriate for business functions where high innovation through
IT and autonomous IT use are a strategic imperative, for example in the research and
development departments in a technology-intensive industry.

2.4.3 Shared Services Model

In the Shared Services model depicted in Figure2.1, the IT resources are highly cen-
tralized, while the IT decision rights are primarily located at the division level. That is,
divisions share the usage of centralized resources to capture advantages associated with
a Centralized modelincluding economies of scale and scope and joint IT architecture
planningwithout giving up major decision rights to a corporate IT unit. The business
divisions typically also participate in steering committees and other decision-making
bodiessuch as cross-divisional IT boards responsible for IT architecture, IT application
prioritizations, and infrastructure managementto set priorities for all of the divisions
using the centralized IT resources. Shared services units are financially more autonomous
than a purely centralized model and are responsible for their own results (Schulz et al.
2009). IT organizations that transition to this model therefore often need to devote
significant efforts to productize their IT services on a competitive cost basis, so that they
can retrieve their costs with chargebacks to their business customers (Ryan and Raducha-
Grace2009). External service providers may also be contracted for, and governed by, the
shared services unit, especially for infrastructure services. However, depending on their
size and maturityand policies of the overall organizationbusiness divisions may also
have sufficient IT demand capabilities (and authorization) to independently contract out
to external parties and thus circumvent the shared services unit.
Some of the early roots for this model can be seen in the writings byVon Simson(1990)
and others in the 1990s, when organizations sought to better balance the advantages of
a centralized model with those of a decentralized model with hybrid approaches. One
hybrid approach was to create a federal model with IT application rights and resources
residing within the divisions or business units, but IT operations (rights and resources)
in a corporate IT unit. In contrast, in a pure Shared Services model, IT decision
rights are at the division level, but IT (both application and infrastructure operations)
resources are centralized. The global implementation of enterprise systems beginning
in the late 1990s, which required both centralized application maintenance and process-
based customizations, has been one of the catalysts for a wider acceptance of the pure
shared services model (Brown2003).
In many corporations today, shared services have therefore become a dominant model
to organize and deliver IT as well as other enterprise support functions (e.g., account-
ing, physical facilities management), which are therefore sometimes co-located with IT
(Schulz et al.2009). Companies thereby aim to combine the benefits of centralization
(economies of scale and scope) for IT applications and operations, with the benefits of
outsourcing (e.g., customer focus, quality orientation, and increased variable versus fixed
costs at the division level)without sharing the potential drawbacks of outsourcing to

22
 2.4 IT Organization Archetypes

an external supplier (e.g., supplier sustainability, loss of internal know-how, regulatory

compliance and data security concerns, etc.).
Sometimes this model is also seen as an opportunity to generate additional business,
or as a strategic step before entirely outsourcing IT operations. Until the mid 2000s,
many major corporations set up such IT subsidiaries with the primary goal of generating
external revenues during a time of tremendous IT expansion in developed countriesa
strategic trend that from todays perspective, with few exceptions, can be counted as a
failure (Kreutter and Stadtmann2009). Reasons for why many of these captive players
could not sustainably hold ground in an external market include the changing capability
requirements for internally versus externally competing service providers, and the rise of
mature IT outsourcing firms that utilize cheaper labor.
Some of the inherent challenges of this model also relate to the lengthier channels of
communication from business demand to IT supply units (delivery), which may need
to be coordinated across multiple division (and country) boundaries. For this reason,
sophisticated governance mechanisms, including service level agreements by business
units, as well as strong demand-side IT capabilities, are required in order to implement
this model successfully (Peterson et al.2000;Van Grembergen2004).

2.4.4 Corporate Coordinator Model

In the Corporate Coordinator model, IT-related tasks are performed externally or by
divisional resources (i.e., by divisional IT units or non-IT business users themselves),
while a central IT authority (office of the CIO, or in some cases a CTO5) governs through
IT decision rights and aligns the IT resource investments with an overall IT architecture
strategy. In the pure form, the office of the CIO is empowered to develop and enforce
standards and monitor adherence via the CIOs direct report to corporate management, but
does not possess dedicated resources to provide IT supply. Corporate IT standards differ in
extent and range, from the usage of certain technology platform and application standards,
to guidelines for risk management and security controls. The reliance on committees
and other coordination mechanisms to balance corporate and cross-functional priorities is
similar to the Shared Services model. However, in a Corporate Coordinator model, these
governance bodies are under the CIO, who has greater decision making rights. For
example, large IT development projects and sourcing arrangements to be managed at the
division level may require pre-approval from the CIO.
External providers are contracted centrally by the office of the CIO or by divisional IT
groups, depending if the service being sourced has firm-wide impacts (e.g., infrastructures
and communication) or only divisional impacts (e.g., consultants and IT specialists in
a project context). The CIO office acts as the mediator of external IT services, which
are charged back to the divisions based on the costs of provision. Financial autonomy
5
The Chief Technology Officer (CTO) role has evolved from research and development (R&D) man-
agement positions in technology-based industries and has recently also attracted more attention as a
point of strategic responsibility for long-term goals and guidelines for the use of information technology
within organizations (Hunter2011).

23
2 Foundations of IT Governance and Organization Design

of the internal, divisional IT units is generally low, costs are accrued to divisional IT
cost centers that are consolidated in divisional budgets, and no chargebacks take place at
the division level. However, global cost transparency is warranted through oversight by
the CIO and a global portfolio of divisional and corporate IT projects. For IT supplier
steering and internal as well as external coordination, the central CIO office needs to
develop strong demand capabilities (e.g., IT planning and change management, market
responsiveness, and external relationship management). IT supply typically takes place
through external suppliers or through divisional IT resources (as applicable).
The Corporate Coordinator model in its pure form is appropriate for several particular
contexts, of which we highlight three. First, establishing a CIO office is often used
strategically as a first step to advance from very decentralized configurations to more
centralized governance and transparency, before actually centralizing resources, consol-
idating infrastructures and achieving global scale. Second, for some business models
which are based on replication (i.e., different entities with low data integration needs
but similar business processes), a Coordinator model is the appropriate choice, due to
its ability to leverage standardization potentials and economies of scale in IT sourcing,
without integrating IT architectures (Ross et al.2006, p. 35). Examples for such business
models are diversified conglomerates as well as franchise companies.
Finally, the CIO office as a mediator of external IT services enables the ongoing IT
outsourcing and industrialization trend. That is, the more (diverse) services are
procured from the external market, the higher is the need for expert buyers to steer and
manage these providers in order to achieve the desired benefits (e.g., costs, flexibility and
innovation goals). Thus, establishing a Corporate Coordinator model can be a viable
alternative to building the distributed and costly demand capabilities in the business
divisionsas required for the Shared Services model.
The key challenge of the Corporate Coordinator model is its difficulty in effectively
implementing centralized IT governance to leverage economies of scale and standardiza-
tion via negotiations across division heads. This may explain why this archetypeas a
model for the entire IT organizationis still uncommon today in practice.
The four IT organization archetypes and their key characteristics are summarized in
Table2.2.

2.5 Motivation for Further Research

Past research has proposed traditional business drivers such as a firms competitive
strategy and structure as influencing the choice of the archetype of an IT organiza-
tion (Agarwal and Sambamurthy2002;Brown and Grant2005). For example, more
globalized firms seeking responsiveness to local markets are likely to decentralize some
IT rights and responsibilities, while smaller firms striving for economies of scale are likely to
centralize their IT decision rights and resources (Sambamurthy and Zmud1999;Weill and
Ross2004a;Huang et al.2009). However, more recent literature also emphasizes the
complementarity between organizational and technological architecture (Tiwana and

24
 2.5 Motivation for Further Research

Table 2.2:Key characteristics of four IT organization archetypes

Centralized Decentralized Shared Services Corporate
Coordinator
IT decision CIO with Business unit Business unit leaders Central CIO office
allocation senior leaders (federally) and enforcing standards,
management (separately) central CIO local implementation
support
IT resource IT resources in IT resources in IT resources in IT resources in
allocation corporate IT local divisions shared IT unit, few divisions or external,
IT demand resources few strategic IT
resources in CIO
office
Coordination Business Divisional IT Divisional IT Executive board,
mechanisms relationship heads, managers, central divisional IT heads,
(structural) managers, IT divisional account managers, architecture board
steering management cross-divisional IT
committee boards boards, e.g. IT
architecture board
Financial Cost or Cost center or Break-even,profit or Chargebacks for
autonomy break-even accrued to investment center, external IT services,
center, simple other budgets, productized cost centers for
chargebacks no chargebacks chargebacks divisional IT, global
(for small monitoring
divisions)
Sourcing ESPs governed ESPs governed ESPs governed by Firm-wide ESP
arrange- by corporate IT by divisional IT corporate IT or contracts governed
ementsa divisional IT by CIO office,
specialist ESPs by
divisional IT
Capabilities Good demand Demand from Ideal split of IT Demand capabilities
and skills capabilities in business, demand and IT in CIO office, supply
corporate IT supply supply capabilities capabilities in
needed capabilities in realized division IT (or
division IT externally)
Strengths Standardization, High Economies of scale Expert sourcing by
resource responsiveness and responsiveness, CIO office,
pooling, and local customer- standardization,
efficiency innovation, orientation, IT cost global IT cost
strategic transparency transparency,
independence strategic
independence
Common Lack of Lack of More complex Difficult to empower
challenges business value efficiency, low governance, longer CIO office, lack of
contribution, cost communication strategic IT
low flexibility transparency, channels, IT supply competence in
silo structures competes externally, business divisions
conflicting sourcing
governance
a
ESP = External service provider

25
2 Foundations of IT Governance and Organization Design

Konsynski2010).
We conjecture that recent technology trends such as cloud computing and IT con-
sumerization are likely to affect the IT organization models for both the IT demand
and supply sides. More specifically, cloud computing and the Internet-based delivery of
applications and components as a service will further push the border of what is core and
what is commodity across enterprise application landscapes (Bento and Bento2011).
Thus, on the application level, business units are more likely to manage their own cloud
applications in a more decentralized fashion and thus circumvent centralized investment
procedures (Winkler et al.2011a). At the back-end, fewer IT resources will be needed
for the operating infrastructure. However, managing the technological architecture and
integrating cloud-based services with internal and external infrastructures will pose in-
creasingly important challenges and the need for new capabilities.
Consumerization of IT superimposes the cloud wave. Employees with increasing IT
skills and access to sophisticated client devices for personal use expect to find IT tools
in their workplace that they already use in their home environments (Bernnat et al.
2010). As an answer to these new expectations, some companies have created policies for
allowing employees to bring their own devices, such as smartphones and tablet PCs, into the
work environment and integrate them. This represents a paradigm shift inasmuch as
employees are subsidized for using their own hardware and applications. Data security
and other related risks need to be diligently addressed by enforcing appropriate firm-wide
guidelines.
These and other technology trends suggest greater decentralization of IT responsibil-
ities and more hybrid IT governance designs in the future. More application as well as
infrastructure decision rights (e.g., on mobile device use) will shift to tech savvy business
users, while IT operations responsibilities are increasingly shared between internal and
external suppliers. Managing the diverging ecosystem of user IT demand and entire
supply chains of IT service provision will be one of the key IT governance challenges in
the future (McDonald2007). Enterprise-level organizational models that enable a better
integration and coordination across users, IT units and multiple suppliers will need to
be developed, which we expect to be reflected in future versions of standard IT reference
models (Pardo et al.2011;Marrone and Kolbe2011). 6
Beyond the technology contingencies, other perspectives also appear particularly fruit-
ful for investigating the changing shape of contemporary IT organizational configurations.
First, industry-specific approaches have largely been neglected in the past. Organizations in
the public sector, for example, national and local governments as well as non-profits in
healthcare and other industries, hold different principles for creating public versus
private value, which may also call for different principles of IT governance (Weill and
Ross2004a, pp. 185-214,Sethibe et al.2007). Second, given the increasing dispersion of
IT value creation across organizational ecosystems, the understanding of organizational

6
For example, in its 2011 version ITIL has introduced additional strategic processes and liaison roles
to address increasing coordination needs, such as a service strategy manager, a business relationship
manager, and a demand manager.

26
 2.6 Summary

configurations needs to be broadened to span entire IT value networks (Leimeister et al.

2010;Iyer and Henderson2012). This also implies that the extensive, yet separate
literature strands on governance of (internal) IT functions and governance of (external)
outsourcing relationships need to be united under a common frame. Third, such gov-
ernance arrangements may significantly vary depending on the kind of IT subfunctions
considered. Various authors have begun to investigate IT organization and governance
phenomena regarding certain subdomains, such as governance in system development
projects (Tiwana2009), application governance (Winkler et al.2011a), data governance
(Khatri and Brown2010), and infrastructure sourcing governance (Xue et al.2011).
Taking such modular views and aligning these with overall (networked) governance
schemes appears a promising field for future researchers. Finally, having argued that
organizational configuration is a dynamic phenomenon influenced by business and tech-
nology developments, we conclude that more longitudinal research is needed to study IT
organization design phenomena.

2.6 Summary
Organizing and designing the information technology (IT) function is a critical manage-
ment issue, and one that is influenced by business factors and technology trends. In
this chapter we focused on the organizational embedding of the IT function in the wider
enterprise. We presented a 22 matrix of archetypes for the IT function (Centralized,
Decentralized, Shared Services, and Corporate Coordinator models) that differ based on
the centralization versus decentralization of IT decision rights and allocated IT resources.
Then we described these archetypes based on four other design dimensions: coordination
mechanisms, financial autonomy, sourcing arrangements, and IT-related capabilities and
skills. Finally, we argued that current technology trends, including Cloud computing
and IT consumerization, increase the need for corporate IT coordination and thus are
likely to lead to more hybrid models in practice. We presented different perspectives
that appear particularly fruitul to explore contemporary IT governance phenomena.

27
3 Methodological Foundations
In this chapter I explain the methodological foundations for the research conducted in
this thesis. The methods applied in this research acknowledge the existence of different
epistemological paradigms. They can be broadly classified by qualitative and quantitative
empirical methods, depending on the nature of data that is used for analysis (i.e., word
versus number data).

3.1 Epistemological Preamble

The question of what is knowledge and how we acquire knowledgeepistemology

has has been a focal point of philosophy for more than a century since the period of
Immanuel Kant (Strig2011, pp. 757-762). In his Kritik der reinen Vernunft, Kant
arbitrated between the two conflicting and emerging positions of his time. These were
the strict empiricism, which demanded the creation of new knowledge (inductively) only
from what is directly observable, and the rationalism, which proclaimed reason and
(deductive) thinking to be the primary source of new knowledge. Parallel to disruptions
in mathematics and physics (for example, the proposition of a non-euclidean geometry),
later philosophers (the Vienna circle) arrived at a neopositivist (or logical positivist)
view, stating that science should describe and intend to explain what is positively
given, [Link] the tradition of empiricismthe perceptual observations by the senses
and impressions. However, in case certain laws emerge from these research activities,
these are thought of as models and results of our own thinking, rather than laws of
nature that determine our reality. The controversy, of to what extent certain laws can
actually be verified, was a vital contribution added by Karl Popper, who introduced the
principle of falsification. This principle can be regarded as the basis of modern science
in that it states that hypotheses about a population of subjects can never be veryfied, but
only falsified. This implies that what we consider as knowledge always remains with a
hypothetical and somewhat provisional character, as Popper himself acknowledged
(Strig2011, p. 778).
However, the propositions by Popper have also been subject to criticism in several
directions, particularly regarding the mapping between theoretical laws and observations,
as well as the role of language. Even physicists like Albert Einstein noted that the notions
and terms we use to denominate real-world phenomena are all creations of our own think-
ing and cannot be derived inductively from perceptual observations, and therefore the
worlds of the observable and of the theoretical are inherently separated (von Kutschera
1972, p. 489). This criticism can also be related to the emergence of constructivism, a

29
3 Methodological Foundations

philosophical direction originating from different scientific domains that emphasizesin

simple wordsthat what we perceive as our reality may be ultimately the result of our
own construction. In general, positions that take a stance against a positivist view are
summarized as antipositivist (or also interpretivist) views. Originating from sociology,
interpretivism focuses on understanding the meanings that social actions have for the
individuals being studied, rather than aiming to derive generalizable laws (Macionis and
Gerber2011, p. 32). Given that we first need to understand and describe the phenomena
we observe before we can derive lawful relationships, interpretive research today is widely
regarded as a complement to (neo-)positivst positions (Bryman2004). Early advocates of
such paradigmatic pluralism are frequently seen in Paul Feyerabend and Thomas Kuhn.
However, Kuhn also emphasized a revolutionary view of scientific progress where, after
periods of normal science, a paradigm can disruptively shift to another (Hoyningen-
Huene2002). In this sense, the emergence of different research paradigms can also
be viewed pragmatically, i.e. by asking which paradigm is more suited to explain the
phenomenon of interest.
In the Information Systems (IS) field, interpretive studies have become more common
after the seminal work ofOrlikowski(1993). Although there is often a tendency to
equate a positivist paradigm with quantitative methods and an interpretive paradigm
with qualitative methods, this equation obviously appears somewhat crude (Mingers
2003, p. 236). This is because the paradigmatic view taken by a researcher is assumed to be
independent from the nature of the data used. To make two simple counter-examples, the
positivist case study (Yin2003) is typically based on qualitative data, conversely
grounded theorya classic interpretive approach (Glaser and Strauss1967)explicitly
encourages to include both qualitative and quantitative data in the analysis. Besides
these dimensions, further dichotomies can be used to characterize a research approach,
particularly intensive vs. extensive, data-driven vs. theory-driven (e.g.,Mingers2003),
exploratory vs. confirmatory (e.g.,Boudreau et al.2001), process-theoretic vs. variance-
theoretic (e.g.,Newman and Robey1992), and behavioral vs. design-oriented (e.g.,
March and Smith1995;Hevner et al.2004).Gregor(2006) provides a taxonomy of five
classes of theory that can emerge from using different epistemological paradigms and
summarizes the prevailing approaches in the IS field. This taxonomy also provides an
argument for how different types of theory can be linked. For an in-depth discussion of
these issues I may refer to the given literature.
The majority of the studies in this thesis assume a neopositivist perspective. That is,
both qualitative and quantitative approaches aim to provide evidence and find lawful re-
lationships between constructs, which potentially first need to be identified and described.
In this sense, the qualitative (intensive) approaches seek to study in depth a small number
of cases, while the quantitative (extensive) studies aim to make empirical generalizations
based on an a large number of cases. However, in the sense of neopositivism, we ac-
knowledge that all relationships and models provided by this research only represent
one potential way of interpreting real-world phenomena (and not natural science like
laws). Given this epistemological premise, we will not make explicit reference to the
adopted research paradigm for each study, unless the study deviates from this paradigm

30
 3.2 Qualitative Methods

(which may be the case for Chapters4.4and5.3). In particular, Chapter4.4takes an

action design-oriented perspective that is more based on a paradigm of intervention than
observation (Sein et al.2011). Chapter5.3takes a process-theoretic perspective that can be
viewed closer to interpretivism (Newman and Robey1992).

3.2 Qualitative Methods

In the scope of this thesis, we may distinguish four types of qualitative research ap-
proaches: interviews, content analysis, grounded theory, and case studies. Note that
these four types are (by far) not exhaustive; for a broader overview see (Myers1997).

3.2.1 Interviews
Interviews are certainly the most common method in qualitative research for data ac-
quisition and typically used in combination with further analysis techniques. Interviews
should be well prepared by the researcher and guided by a script that may take an
unstructured, semi-structured, or structured (i.e., survey interview) form (Warren2002,
pp. 83-101). Accordingly, the researcher may consider the use of open-ended versus close-
ended questions.
In the IS field, interviews can be conducted, for example, with users of an information
system or experts regarding a certain phenomenon. The selection of subjects obviously
implies an important influence for the further direction and the results of a study (Warren
2002, pp. 87). Semi-structured interviews are the most common form in IS (Myers and
Newman2007). However, despite the significance of this data acquisition method,Myers
and Newman(2007) find that the use of interviewing is often not sufficiently reported
in IS studies and the potential challenges are often overlooked. Common pitfalls relate
to missing positioning of the interviewer and interviewee (e.g., clarifying the roles of the
interviewer and interviewee, establishing a trust relationship, explaining reasons for the
interview), the communication between interviewer and interviewee (e.g., language and
interpretations, biases resulting from the interviewers perception, variety of voices) as
well as to the content of the interview itself (missing flexibility in the script, omitting
potentially interesting topics, missing sensitivity to the context) (Myers and Newman
2007). Interviews are typically recorded (which may as well impose some biases) and
transcribed for further analysis.

3.2.2 Content Analysis

Content analysis refers to a bundle of techniques that aim to synthesize aggregated
insights from amount of qualitative data, e.g. transcribed interviews, organizational
documents, news feeds, as well as pictures or videos. These techniques have in common
that they try to reduce the complexity of information contained in the underlying data
by relating fragments from it to an (either predefined or emerging) set of categories

31
3 Methodological Foundations

(represented as codes). This process termed coding typically delivers a (hierarchical)

system of categories (i.e. main categories, categories, subcategories, etc.). Coding is
commonly seen as the first step before the codes (more precisely, the pieces of content
that are assigned to a code, i.e. for text material the quotations) undergo further
analyzes. These may be frequency analyzes (i.e., counting the number of occurrences of
specific codes), valence analyzes (i.e., assigning an emotional tone to the each occurrence,
e.g. either positively or negatively), or intensity analyzes (i.e., measuring the degree
of an occurrence on an appropriate scale). Reliability of human coding procedures is
of paramount importance, given that the goals is to produce relatively objective (or at
least inter-subjectively comprehensible) results. (Neuendorf2002, p. 141). Therefore it is
recommended that two or more coders perform coding and that coefficients of intercoder
reliability are reported (Neuendorf2002, p. 148).
The content analysis methodology also explicitly encourages researchers to conduct
further statistical analyzes on these evaluated set of codes. In this sense it can be
practically regarded as a quantitative analysis method based on qualitative data. For
example, joint occurrences of codes from different main categories, or occurrences for
different subjects (e.g. interviewees) are relationships that a researcher might want to
test. Recently also a number of methods from machine learning are used to aid researches
in conducting coding and analysis procedures for large sets of unstructured qualitative
data, specifically latent semantic analysis and data mining methods (e.g.,Indulska et al.
2012). However, content analysis is often charged with not being sufficiently qualitative
in that counting codes and applying statistical inference obscures the richness provided
by qualitative data and its interpretation (Morgan and Others1993).

3.2.3 Grounded Theory

Grounded theory (GT) was developed byGlaser and Strauss(1967) in the 1960s, a
period where logico-deductive thinking gained dominance and qualitative research was
often considered unscientific (Charmaz2003, p. 251). With their book The Discovery
of Grounded Theory, Glaser and Strauss intended to provide a new methodological
framework that especially addressed young students (the kids) who were still free
from preconceptions (Legewie and Schervier-Legewie1995). This framework provided a
procedure to generate substantive theory that is systematically derived from, or grounded
in, data, rather than being a new theory itself (therefore the name GT can be regarded as
somewhat misleading, as Strauss admits in a later interview withLegewie and Schervier-
Legewie1995).
GT foresees a rather prescriptive set of guidelines for both data acquisition and
analysis. First, acquisition and analysis are not thought of in a sequential, but rather
in a simultaneous manner where the researcher constantly compares new to old data
and generates new concepts from it, until a state of theoretical saturation is reached.
During this processcomparable to content analysisthe data is coded in three logical
iterations: open coding, axial coding and selective coding (whereas axial coding has been
added later byStrauss and Corbin1998). In simple words, data fragments are assigned

32
 3.2 Qualitative Methods

use
Categories Coding paradigm

are grouped in

Concepts Properties
possess
are aggregated to have

Open codes Dimensions

are assigned to

Data fragments

Figure 3.1:Grounded theory elements (authors representation, simplified entity rela-

tionship notation)

to codes which are then aggregated to concepts in open coding. In this activity, the
method demands theoretical sensitivity to the phenomena observed, particularly by
constantly comparing observations with other (theoretical) occurrences. Axial coding
relates concepts to each other by making use of a basic set of categories. Glaser refers
to this as coding families, which can be used flexibly to address the specific type of
phenomenon of interest (Glaser1978, pp. 73-82). Straussian GT is more prescriptive by
recommending a coding paradigm that comprises causal conditions, context, intervening
conditions, action/interaction, and consequences (Strauss and Corbin1998, pp. 96-97).
Selective coding is done after a core category has been identified. Then, all relevant
concepts can be grouped around this core category (related by the coding paradigm or
coding family), while others can be eliminated. The GT coding procedure and how the
different elements are related is depicted in Figure3.1.
Originating from the field of nursery studies (seeGlaser and Strauss1967), GT has
soon expanded into many other domains, such as sociology, education and psychology
(Charmaz2003, p. 252). In the IS field, GT has become more common after the entering
into a leading journal withOrlikowski(1993). Although GT describes a holistic approach
for conducting research (i.e., including a sampling strategy as well as an interpretive
epistemological attitude), the most common application of GT in IS field is analytical,
i.e. as a method for analyzing qualitative data (Matavire and Brown2008). In this
regard, it has also been combined with other qualitative methods, especially with case
study approaches (e.g.,Hughes and Jones2003;Fernndez2005;Strong and Volkoff
2010).

3.2.4 Case Studies

In the social sciences, a case study generally refers to the intensive study of a single
unit of analysis (e.g., an organization, a project, or an individual) oras wellto the

33
3 Methodological Foundations

comparative analysis of multiple units, i.e. a multiple case study. These methods
generally regard the cases as instances which are particularly revelatory to study a
broader class of problems within a wider population. The selection of cases, therefore,
should be based on the richness of information provided by each case, rather than (only)
statistical sampling (Merriam1998).Flyvbjerg(2006, p. 229) notes that atypical or
extreme cases often reveal more information because they activate more actors and more
basic mechanisms in the situation studied. For the selection of multiple cases,Yin(2003)
emphasizes that a replication logic is needed to determine whether the cases are intended
to strengthen or broaden the insights from each other. He speaks of a literal replication
logic when cases are selected to corroborate each other, while a theoretical replication
dictates to select cases that are expected to exhibit different conditions or outcomes. In
practice, it is acknowledged that the case selection and data acquisition also often have
to follow opportunistic criteria since researchers do not always have access to those cases
which are potentially the most revelatory (Patton and Others1990, pp. 182-183).
Case study research includes techniques for both data acquisition and analysis. It
recommends to triangulate between data from multiple sources, particularly interviews,
documentation, archives, direct observation, participant observation or artifacts (Yin
2003, p. 85)in this respect it is similar to GT, where everything is data. In contrast
to grounded theory, data analysis methods for conducting case study research are not
very prescriptive. The main requirement is rather that the outcomes exhibit a clear
chain of [logical] evidence asBenbasat et al.(1987, p. 374) suggests. For a multiple
case analysis,Eisenhardt(1989b) recommends to first perform within-case analysis and
then search for cross-case patterns. Different tactics can be applied, such as selecting
categories or dimensions and then looking for within-group similarities and intergroup
differences. Other tactics are to compare cases pairwise or to divide the data by sources
(e.g. interviews vs. documentation) (Eisenhardt1989b).
In the IS field, case studies are a common and versatile research method that can
follow different epistemological paradigms. A meta-review byMingers(2003) counts
that more than one fourth of the publications in major IS journals use some form of a
case study, which follow either positivist, interpretive or intervention paradigms. In a
positivist sense, case studies can be regarded as the ideal type of test for falsification,
i.e. if just one observation does not fit with the proposition, it is considered not valid
generally and must therefore be either revised or rejected (Flyvbjerg2006, p. 228).
From an interpretive stance, case studies are thought of to provide generalizations to
theory, rather than verifying or testing [Link](1995) outlines that interpretive
case studies can particularly contribute generalizations of concepts, networked concepts,
propositions, as well as implications for a particular domain of action. Finally, intervening
case studies, such as action research approaches, aim to change a real-world problem and
derive a generalizable solution from this, rather than focusing on pure observation (e.g.,
Maartensson and Lee2004;Sein et al.2011).

34
 3.3 Quantitative Methods

Table 3.1:Overview of research methods

Qualitative methods Quantitative methods

Simulationbc
Clustering /
Interviewsb

Structural
theory(b)c

modelingc
Grounded

study(b)c

subgroup
analysisc

analysisc
equation
Content

Surveyb
Case
Chapter4.1 x x x x x
Chapter4.2 x x x x
Chapter4.3 x x
Chapter4.4 x x x
Chapter5.1 x x x
Chapter5.2 x x x
Chapter5.3 a (x) x
Chapter5.4 a (x) x x

a
Data based on a previous study (x)
b
Data generation method
c
Data analysis method

3.3 Quantitative Methods

Table3.1presents and overview of the research methods used in this thesis. The quanti-
tative methods can be distinguished in survey, structural equation modeling, clustering and
subgroup analysis, as well as simulation approaches.

3.3.1 Survey

Surveys are certainly the most widely used method for acquiring quantitative primary
data in the social sciences. They can be regarded as means of mass communication
between a researcher and number of subjects (and vice versa, of course) (Churchill
and Iacobucci2002, p. 270). Surveys typically use a questionnaire that may capture
structured or unstructured data. More interesting for statistical analysis are obviously
structured (i.e. closed-ended) questions with fixed alternatives due to the possibilities for
statistical evaluation. Depending on the phenomenon studied and the planned method
of analysis, the researcher should draw attention on the types of (structured) questions
used. These can include binary, categorical, multiple-choice, scaled, numeric, or ranking
type of questions.
Most common for operationalizing psychometric models are Likert scales (Likert1932),
i.e. questions where the subjects are asked to evaluate their attitude to a statement
on a bipolar and (presumably) equidistant scale (often with 5 or 7 points).Churchill
(1979) proposed a rigorous eight-step procedure to develop scales for measuring latent
variable constructs. This approach focuses strongly on the interrelatedness of construct

35
3 Methodological Foundations

items and has therefore been criticized by later authors (foremostRossiter2002)in

simple wordsfor optimizing reliability at the costs of content [Link]
advocates the use of single-item constructs whenever it is justified by the nature (e.g., a
low level of abstraction) of the construct. However, increasing the number of items per
constructwhile causing greater survey length and potential redundancy of questions
statistically increases the predictive validity of a model (Diamantopoulos et al.2012). A
second important question related to this discussion refers to the use of formative versus
reflective indicators to measure latent variable constructs (see also next Section3.3.2).
The total questionnaire should have a clear structure and follow a logical thread (e.g.,
from the most general to the most specific) in order to sustain the subjects attentiveness
(Churchill and Iacobucci2002, p. 345). Before being administered, surveys should be val-
idated thoroughly and pretested with the potential subjects to increase understandability
and minimize later measurement errors (Hunt et al.1982). Surveys can be administered,
for example, via mail, online, telephone or in person. Researchers typically need to
balance criteria of sampling control (i.e., who answers), information control (i.e., the
quality of information and potential biases), and administrative control (i.e., time and
money) in choosing an appropriate administration method (Churchill and Iacobucci2002,
p. 296). Potential biases in the survey method may result from the sample frame (e.g., non-
representativeness, undercoverage), the responses (e.g., nonresponse, voluntariness,
social desirability, cognitive consistency) and the measurement (e.g., leading questions,
insensitive measures, ambiguity) (Podsakoff and Organ1986;Hartman et al.2002).
Such biases should be taken into consideration, both a priori in the study design and a
posteriori by testing for them.

3.3.2 Structural Equation Modeling

Structural Equation Modeling (SEM) refers to a class of multivariate analysis techniques

that combine the use of latent variables (LVs) with path analytic modelingfor this
reason they are also sometimes referred to as second generation multivariate analysis
techniques (e.g.,Fornell and Larcker1987). Coupling two traditionsa psychometric
emphasis of measuring latent (unobservable) variables by multiple indicators and an
econometric perspective of prediction through a directed graph of relationshipsSEM
enjoys high popularity across many disciplines, also due to the increase in software pack-
ages to perform such analyzes (Chin1998a). This prevalence can also be explained in that
a path model reflects our thinking in chains of causal relationships and thus facilitates
translating such theories into data analysis.1 Mathematically, this analysis method uses
a (usually)2 linear equation system with indicators (i.e., observable) variables connected
via LVs and coefficients, which are then estimated by the algorithm, subject to according
error terms.
1
Misleadingly, structural equation models have sometimes also been named as causal models, although
SEM does only ascertain statistical association, not causality.
2
More recently, also non-linear approaches to PLS have been proposed, e.g. for modeling U- and S-
shaped relationshipships (seeKock2010)

36
 3.3 Quantitative Methods

There are two different approaches to SEM, the covariance-based (CB) and variance-
based SEM (Jreskog and Wold1982). CB-SEM aims to estimate the model parameters (i.e.,
coefficients) by fitting the covariance matrix from the structural equations to the
empirical covariance matrix observed in the sample (Reinartz et al.2009). Therefore,
it is also considered as a hard modeling approach where normal and interval-scaled
distribution assumptions and several hundreds of cases are necessary (Tenenhaus et al.
2005). In contrast, the variance-based partial least squares (PLS) SEM estimates param-
eters by maximizing the variance explained for all LVs by iterating through a series of
ordinary least squares regressions (Reinartz et al.2009). This soft modeling approach
is thought to pose less distribution assumptions and generally works with smaller sample
sizes than CB-SEM (Tenenhaus et al.2005). Besides the fundamentally different goals
of both approachesi.e., confirming a model structure versus exploring and predicting
variable relationshipsthe suitability of both approaches has been widely discussed (e.g.,
Marcoulides and Saunders2006;Reinartz et al.2009;Marcoulides2009;Hair et al.2011).
This thesis makes use of PLS-SEM and briefly states in each study the motivations that
led to this choice.
SEM allows for some specifics in modeling and measuring theoretical relationships.
First, LVs in SEM (both CB-SEM and PLS-SEM) can be measured by reflective or for-
mative indicators, depending on the assumed direction of the effect. Reflective indicators
assume that the questionnaire item reflects the true value of the LV plus an error term.
Formative indicators, in contrast, compose the LV by a linear combination of their values
and certain weights, i.e. the error accrues in the LV value (Cenfetelli and Bassellier2009).
SEM also allows for the conceptualization and use of second-order (or even higher-order)
models where a first-order construct is reflected in, or composed by several second-
order constructs (Wetzels et al.2009;MacKenzie et al.2011). Furthermore, researchers
become increasingly interested in the study of interaction effects such as mediation and
moderation in SEM (Chin1998b). In brief, a mediation effect occurs when the inclusion
of a variable M (the mediator) in between a variable A and B leads to a weaker direct
effect from A to C. A moderation effect occurs when a variable M (the moderator)
influences the strength of a linkage between A and B. For PLS-SEM, interaction effects
can be modeled and tested in different ways (Hayes2009;Henseler and Chin2010).
When estimating a SEM model, a number of quality criteria should be assessed, whose
extensive discussion would go beyond the scope of this introduction. For PLS-SEM, the
measurement model (i.e., all observed variables and their relationships to the LVs) and
the structural model (i.e., the endogenous relationships between LVs) are commonly
assessed separately (Chin1998b). Quality criteria for reflectively measured LVs relate
to convergent validity (i.e., whether the indicators measure the same construct) and
discriminant validity (i.e., whether distinct LVs and their indicators differ sufficiently).
Opposed to this, formative indicators are included in the model by definition due to their
constituting meaning for a construct, so thatsimilar to an ordinary regressiononly
multicollinearity should be ruled out (Cenfetelli and Bassellier2009). The structural
model is assessed based on the parameter estimates for the path coefficients. Unlike
to CB-SEM, PLS-SEM does not directly provide p-values for statistical significance.

37
3 Methodological Foundations

Therefore t-values are generated by a bootstrapping (or jackknifing) procedure that

repeatedly draws pseudo samples from the base sample and thus obtains a distribution of
the parameter estimates (Hair et al.2011). Note that in terms of a falsification (Popper), for
two-sided tests the null hypothesis is that there is no linear relationship (i.e., path
coefficient=0), so that conversely a non-significant path does not inform whether the real
world is free from such relationship.

3.3.3 Clustering and Subgroup Analysis

In this category I summarize a few further group analysis methods that are used in this
thesis in combination with SEM. These methods aid in both the discovery of previously
unknown groups by clustering, as well as the analysis of categorical (i.e., previously
known) subgroups of the sample. Clustering techniques detect groups of objects in a
larger sample based on their similarity characteristics (assessed by appropriate distance
measures,Johnson and Wichern2007, pp. 673-678). Hierarchical clustering (also called
linkage or connectivity-based clustering) methods either start from the individual objects
or the group of all observations and succeed by aggregating (or partitioning) objects
(groups) by iteratively merging (de-merging) the two objects (subgroups) that are closest
to (farthest from) each other (Johnson and Wichern2007, pp. 680-696). Wards linkage
(the type of clustering method applied in this thesis), is a special hierarchical clustering
method that considers the loss of information from joining two groups, usually measured
by an increase in the error sum of squared deviations from the cluster mean (centroid)
(Johnson and Wichern2007, p. 692). Wards linkage typically produces clusters that can
be elliptically shaped and have more equal cluster sizes compared to other hierarchical
methods (Backhaus et al.2003, p. 516). This makes this clustering method somewhat
superior for studying multivariate observational data in the social sciences (Punj and
Stewart1983).
A second method used in this thesis refers to multidimensional scaling (MDS) (Chap-
ter5.4). MDS is similar to clustering in that it assesses objects based on their similarities;
however, in MDS these classes of objects are determined a priori by the sample. MDS
allows for displaying multivariate data by transforming it to a low-dimensional (i.e.,
typically two- or three-dimensional) space and thus visualizing similarity characteristics
(Kruskal and Wish1978;Kappelhoff2001). MDS uses an iterative algorithm. Based
on a start solution (x, y), objects are compared pairwise by their compound similarity
relative to their spacial distance. If two dissimilar objects lie relatively close to each
other, they are moved apart (conversely if two similar objects lie relatively far from each
other, they are moved closer). This procedure is repeated until the configuration of
objects sufficiently reflects the similarity characteristics. The criterion for evaluating the
goodness of a configuration is called STRESS and is usually measured by the explained
variance from a regression of object dimensions on the position vector (x, y). The
stop criterion can be the number of maximum iterations or a minimum increase of
STRESS from one iteration to another (Kruskal and Wish1978, p. 25). Altogether,
both clustering and multidimensional scaling (as well as other analysis techniques) allow

38
 3.3 Quantitative Methods

to explore subgroup characteristics and potential heterogeneity in multivariate a data

such as latent variable factor scores from a PLS-SEM model.
Another criterion for multigroup comparison in SEM is the dissimilarity of path coef-
ficients (Johnson and Wichern2007, p. 678). This approach is different from clustering
or MDS as it focuses on the strengths of effects on a dependent variable in each sub-
sample, rather than the characteristics of the sample (or latent variable scores) itself.
Testing for path differences can be accomplished by a special t-test statistic using the
standard errors from separate bootstrapping procedures, as demonstrated in (Keil et al.
2000, p. 315). More recently, an alternative test has been proposed that relies on the
observed distribution of the bootstrap values instead of the standard errors (Henseler
et al.2009). Both of these tests are applied in Chapter5.3. Finally for completeness
analogue to clustering techniques for the factor valuesmore sophisticated segmentation
techniques, most prominently finite mixture models, also allow to form subgroups based
on (presumably) heterogeneous path coefficients in the sample (Jedidi et al.1997).

3.3.4 Simulation
Simulation as a research method generally refers to studying the behavior of a system
or a process over time, usually by the use of computational means. Similar to statistical
models such as in SEM, simulation approaches use some kind of model with defined input
(i.e., exogenous), auxiliary (i.e., endogeneous) and output (i.e., dependent) variables.
However, simulations produce data rather than analyzing or testing it. The relationships
between these variables can take any mathematical complexity, e.g. represented by non-
linear, conditional, stochastic, and differential equations. Differential equations are
especially important since they make simulation models time-aware, i.e. a variable may
depend on the (differential) change, or the accumulated sum (integral) of past values of
this or another variable over time. Therefore simulation models can also exhibit loops,
which poses a complexity that usually makes it impossible to find a closed-form analytical
mathematical representation.
Simulation approaches can target at both a better understanding of a system behavior
and/or the prediction or certain variables of interest. As it is in the nature of modeling,
usually a number of assumptions need to be made regarding input variables and the
proposed variable relationships. Proponents of the simulation method particularly put
forward that the process of modeling and running a simulation model itself increases the
understanding of potentially complex systems (e.g.,Stave2002). Additional analyzes can
be conducted based on a simulation model, for example scenario analyzes (i.e., comparing
simulations with different underlying assumptions), sensitivity analyzes (i.e., quantifying
changes in an output variable to changes in an input variable), or monte-carlo simulation
(i.e., assuming probabilistically distributed input variables and studying the distribution
of output variables). Individual-based simulations (also multi-agent systems, which are
not in scope of this thesis) model the behavior of autonomous agents in order to assess
the effects on a system, rather than directly describing the system as a whole (Bousquet
and Le Page2004).

39
3 Methodological Foundations

In the IS field, simulations are one of the less frequently used outlier methods (Mingers
2003, pp. 243, 245). Notably, other fields that are as well characterized by multi-
disciplinary and social, technical and organizational viewsfor example health care
researchmake distinctively larger use of simulation either as a primary or as a secondary
(i.e. supporting) research method (Brailsford et al.2009). One of the reasons for this
paucity in the IS field may be that reviewers have less confidence in the (approximate)
validity of simulation models, and researchers respectively do not sufficiently validate
their simulation models (Forrester and Senge1978). Nevertheless, proponents argue that
researchers can use simulation approaches to go beyond the role of the individual and
study the different interactions among agents in different organizational levels (Bousquet
and Le Page2004).

40
4 IT Governance and Innovation Adoption
in E-Government

4.1 Innovations in Mobile Government

4.1.1 Preamble

This chapter has been initially published and presented at the German Tagung Wirt-
schaftsinformatik (WI) January 2011 in Zurich, Switzerland (seeWinkler and Ernst
2011). Some of the formulations and statements may deviate from the original paper
due to the translation from German to English.

4.1.2 Introduction
The ongoing technological development of mobile broadband networks also results in a
positive momentum for mobile Government (M-Government). It is expected that the
number of people in Germany who regularly use Internet functions on the mobile phone
will triple to more than 30 million from 2010 to 2012 (Computerwoche2009;Bitkom
2010). While in the past M-Government was largely confined to simple services such as
SMS1 notifications and isolated intra-governmental applications (Trimi and Sheng2008),
new scenarios are emerging for the interaction between government and citizens, such
as location-based reporting services, mobile library services, and intelligent car-routing
systems.
Within the framework this study, the terminology and scope of M-Government will be
understood broadly in relation to different actors and forms of mobility. Implementing
M-Government services and applications can lead to significant changes in administrative
processes. This issue receives additional relevance through current initiatives that are
driven by national level and have both organizational and technological impacts on
local M-Government implementations. Recent examples for such initiatives in Germany
include the introduction of the single service number 115, the implementation of the EU
Services Directive, and the introduction of the German electronic identity card.
Simultaneously to fulfilling such strategic requirements, many municipalities face mas-
sive cost pressures. For example, in North Rhine-Westphalia nearly two-thirds of mu-
nicipalities need to manage their budgets according to a budget-balancing concept since
they are significantly underfunded (NRW2009). This limits the scope of action for
1
Short Message Service

41
4 IT Governance and Innovation Adoption in E-Government

implementing innovative IT projects and leads to a more stringent review of economic

efficiency. Accordingly, the core question of this study is whether M-Government can
be regarded as a further stage of E-Government and thus a path to greater municipal
efficiencyor whether it remains to be a marginal issue for those municipalities that still
have some room for maneuver for implementing innovative IT projects?
While business related research often studies innovation adoption from the perspective
of the enterprise (IT) decision maker, the E-Government literature has mainly viewed the
citizen in the center of the adoption decision. Thus, applications that are used by
government employees and achieve an intra-governmental benefit have often been
neglected. Our study represents a novel approach insofar as we put the municipal IT
decision makeras an important stakeholder and source of stimulusinto the center
of attention to explain municipal innovation adoption. Consequently, M-Government is
understood as a bundle of potential services and applications and we consider a variety
of possible application scenarios, rather than a single (citizen-centric) technology. This
article makes a relevant contribution by pointing out (1) the organizational factors and
conditions that affect a municipality in the implementation of mobile services, (2) how
much these factors affect the perceived potential of mobile government services, and
(3) what effect this has on the municipalitys investment behavior.
The remainder of this chapter is structured in six sections. After a theoretical founda-
tion and hypotheses development in Section4.1.3, we present the results of a qualitative
pre-study in Section4.1.4. Building on this, we explain the methodology for the empirical
study in Section4.1.5and analyze the results in Section4.1.6. The last Section (4.1.7)
summarizes this work, discusses practical implications and provides an outlook for future
research.

4.1.3 Theoretical Foundations and Hypotheses Development

In the following we explain the terminological and theoretical foundations for our research
on M-Government and develop the research model, see Figure4.1.

Forms of Mobile Government

Following the definition ofKushchu and Kuscu(2003), M-Government can be understood

as a strategy and its implementation involving the utilization of all kinds of wireless and
mobile technology, services, applications and devices for improving benefits to the parties
involved in E-Government. These parties do not only include citizens, but also involve
businesses and employees in public institutions, as well as the governments as such.
According to the common terminology that has been established for E-Government,
we can also distinguish M-Government according to (mobile) government-to-Customer
(G2C), government-to-business (G2B) and government-to-government (G2G) patterns,
whereas the latter has also been termed as IEE (internal effectiveness and efficiency,
Trimi and Sheng2008). M-Government can thus be understood as an extension and
subset of E-Government.

42
 4.1 Innovations in Mobile Government

Technologies used in mobile government go far beyond the capabilities of telecommu-

nications. Wireless networking, Bluetooth, CCTV (closed circuit television), location-
based services, RFID (radio frequency identification), biometric identification, traffic
monitoring, smart cards and NFC (near field communication) applications are just some
examples of mobile (i.e., non-stationary) technologies, which are not necessarily used
in conjunction with a mobile phone (Kumar and Sinha2007). In a study byWinkler
et al.(2009) we presented eight application clusters for M-Government services in an
urban context, which can be arranged on a continuum between the public value and
private economic benefits. This framework covers the sectors of public administration,
public safety, education, health, transport, environment and infrastructure, tourism and
culture, as well as applications for private households. This classification will also be used
in course of this study as a framework to operationalize the concept of M-Government.

Behavioral and Theoretical Foundations

The literature on E-Government adoption and usage primarily draws on conceptual

models that are based on innovation diffusion theory byRogers(1962), Fishbein and
Ajzens (1975) theory of reasoned action, and Davis (1989) technology acceptance model.
Since the latter focuses on specific characteristics (particularly perceived usefulness and
ease of use) of a concrete technological innovation, rather than a bundle of applications,
we presume that it is less suitable for a holistic investigation of M-Government adoption.
Therefore, we limit our theoretical perspective to the former two approaches.
According toRogers(1962), an innovation is defined as the acceptance of an idea
or a practice over time by organizations or individuals that are connected in the form
of communication channels, social structures and a system of culture and values. The
process of innovation goes through five phases of knowledge, persuasion, decision, im-
plementation and confirmation. In this study we assume that developments in the field
of M-Government are at presentin contrast to E-Governmentpredominantly still in
the first three phases. Therefore, the study conceptually targets at the act of persuasion
and decision for (or against) certain M-Government services.
The process of decision can be explained in more detail with the aid of the theory of
reasoned action (TRA,Fishbein and Ajzen1975). According to TRA, the behavioral
intention of using a particular innovation is a mediator between the (objective) attitudes
regarding this innovation and the behavioral outcome (i.e., adoption or non-adoption).
The TRA was originally developed to explain the behavior of individuals, while in the
complex structure of a local government there are presumably a number of actors involved
in a decision for or against technological innovations. Nevertheless, we still consider
the TRA to be applicable to a bundle of attitudes and behavioral intentions, provided
that these can be adequately measured. In doing so, we follow several examples in
the IS literature on organizational adoption, e.g. (Benlian et al.2009). However, on
an organizational level we assume that the influence exerted by subjective norms loses
relevance for the opinion of a group of individuals (Fishbein and Ajzen1975).
With regards to the provision of mobile services, we interpret the attitude of the

43
4 IT Governance and Innovation Adoption in E-Government

Persuasion phase
IT efficiency Decisio n phase
goals
H3 (+)

IT innovation H4 (+) Perceived H1 (+) Perceived service H2 (+)

Planned investment
goals potential attractiveness

H5 (+)
IT sophistication

Figure 4.1:Research model

municipal decision-makers as an aggregate measure of the perceived potential benefits of

M-Government. These may be influenced by the municipalities strategic goals and the
organizational context. The behavioral intention of adopting (i.e., using internally and/or
offering mobile services to citizens) shall be operationalized as an aggregate measure of
the perceived attractiveness of specific M-Government offerings. The intended behavior
of a municipality to introduce a mobile service should then ultimately be reflected in the
planned investment in M-Government services. Altogether, we postulate the following
hypotheses:
H1: There is a positive relationship between the perceived potential of M-Government
in general and the perceived attractiveness of specific service offerings.
H2: There is a positive relationship between the perceived attractiveness of specific ser-
vice offerings and the planned investment in M-Government services.

Antecedents of Perceived M-Government Potential

Mobile Government has only more recently become subject of academic research so that
we acknowledge a general lack of empirical works in this field (Kushchu2007, p. 1).
For this reason, we make reference to the related E-Government literature as well as to
studies in the field of strategic IS and IT investment decisions to identify appropriate
preconditions and factors influencing perceived M-Government potential.
There are a number of empirical studies in the E-Government field that examine the
offering and the acceptance of innovative services, foremost government websites (see
Patel and Jacobson2008for an overview). The large majority of these studies put the
citizen as an end-user in the center of the adoption decision. Consequently, studies
focus on individual antecedent factors such as trust in E-Government, IT experience
of computer users and IT skills. Also, demographic characteristics such as gender and
educational level are often considered as key factors. The work byMoon and Norris
(2005) represents an exception inasmuch as they study organizational factors such as the
size of the municipality and type of government (i.e., council-city manager versus mayor-
council governments) to explain different levels of maturity in introducing E-Government

44
 4.1 Innovations in Mobile Government

across a sample of U.S. municipalities.

Further organizational antecedent factors may be derived from the the IS and IT
alignment literature. The IT functions of municipal government are comparable to
those in private sector firms, as both have to meet certain targets depending on the
strategic goals and operational (business) requirements. Thus, we assume that there are
also certain strategic guidelines and conditions for the introduction of M-Government
services, depending on the situation in each particular municipality. Typologies of such
IT strategies in the literature have often employed (parsimonious, yet not too simplis-
tic) tripartite models (Denford and Chan2007). For example, (Sabherwal and Chan
2001) mirror the well-known Miles and Snow typology of business strategies (defenders,
analyzers, prospectors) to characterize an IT/IS strategy by the three attributes IS for
efficiency, IS for flexibility and IS for comprehensiveness. Later authors get to similar
partitions (Denford and Chan2007). We combine the idea of having a triad of strategic IT
attributes with the specific characteristics of E- and M- government adoption and pos- tulate
three dimensions related to efficiency goals, innovation goals and IT sophistication as
important antecedents of M-Government acceptance.
Efficiency goals relate to the municipalitys motif to support administrative processes
through the use of mobile applications and thus ultimately save time and money. De-
pending on the budgetary situation of the considered municipality, we expect a variation
in the in level of this dimension. The degree to which a municipality pursues efficiency
goals primarily reflects the economic perspective in M-Government acceptance.
Innovation goals express the degree to which a community aims to extend its service
offering in terms of effectiveness (yet not necessarily efficiency). Cities do not only
compete with each other, but are also exposed to different expectations of their customers
(seeKushchu2007). We argue that in this dimension individual drivers of acceptance,
such as the increasing technological affinity of citizens and employees and their evolving
IT skills, also play an important a role (Patel and Jacobson2008). New service offerings
supported by mobile technology can help satisfy the continuous customer demand for
innovation. Thus, this dimension addresses in particular the external perspective.
IT sophistication captures the IT-related prerequisites that are present in a munic-
ipality. This dimensions refers to both the physical and the soft infrastructure, as
Kushchu(2007) notes. Thus, equally important to existing systems and networks are
therefore institutional arrangements and a technological vision for M-Government. This
claim is consistent with the findings ofTornatzky and Klein(1982) who state that
compatibility with existing structures is crucial for a variety of types of innovation.
We expect that municipalities that already make substantial efforts for E-Government
implementations also have significant synergies when adopting M-Government services
(Kushchu2007). This dimension thus represents in particular the technological and
organizational perspectives.
Altogether, these considerations lead us to the following three hypotheses, which are
summarized in the research model shown in Figure4.1.
H3: There is a positive relationship between the efficiency goals of a municipality and

45
4 IT Governance and Innovation Adoption in E-Government

the perceived potential of M-Government services.

H4: There is a positive relationship between the innovation goals of a municipality and
the perceived potential of M-Government services.
H5: There is a positive relationship between the IT sophistication of a municipality and
the perceived potential of M-Government services.

4.1.4 Qualitative Pre-Study

To assure content validaty of our research hypotheses and operationalize the proposed
constructs we conducted a series of interviews and performed a content analysis.

Participants and Method

This pre-study analysis was oriented in the method proposed byNeuendorf(2002) and
conducted in two phases. We first established a categorical system according to the
variables and dimensions in our research mdoel (Figure4.1). In the first phase (test
phase), three interviews were conducted with E-Government experts from the municipal
administration of a large German city as well as a technology foundation that collaborates
with public and private institutions. For the purpose of triangulation, it seemed appro-
priate at this stage to select interview representatives from both the departmental (i.e.,
demand) and IT (i.e., supply) sides in the same municipal context. Table4.1provides
details about the job positions of the interviewees. We used an interview guideline based
on the categorical system which included questions regarding the hypothesized aspects
of each category as well as open-ended questions. The interviews were conducted as
presence meetings interviews of approximately 60 minutes and recorded digitally. The
analysis of interview transcripts was performed using the software [Link] for qualitative
analysis. Following an inductive approach, the category system was revised and further
sub-categories were developed.

Table 4.1:Interviewees and city sizes

Phase Position City size
1)Test Head of IT Competence Center > 200.000
phase Head of Department of Media, Information > 200.000
and Communication Industries
Head of Information and Communication > 200.000
Technology (ICT)
2)Coding Head of Staff Function E-Government 180.000
phase Head of Department (for Personnel, 70.000
Organization, IT, and Education)
IT Organizer 60.000
Head of Data Processing 60.000

In the second phase (coding phase) we conducted interviews with IT executives from
four other German municipalities. These phone interviews were conducted on the basis

46
 4.1 Innovations in Mobile Government

of the revised category system with a slightly adapted guideline (time approximately 45
minutes). Similar to the first phase, the coding of interview transcripts was performed
independently by two coders and discussed in the case of disagreements. After revising
the total 488 codes, we counted an agreement of 71 percent, which represents a good
intercoder reliability (Neuendorf2002, p. 141).

Operationalization of Antecedents

The subcategories resulting from the content analysis and number of codes are presented
in Table4.2. Since for the further analysis we only considered the three most frequently
mentioned codes, our description will be limited to the these subcategories. Further sub-
categories referred to, for example, the availability of partners, intercommunal networks
and synergies with other municipalities.
According to the interview results, the dimension of efficiency goals can be disag-
gregated into indicators that we termed process improvement, win-win situation and
efficiency improvement. Process improvement refers to the motivation to use mobile
services to improve certain processes in urban administration. Win-win situation em-
phasizes the aspect that this should not only achieve a benefit for the municipality, but
also simultaneously for its customers (i.e., citizens and local businesses). Efficiency
improvement particularly encapsulates such motivations that explicitly result from the
need to save on costs.
Innovation goals are at first reflected in the external demand for providing novel
services, referring to the pressure explicitly exerted by the municipalitys customers.
Secondly, it comprises one indicator that we label service enhancements summing up
the motivations to offer new mobile services from out of the administration in order
to expand the current serivce portfolio and present itself as a modern and innovative
municipality. Finally, we also attribute the increasing mobile consumerization to this
dimension, describing the development to which mobile devices and according service
services become part of peoples (i.e., citizens and employees) everyday lifes, which also
poses new opportunities and challenges to be addressed by municipalities.
As the first dimension of IT sophistication, E-Government platform captures the degree to
which certain administrative processes are already supported (or enabled) by IT so that
the municipalitys services are accessible online. Building on a sound E-Government
platform, we expect, will also facilitate the implementation of novel mobile services, since
existing interfaces and technologies can be reused. In this context, interviewees also
pointed to the importance of the existence of a comprehensive service strategy and an IT
strategy with E-Government and M-Government elements, as these would facilitate the
alignment of multiple stakeholders and enable a long term planning of the municipality.

47
4 IT Governance and Innovation Adoption in E-Government

Table 4.2:Operationalization of antecedent factors

Dimension Subcategory Question (How do you rate the motivating influence of the
following factors to realize m-government services in your city?)
Efficiency [Link] the need of the administration to improve the work processes
goals (EF) improvement
[Link]-win achieving a win-win situation for the municipality and
situation potential users
[Link] the pressure to save money by increasing administrative
improvement efficiency
Innovation [Link] the expectations from citizens and business to improve the
goals (IN) demand administration
[Link] the development of new municipal service offerings enabled by
enhancements mobile technologies
[Link] con- the increasing technological affinity in the population
sumerization
IT sophisti- 1.E-government administrative processes that new mobile services are based on,
cation (IT) platform are already implemented electronically
[Link] your municipality has a comprehensive plan for future service
strategy offerings, which considers the use of modern information and
communication technology channels
[Link] strategy your municipality has a comprehensive IT strategy that
contains E-Government and M-Government elements

4.1.5 Empirical Study

Questionnaire Design

To test the proposed model, a comprehensive questionnaire was developed and validated.
At the beginning of the questionnaire, we asked for context information regarding the to
the city (inhabitants and municipal budget), and demographic information of the
participant (age, job position, etc). All the following items were operationalized as 7-
point Likert scales. The antecendent factors efficiency goals, innovation goals, and IT
sophistication are understood as formative constructs that result from the sub-category
indicators described above (4.1.4). Based on the qualitative pre-study, each of these indi-
cators was operationalized on a scale from 1=no influence to 7=very high influence,
see Table4.2.
The main part of the questionnaire was a list of 60 possible mobile service offerings,
each with a brief description that needed to be rated according to their perceived attrac-
tiveness for the municipality on a scale from 1=not attractive to 7=very attractive.
These partly very innovative application scenarios were extracted from the academic
and practitioner literature and grouped according to the application clusters introduced
earlier (Section4.1.3). Table4.3(page51) shows a selection of these services.
The total perceived M-Government potential was also assessed based on the application
clusters presented Section4.1.3and operationalized, i.e. by 8 items on a scale from
1=no potential benefit to 7=very high potential benefits for the municipality. We

48
 4.1 Innovations in Mobile Government

opted for replicating this structure to ensure that the respondents understanding of M-
Government was congruent with the groups of concrete rated mobile services, and thus
supported content validity of both constructs. The planned investment in M- Government
services was assessed directly by asking for the estimated total investment of the
municipality in mobile services within the next three years as a single 7-point item with
the interval limits <50; 100; 250; 500; 1,000; 5,000 and >5,000 thousand Euros.
The content validity of the questionnaire was checked carefully. The questionnaire
items were first revised by several colleagues and experts in measurement theory and
statistics. Following the method proposed byHunt et al.(1982), the initial version of
the questionnaire was then pretested in meetings with some of the interview partners
involved in the pre-study, which lead only to minor changes in the formulations of the
service descriptions and antecedent factors. The full online questionnaire is presented in
Appendix1.

Sample

The actual survey took place between May and June 2010 and was conducted as an
open online survey. From the list of participants of one of the largest E-Government
conferences, we extracted the electronic addresses of the mayors and IT executives of the
187 German municipalities with more than 50,000 inhabitants and completed these by
an Internet search when necessary. Since we could not assume that the IT executives
are always the right contact person for the topic of M-Government, we sent an initial
invitation to the mayors offices with a request to forward our message to an appropriate
contact person. As the only incentive to participate, we offered the participants to
provide them with the results of the study (which we did afterward). A few days after
the initial invitation, separate reminder emails were sent to the mayors and the municipal
IT executives, excluding those who had already answered. For large municipalities (i.e.,
those with >100,000 inhabitants) we also made reminder calls via phone. Furthermore,
we answered several, mainly technical, inquiries on the phone.
Of the 187 invited municipalities, 78 representatives had begun to fill in the question-
naire. Out of these, 28 incomplete records had to be discarded, leaving 50 valid responses
(response rate 27 percent) that entered the analysis. The majority of respondents (42)
stated to have IT-sided roles and 8 to work on departmental side of the municipality.
The distribution of city sizes and position of the respondents are presented in Figure4.2.
According to the criteria provided by (Kromrey2006), the data can be regarded as
representative sample for Germany.

Descriptive Results of Service Attractiveness

The descriptive analysis of the rated service attractiveness by mean (M ) and standard
deviations (SD) provides us with a detailed picture of the preferences of IT decision-
makers. While the M value can be interpreted as a compound rating of the attractiveness, the
SD reflects in how far the respondents agree or disagree. In the first application

49
4 IT Governance and Innovation Adoption in E-Government

120 Department
Head of director
100 Gemany total (n=187)
sub-department 3
Sample (n=50) 4
80 Other
Top level IT
4
head
60 IT staff
22
2
40

20 IT group leader
11 IT department
0 head
50-100,000 100-200,000 >200,000 4

Figure 4.2:Sample description (city inhabitants and respondent position)

cluster (public administration), mobile work management, i.e. the use of mobile devices
for data collection and disruption-free processing, for example, in the offices for food
and veterinary inspections, has been evaluated with M =4.86 as the most attractive
application scenario.
By far, the highest attractiveness of all applications is in mobile firefighter support
systems. Mobile access to information such as building plans, maps, event and object
information, as it is already implemented in some municipalities, is apparently considered by
all municipal officials as an extremely useful application. The second-highest rating is
given to digital authentication, which is currently expectedin combination with the
electronic ID cardto allow for secure identification via telephone and Internet and thus
to enable a number of new transactional E-Government services for municipalities. By
far, the highest variance (SD) can be noted for municipal wireless networks. This possibly
reflects the still different opinions regarding the economic value of publicly subsidized
wireless networks (which potentially compete to commercial broadband offerings, see also
Winkler et al.2009for a broader discussion).
Table4.3shows the rated service attractiveness, sorted in descending order by mean
values (n=50). For brevity, here we only exhibit the top three services plus the the
ones with the lowest mean attractiveness scores per application cluster. The complete
list of services descriptions and attractivenesses is presented in (Ernst2010, available on
request).

4.1.6 Model Tests and Results Discussion

Methodology

We estimate and test the postulated model using the variance-based partial least squares
path analysis (PLS). All calculations were performed using the software products SPSS
and SmartPLS (Ringle et al.2005). The PLS approach is particularly suitable in the
present case for several reasons. First, the given question can be regarded as rather

50
 4.1 Innovations in Mobile Government

Table 4.3:Descriptive results of service attractiveness (excerpt)

M-Government service M SD M-Government service M SD
Public administration Transportation and traffic
Mobile work management 4.86 1.69 Mob. payment in public transport 4.48 1.53
Mobile payment at municipality 4.50 1.68 Car parking guidance service 4.42 1.75
City information services 4.48 1.59 Intelligent transport system 4.25 1.59
Mobile voting 3.04 1.70 Automatic city-toll 3.10 1.83
Public safety Environment and Infrastructure
Mobile firefigther support 5.57 1.53 Intelligent building management 4.80 1.58
Digital authentication 5.14 1.71 Intelligent street lighting 4.62 1.51
Mobile police support 4.79 1.69 Air pollution information 4.59 1.52
Person tracking 3.32 1.64 Intelligent garbage disposal 3.48 1.60
Education Tourism and Culture
Electronic library card 5.04 1.69 Mobile tourist guide 4.64 1.65
Mobile library 4.70 1.40 Mobile ticket booking 4.35 1.66
Electronic student card 3.91 1.74 Information kiosks 4.00 1.62
Educational information system 3.06 1.63 Mobile TV 2.88 1.73
Health Private households
Barrier-free accessibility 4.16 1.76 Municipal wireless network 4.74 2.02
Medical information service 3.69 1.84 Mobile services for the elderly 3.81 1.68
Mobile telemedicine services 3.69 1.87 Networked home environment 3.58 1.67
Dependents information service 3.21 1.72 Pet tracking service 2.85 1.59

exploratory since we can hardly build on existing theory. In contrast to covariance-based

approaches, PLS poses no requirements for the distribution of the underlying variables
and is more suitable when the focus is on theory development (Chin1998b).
Second, the research model contains formative constructs, which are more difficult to
model in classical covariance-based approaches (Panten and Boow-Thies2007). Hence,
the variables efficiency goals, innovation goals and IT sophistication are constituted (i.e.,
formed) by the indicators operationalized in our pre-study (Section4.1.4). Especially in
research on success and influence factors, formative constructs are often better suited to
represent the causal effect between indicators and the construct (Albers and Hildebrandt
2006). Consequently, the antecedent factors represent a weighted index of their substan-
tial indicators. The weights resulting from this correspond to the beta coefficients in
a standard regression model and typically have smaller absolute values than reflective
indicator loads.2
The required sample size for testing of PLS models is not without controversy in the
recent [Link](2010) argues that PLS can deliver meaningful results even at
a sample of 20 cases. On the other hand, (Marcoulides2009) indicate that samples of
2
It should be noted that for robustness, we also tested the model in two additional variants, with all
constructs modeled reflectively as well as formatively. In neither case did major differences arise in
the interpretation of the path coefficients and statistical significances, which is consistent with the
observations made by (Albers and Hildebrandt2006).

51
4 IT Governance and Innovation Adoption in E-Government

this size are not suitable to reliably detect weak path coefficients. Depending on the
degrees of freedom of the model, the heuristic ofChin(1998b) has been established
stating that the sample should be at least 10 times as large as the largest number of
formative indicators of a latent variable, or as the largest number of predictors of latent
endogenous variables. Both numbers are equal to 3 in the present model, so that with
n=50 we fulfill this heuristic. We follow the approach byChin(1998b) and assess the
measurement model first before we test our model hypotheses.

Measurement Model

To assess the measurement model validity, formative and reflective constructs are to be
considered separately.

Formative Constructs The formative variables efficiency goals, innovation goals and
IT sophistication must be tested for multicollinearity (Panten and Boow-Thies2007).
For this purpose, we calculated a Pearson correlation matrix as well as the tolerance
values from reciprocal regressions for each triple of construct indicators. Despite some
significant correlations of up to r=0.6, all tolerance values are well above the threshold
of 0.1. This indicates that multicollinearity can be ruled out as a measurement issue for
all nine indicators so that they do not need to be merged in more aggregated indices, see
Table4.4. Since formative indicators are included in the model indicators because of
their content-related relevance, no further assessment of convergent and discriminant
validity is required (Panten and Boow-Thies2007).

Table 4.4:Tolerance values of formative indicators

Efficiency goals Innovation goals IT sophistication
EF1 0.619 IN1 0.624 IT1 0.503
EF2 0.634 IN2 0.662 IT2 0.390
EF3 0.845 IN3 0.484 IT3 0.391

Reflective Constructs The evaluation of the reflective variables follows according to the
logic ofHomburg and Giering(1996) in three steps. In the first step, exploratory factor
analyzes (EFAs) are conducted with the indicators associated to each of the variables in
order to potentially purify the measures and assure unidimensionality. Each indicator
should exhibit a loading greater than 0.4 and no load on a second factor, and each
factor should explain at least 50 percent of the variance of its associated indicators to
be included in the analysis (Homburg and Giering1996). Then, in the subsequent steps,
convergent and discriminant validity can be assessed.
The variable perceived M-Government potential was operationalized by eight indica-
tors, one for each application cluster. The EFA of these eight indicators shows that
the perceived potential for tourism and culture (0.78) and for transportation and traffic

52
 4.1 Innovations in Mobile Government

(0.51) clearly load on a second factor. Apparently, municipal IT decision-makers have a

diverging opinion to these applications, whicharguablylie outside the core responsi-
bility of the municipal administration. This is, in both of these application clusters the
respondents had to rate application scenarios that are much more oriented towards the
private sector (e.g., mobile transport ticketing, car sharing, as well as mobile tourist
guides, mobile TV, etc.). In line with (Homburg and Giering1996), we conclude that
these two indicators are not suitable for measuring the attitudes toward the potential
attractiveness of mobile government as a whole and thus remove them from the analysis.
The remaining one-dimensional factor explains, on average, 64 percent of the variance in
the remaining six indicators.
The variable for measuring perceived service attractiveness is derived from list of rated
services. For this purpose, we first created eight indices by averaging the indicators
of the perceived attractiveness per application cluster. To ensure congruence with the
(now purified) variable of perceived M-Government potential, we removed the indices for
tourism and culture as well as transportation and traffic analogously. The EFA of this
construct produces a single factor that explains, in average, 69 percent of the variance
the six remaining indices.
Convergent validity of the obtained constructs is first assessed by checking the internal
consistency and Cronbachs alpha. The values for both constructs are well above the
required threshold of 0.7. Since alpha still depends on the number of indicators (here
six), we consult the composite reliability (CR) as a further criterion, which measures
how well the constructs are represented by the associated indicators. Values for both
constructs are well above the threshold of 0.6 (Panten and Boow-Thies2007), which
supports convergent validity of our measurement model, see Table4.5.

Table 4.5:Convergent validity criteria

Construct AVE Alpha CR
Perceived M-Government potential 0.626 0.899 0.921
Perceived service attractiveness 0.709 0.930 0.944

Discriminant validity refers to the extent to which the indicators of different latent
constructs are separable. The Fornell-Larcker criterion demands that the latent variable
correlations with other constructs should be less than the root of the substantively aver-
age variance extracted (AVE) (Fornell and Larcker1981). As we can see in Table4.6, this
criterion is fulfilled for all latent variables.3 The second criterion of discriminant validity is
that the factor loadings of the indicators by their substantial constructs should be
higher than the cross-loadings from other constructs, which is also fulfilled.

Common Method Bias and Demographic Distortions In order to assess whether a

majority of the observed variance results from the measurement method (i.e., common
3
For formative constructs, this criterion is not applicable and has only been demonstrated to demon-
strate robustness in the reflective case.

53
4 IT Governance and Innovation Adoption in E-Government

Table 4.6:Discriminant validity criteria (reflective case, root AVE on diagonal)

EF IN IT MP SA
Efficiency goals (EF) 0.790
Innovation goals (IN) 0.390 0.826
IT sophistication (IT) 0.382 0.264 0.888
M-Government potential (MP) 0.466 0.512 0.532 0.791
Service attractiveness (SA) 0.428 0.355 0.412 0.812 0.842

method bias), we performed a one-factor test according toHarman(1976). The test, i.e.
an EFA with all model variables, resulted in five distinguishable factors that reflect the
five factors of our model, whereas the first factor explains only 31 percent, rather than
the majority, of the total variance in the model. This suggests that the presence of a
common method bias can not be the main reason for the correlations in the measurement
model (Podsakoff and Organ1986).
To assess whether a distortion of the variables results from demographic characteristics
in our sample, we performed pairwise Spearman rank correlation tests of the variables of
the measurement model and the number of the inhabitants, as well as other characteris-
tics of the respondents. Contrary to the observations made by (Moon and Norris2005),
we find no significant correlation between the size of the municipality and the measured
variables. The same applies to the respondent demographics.

Structural Model Assessment

The results of the PLS analysis are presented in Figure4.3. Statistical significance was
assessed by t-tests based on a bootstrap procedure with 5,000 resamples. To evaluate
the results, the explained variances R2 and path coefficients can be interpreted simi-
lar to those in the simple regression (Panten and Boow-Thies2007). We begin our
interpretation with the hypothesized antecedent factors.
The results support the hypotheses that innovation goals (H4) and IT sophistication
(H5) are positively related with the perceived M-Government potential. Based on the
available data, the impact of IT sophistication can be viewed only as slightly stronger.
This finding provides a strong indication that those municipalities are primarily mo-
tivated for M-Government which also have a a more general ambition to implement
innovations in E-Government. M-Government therefore can not be viewed as an isolated
part of the municipal technology strategy, but rather adds up to other E-Government
activities. In this regard, the incentive to offer new services and respond to changing
customer needs (i.e., innovation goals) and past experience in E-Government (IT sophis-
tication) play an almost equally important role.
However, albeit a path coefficient of 0.19, we find no significant support for influence of
efficiency goals (H3, t=1.56). According to our nomological framework, this means that,
although many M-Government applications obviously aim at process support, efficiency
improvements and the bilateral benefits for citizens and administrations, this potential

54
 4.1 Innovations in Mobile Government

Process improvement 0.35ns

0.59ns
Win-win situation Efficiency goals

Efficiency 0.19 ns (H3)

improvement 0.30ns

External demand 0.04ns

0.34** (H4) Perceived 0.81*** (H1) Perceived service

0.35ns
Service enhancements Innovation goals potential attractiveness
R2=0.46 R2=0.66
Mobile
consumerization 0.78*

E-government 0.33ns 0.37** (H5)

platform
0.80**
Service strategy IT sophistication

IT strategy 0.48ns *p<0.10; **p<0.05; ***p<0.01; ns=not significant; n=50

Figure 4.3:Structural model results

is not (or at least not strongly) associated with the general perceived potential of M-
Government. In other words, efficiency goals seem to be not (yet) a strong driver for
positively evaluating the general M-Government potential and thus neither for planning
concrete service implementations.
Altogether the antecedent factors explain R2=46 percent of the variance in perceived M-
Government potential, which can be interpreted as a good level of determination given
the exploratory direction of this study (Homburg and Giering1996). The perceived M-
Government potential mediates the influence of the antecedent factors on the perceived
service attractiveness. With a path coefficient of 0.81, it is strongly and significantly
associated with the perceived service attractiveness (H1). This strong coefficient appears
remarkable, given that these two constructs originally point at two different conceptual
and nomological levels (i.e., one at the general perception of M-Government as a phe-
nomenon, and the other at an aggregate measure of the attractiveness of a number
of concrete application scenarios). This, and the R2=66 percent of variance explained
in perceived service attractiveness, indicate that perceived M-Government potential is
a reliable predictor for the concrete evaluation of a broad spectrum of M-Government
services. This also underlines the high level of content congruence and statistical validity
of these two novel constructs.

Cluster Analysis

The different influences exerted by the antecedent factorswhich are only partly consis-
tent with the results of our qualitative pre-studymay indicate heterogeneous motiva-
tional structures of the municipalities in our sample. As some authors note, existing
heterogeneity in a sample may be obscured in the aggregation of an unidirectional

55
4 IT Governance and Innovation Adoption in E-Government

regression model (Backhaus et al.2003, p. 511). Such information may be recovered

by an a-posteriori segmentation of the sample into smaller groups. We perform such
analysis by a Wards hierarchical clustering based on the linkage of the (standardized)
factor scores of the antecedent variables efficiency goals, innovation goals and IT sophis-
tication. Examining the gradient of the distance measured over the clustering steps, we
note a strong slope change of the cluster distances when reducing from four to three
clusters. Hence, we set the optimal number of clusters to four. The resulting clusters
with different sizes may be labeled according to their characteristic mean factor scores as:
Innovators (nA=8), IT experienced (nB =6), Efficiency-oriented (nC =23), and Laggard
municipalities (nD=13), as depicted in Figure4.4.
Innovators IT experienced Efficiency-oriented Laggards
(nA=8) (nB=6) (nC=23) (nD=13)
300k
1,5 EUR

100k
0,5
EUR

-0,5

-1,5

Efficiency goals Innovation goals IT sophistication Service attractiveness Planned investment (1200k)

Figure 4.4:Mean factor scores per municipality cluster

We may call such municipalities Innovators that see themselves particularly sensitive to
the expectations of their customers and hence, besides striving for operational efficiency,
draw new ideas for M-Government services and applications from the goal to innovate.
Apparently, this may not necessarily be the municipalities with the highest level of IT
sophistication. In contrast, IT experienced 4 municipalities seem to have a good basis for
innovative IT projects, yet no urgent efficiency goals. These municipalities potentially
still perceive a lack of attractive services offerings (see also next subsection). The largest
segment can be classifiedprobably the light of their budgetary situationsas efficiency-
oriented municipalities, which must limit their IT expenditures to the bare essentials
and thus are motivated rather moderately for M-Government endeavors. The cluster of
municipalities that exhibit a below-average level of all three drivers can be characterized
as Laggards. These communities are evidently behind the other three groups in the
process of innovation diffusion and M-Government adoption.

Relationship between Goals, Perceived service attractiveness and Investment

Behavior

To test the postulated relationship between the motivational structure of municipalities and
the perceived service attractiveness (H1) on the one hand, and the planned invest-
4
Note: This cluster has been termed IT experienced (instead of IT sophisticated) for better coherence
with the original version of this paper, which has been published in German.

56
 4.1 Innovations in Mobile Government

ment in M-Government (H2) on the other hand, we perform an analysis of variance and
test post-hoc whether the identified clusters differ significantly from each other regarding
the factor means (M ) of the target variables (see Table4.7).
In terms of perceived service attractiveness, it shows that the null hypothesis of equal
cluster group means can be refuted (F =3.73; p=0.02). To determine between which
clusters these differences exist, we perform post-hoc tests using to Fisher LSD (least
significant difference) tests. At a p<0.1 confidence level, we find that innovators dif-
fer significantly in their perceived attractiveness (MA=0.77) from efficiency-oriented
(MC =0.10; pAC =0.08) and laggard municipalities (MD=-0.61; pAD=0.02). Efficiency-
oriented and Laggards also differ significantly from each other (pCD=0.04). IT expe-
rienced municipalities exhibit a slightly below-average perceived service attractiveness
(MB =-0.08) and no significant group mean differences from any of the other clusters,
which indicates a high variance within this cluster.
The analysis of variance also indicates unequal group means across clusters in terms
of the planned investment (F =2.54; p=0.07). LSD post-hoc tests show that innovators
(MA=175,181 EUR) differ significantly from Laggards (MD=78,680 EUR; pAB =0.06)
and from Efficiency-oriented (MC =71,543 EUR; pAC =0.03). The latter also fall signifi-
cantly below the level of investment of the IT-experienced municipalities (MB =158,333
EUR; pBC =0.09).

Table 4.7:Municipality clusters and group differences

Cluster variables ServAttract Investment
Cluster n EF IN IT M LSDa M (EUR) LSDa
Innovators (A) 8 0.65 1.33 -0.09 0.77 C, D 175,181 C, D
IT experienced (B) 6 -0.71 0.75 1.27 -0.08 158,333 C
Efficiency-oriented (C) 23 0.62 -0.19 0.26 0.10 A, D 71,543 A, B
Laggards (D) 13 -1.17 -0.83 -0.99 -0.61 A, C 78,680 A
a
Group differences assessed by Fisher least significant difference post-hoc tests, p<0.10

4.1.7 Conclusion
Based on the attitudes of municipal IT decision-makers, this study indicates that there is
a link between organizational goals of a municipality and the perceived attractiveness of
novel M-Government services as well as the municipalitys investment intentions. Repre-
senting a sum of opinions, we argued theoretically that the extent to which such strategic
goals are pursued by the IT decision-maker lead to varying degrees of motivation for or
against introducing M-Government applications. Based on the literature and a series
of interviews, we typologized and operationalized such strategic conditions in a triad
of efficiency goals, innovation goals, and IT sophistication. The results of an empirical
study and PLS path analysis support a significant influence of the latter two factors,
suggesting that municipalities possess heterogeneous profiles regarding their motivational
structures. This view is supported by a post-hoc cluster analysis, in which we discovered

57
4 IT Governance and Innovation Adoption in E-Government

four types of motivation profiles: Innovators, IT experienced, Efficiency-oriented and

Laggards. In regard to the dependent variables, we were able to show that, within the
next three years, Innovators and IT experienced plan to invest about the double in M-
Government services compared to Efficiency-oriented and Laggard municipalities. Also
it seems that particularly IT experienced municipalitiesbased on the assessment of
the 60 potential services and applications in our studystill perceive a relative lack of
attractive M-Government service offerings.

Implications for Practice

The findings in this study not only provide relevant implications for the public admin-
istrations, but also for service providers and the political level. First, we demonstrate
thatsimilar to private sector organizations (Rogers1962)cities possess heterogeneous
innovation profiles. Unlike private businesses, however, municipalities are facing far less
competition. Therefore it appears reasonable that the experiences made by Innovator and IT
experienced municipalities serve as an example for the large group of non-innovators and
that a knowledge transfer takes place. Therefore, municipal IT decision makers can use
the proposed typology and clusters to determine the role and position of their own
municipality within the (M-Government) innovation adoption process. Although our
subject of study focused on innovations particularly in M-Government, we put for- ward
that the presented model of antecedents factors and investment outcomes is also
applicable to other technological and organizational innovations in public administration.
Furthermore, our work provides a comprehensive overview about the service offerings
that are most attractive for investment from the perspective of municipal IT decision
makers. This not only provides guidance for political stakeholders, but also serves as
a valuable market information for private service providers that push into this young
segment of the E-Government market. A point we view critically, however, is the
relatively large cluster of Laggard municipalities with their below-average motivational
profile for all three dimensions, along with the finding that efficiency goals (so far) are
not significantly related to M-Government adoption. This suggests that M-Government
is generally not (yet) viewed as a route to greater efficiency in public administration, but
rather as a luxury for those customer-oriented governments which can still afford to
realize innovative E-Government projects. This in turn raises the question whether the
political level should enforce certain mobile service offerings (e.g., those rated with a
high attractiveness in our survey) and/or according standards to a greater extent, as it
has been done in the past, for example, in the implementation of the EU Services
Directive and the electronic identity card.

Limitations and Future Work

A few limitations of this study merit consideration when interpreting the findings. First,
we tried to counteract the inherent limitation of generalizability by addressing a repre-
sentative sample of cities across Germany. A second limitation is given by thedespite

58
 4.1 Innovations in Mobile Government

the good response rate of 27 percentrelatively small sample size. As discussed in

Section4.1.6, a greater sample size may have led to better statistical significance of the
endogenous model factors (in particular efficiency goals).
Further, it was assumed that the evaluation by an individual IT decision maker suffi-
ciently reflects the sum of the opinions of relevant stakeholders in a municipality. This
may have led to measurement errors especially for large municipalities. In addition, our
theoretical framework and the presented research model unavoidably represent a gross
simplification of reality. In order to address the latter two limitations, we asked the
respondents at the end of the survey for their willingness to participate in a secondary,
more qualitatively-oriented study. In this way the authors hope to identify further
influencing factors for M-Government on a more fine-grained level and contribute these
to the body of knowledge in a future research.
Despite these limitations, this work provides a relevant contribution by being one
of the first empirical studies that directs the focus of M-Government adoption on the
relevant decision makers in municipal administrations. For this purpose, we used a com-
prehensive conceptualization and operationalization of M-Government which explicitly
does not neglect the internal, efficiency-oriented perspective on M-Government benefits.
Altogether, this work complements and enhances the common end-user level M- and E-
Government research streams by taking an organization-oriented approach to study this
highly relevant phenomenon.

4.1.8 Summary
Municipalities face increasing fiscal stress and the pressure to save on costs. At the same
time, with the growing popularity of the mobile Internet, new innovative application
scenarios of Mobile Government (M-Government) are emerging. In this chapter we an-
alyzed a broad spectrum of M-Government services from the perspective of municipal IT
decision-makers and investigated the organizational factors that influence the perceived
attractiveness of these services. Based on the extant literature and a series of interviews,
we developed a model that proposes three main motivational dimensions (efficiency goals,
innovation goals, and IT sophistication) to affect municipal mobile government adoption.
We tested the model empirically in a survey with 50 German municipalities by using
a partial least squares path modeling technique and clustered the cases according to
the proposed dimensions. The results indicate that municipalities with a high degree
of perceived M-Government service attractiveness differ significantly from those with a
low degree. Based on these findings, we argue that municipalities can be characterized
either as Innovator, IT experienced, Efficiency-oriented, or Laggard types regarding their
motivational structures.

59
4 IT Governance and Innovation Adoption in E-Government

4.2 Towards Transformational IT Governance

4.2.1 Preamble

This chapter has been published earlier as a conference paper at the 19th European
Conference on Information Systems (ECIS), June 2011 in Helsinki (seeWinkler, Lvova,
and Gnther2011b) and as an extended version in (Winkler2012)

4.2.2 Introduction

With the rise of the mobile internet, smart internet connected devices are becoming part
of our everyday life. The fact that in many countries mobile broadband penetration has
already surpassed fixed broadband subscriptions may only serve as a rough indicator
(ITU2010). Just as for private businesses (e.g.,Siau et al.2001), these technological
advances also offer the possibility for public sector institutions to rethink the interaction
channels with customers as well as with employees and enhance current e-government
practices by appropriate mobile government (m-government) services (Kushchu and
Kuscu2003;Trimi and Sheng2008).
However, such enhancements may imply transformational steps and require fundamen-
tal changes to internal organization and governance (Weerakkody et al.2007;Janssen
and Shu2008) so that, analogously to private economy, public sector institutions take
a different pace in innovation diffusion (Winkler and Ernst2011). Drawing on findings
from a recent survey, our research question is: What exactly makes public sector or-
ganizations differ in their adoption of mobile government services? Present literature
tends to focus rather on technocratic aspects of m-government adoption, often lacks in
empirical foundation and is hardly able to embrace the relevant contextual differences
between the objects of analysis (e.g.,Kushchu and Kuscu2003;Sandy and McMillan
2005;Al-khamayseh et al.2006;Kushchu et al.2007;Kumar and Sinha2007;El-Kiki
and Lawrence2007). We combine grounded theory and quantitative content analysis to
develop a comprehensive framework on m-government adoption and apply this framework
in four cases to determine the relevant contingencies in this context.
This chapter contributes to current research inasmuch as a) we empirically derive
factors which influence m-government adoption and b) generate appropriate theory to
explain differences in the adoption between different entities. Since mobile government
initiatives primarily occur at the local level (Borucki et al.2005), we chose the municipal
decision makers as the object of analysis. In the remainder of this chapter we first
review related work. Then we explain the methodological approach in Section4.2.4.
Section4.2.5presents the derived adoption framework which is then used for the com-
parative case studies in Section4.2.6. Section4.2.7summarizes the results and proposes
future work.

60
 4.2 Towards Transformational IT Governance

4.2.3 Related Work

Mobile Government Adoption

Following the definition byKushchu and Kuscu(2003), m-government can be defined

as a strategy and its implementation involving the utilization of all kinds of wireless
and mobile technology, services, applications and devices for improving benefits to the
parties involved in e-government. Akin to e-government, m-government may have dif-
ferent foci. For the purpose of this work we distinguish between three main interaction
patterns: government-to-citizen (G2C), government-to-business (G2B), and government-
to-government (G2G) (Trimi and Sheng2008). Although rarely causing structural
changes (Borucki et al.2005), the use of mobile technology may still help to transform
governments (Kumar and Sinha2007). In this sense, m-government may be related to the
upcoming research stream of transformational government (Irani et al.2008;Weerakkody
et al.2008).
Challenges and success factors of m-government adoption are widely discussed in e-
government research (seeNapoleon and Bhuiyan2010for an overview). As the most crit-
ical issues, privacy and security as well as accessibility concerns are frequently mentioned
(Kushchu and Kuscu2003;Sandy and McMillan2005;Al-khamayseh et al.2006;El-Kiki
and Lawrence2007;Kushchu et al.2007;Kumar and Sinha2007). Moreover, we find
a long list of technical issues related to infrastructure development, payment infrastruc-
tures, and compatibility as well as legal issues (Kushchu and Kuscu2003;Kushchu et al.
2007), and user-related issues (such as preferences, quality, user friendliness, convenience,
acceptance and education) (Sandy and McMillan2005;Al-khamayseh et al.2006;El-Kiki
and Lawrence2007) as well as cost (Sandy and McMillan2005) as critical factors for
mobile government adoption.
However, the rather young literature on m-government mostly draws upon conceptual
work and single case studies and still lacks empirical foundation (Napoleon and Bhuiyan
2010). Thus, adoption factors identified in existing research can provide general norma-
tive guidelines, but do not to account for the individual differences between organizations.
A recent study byWinkler and Ernst(2011) indicates that public sector institutions differ
in innovation adoption. They segment municipalities into four empirical clusters: the
Innovators, Hybrids, Efficiency-oriented and Laggards. Out of these clusters, only the
former two may be considered as early adopters. In this chapter, we build upon this
classification to explain the differences between such public sector organizations.

IT Governance in the Public Sector

One major barrier for m-government adoption also stated by experts refers to governance
during initiation of an m-government project (El-Kiki and Lawrence2007). Governance
has equally been mentioned as one of the core elements in government transformation
(Janssen and Shu2008). From an organizational standpoint, any public institution can be
horizontally divided in departmental areas and central units that perform cross-functional

61
4 IT Governance and Innovation Adoption in E-Government

35 8
Department
30 Survey (n=50) Age
6 Head of director
25 Interviews (n=12) Top level
sub-department
20 IT head
4
15
10 Other
2
5
0 0 IT department
50-100.000 100-200.000 >200.000 30-39 49-49 50-59 60-69 IT group leader head

Figure 4.5:Interview sample description

tasks such as information systems management (Brown1999). A core aspect of IT Gov-

ernance is to define the distribution of decision-making authority and responsibilities
between such entities. Patterns for the distribution of decision-making authority can be
roughly classified into centralized, decentralized and federal archetypes (Sambamurthy
and Zmud1999;Weill and Ross2004a).
Governance mechanisms, which can be implemented at structural, procedural and
relational level, have been identified as a key to align IT with business organizations
and achieve IT performance in private sector companies (e.g.,Weill and Ross2004a). Yet,
little research has been conducted on mechanisms that contribute to IT governance within
public sector organizations (Ali and Green2007). For instance, in a cross-industry survey
reported byWeill(2007) public sector organizations clearly score the lowest IT governance
index. This topic is recently also gaining practical relevance (e.g.,Hoch and Payn2008), as
the fundamental differences between public and private sector may also call for different
principles of IT governance (Sethibe et al.2007).

4.2.4 Research Method

Data Acquisition

We acquire data from 12 in-depth semi-structured interviews with municipal IT decision

makers. This object of analysis appears particularly suitable, since most m-government
initiatives occur at the local level by the involvement of very few departments and
workers, asBorucki et al.(2005) note. Interview contacts have been drawn from a
prior survey, which features a representative sample of German municipalities (n=50)
Winkler and Ernst(2011). Interviews took place between June and August 2010 and
followed a common guideline (a. context, b. mobile applications c. drivers and inhibitors,
d. stakeholders). Average duration of an interview was 54 minutes amounting to 10.8
hours total interview data, respectively 140 pages of transcription. The city size, demo-
graphics (100% male) and job positions of the interviewees are depicted in Figure4.5.

62
 4.2 Towards Transformational IT Governance

Coding Procedure and Content Analysis

We combine grounded theory with quantitative content in a five-step approach to analyze

the interview material. Grounded theory is contrary to other research methods as it seeks
to systematically develop theory, rather than verifying or testing it (Strauss and Corbin
1998;Glaser1992). Straussian grounded theory follows an interwoven process of open
coding, axial coding and selective coding. It is thought to be more prescriptive compared
to the Glaserian approach (van Niekerk and Roode2009). Our coding and calculations
were performed using a software for qualitative data analysis ([Link]) as well as a
common spreadsheet program.
1. We incrementally performed open coding of the given interview material by two
coders and retrieved an initial list of 173 codes. Average groundedness of a code, i.e.
number of quotations, is 8.3. Each code captures in average 28.3 words, which is a
calculatory 53% of the total interview material.
2. Deviating from the Straussian procedure, we first arranged these codes according to
basic categories (coding paradigm), before aggregating codes to mid-level concepts and
categories. As the core phenomenon (the adoption of m-government services) was clear
from the beginning of this study, we use an adapted Strauss coding paradigm (Strauss
and Corbin1998), which jointly interprets causal and intervening conditions as drivers
(i.e., conditions with a positive influence) and inhibitors (i.e., conditions with a negative
influence). The resulting coding paradigm is illustrated in Figure4.6.
3. We merged codes according to perceived semantic distance and partly sorted out
codes with very few quotations. Simultaneously, appropriate mid-level categories were
found within two of the basic categories to semantically structure these concepts. The
result is a condensed hierarchical framework with total 42 concepts in four categories
(see Figure4.7).
4. We determined the intercoder reliability (Neuendorf2002) of the condensed code set
by recoding a sample of three interviews with switched roles. We considered an overlap in
the quotation in both document versions as a hit, and calculated the reliability according
to the formula r = 2 hits / ( total codesCoder A + total codesCoder B ). The resulting
reliability of 58% can be regarded as a good result considering the number of different
codes. For the subsequent analysis we joined both codings to a consolidated version.
5. For a valuation of the conditional variables, both of the coders went through all
quotations belonging to the conceptualized conditions. We rated the relevant 1,273
quotations on a three-point scale, where -1 represents an inhibiting, 0 a neutral, and
+1 a driving influence on m-government adoption. The reliability of this rating was
=78% measured by Cohens Kappa, which can be regarded as a very good agreement
(Neuendorf2002, p. 141). We calculated the influence of each variable as the average
ratings of both coders. For the purpose of illustration, the resulting valuations have been
transformed to an equidistant five-step scale (interval boundaries at -1.0; -0.6; -0.2; 0.2; 0.6;
1) where represents a strong inhibitor, a weak inhibitor, a neutral
influence, + a weak driver and ++ a strong driver.

63
4 IT Governance and Innovation Adoption in E-Government

Context Causal and Intervening Strategies and Actions Consequences

Conditions
Specific set of Events and incidents that Phenomenon Actions and interactions Outcomes in
properties pertaining lead to the phenomenon Adoption of directed to managing response to the
to the municipality as well as facets of the m-government and responding to the phenomenon
broader structural context phenomenon
that drive or inhibit it.

Figure 4.6:Adapted coding paradigm fromStrauss and Corbin(1998)

Case Study Approach

To investigate the variance in m-government adoption more in detail, we selected four

cases from the sample. In qualitative research, the selected cases should be especially
critical or revelatory with regard to the phenomenon (Benbasat et al.1987;Merriam
1998). The replication logic either dictates to select cases which are expected to yield
similar results (literal replication), or opposite results (theoretical replication) (Yin2003,
pp. 46-53). In this research, we followed the latter approach and based our case selection
on three main considerations. First, we decided to choose one case from each of the
empirical clusters provided byWinkler and Ernst(2011), as these groups emerged from
statistic clustering which generally enhances variance. Second, within these clusters we
preferred such revelatory cases which are characterized by a good clarity, quality and
openness of the interviewee. Third, as selection should also be based on organizational
criteria (Benbasat et al.1987), we took into account municipality sizes and selected two
larger and two medium sized cities. This allows for both, pairwise comparison and testing
for size differences between these cases. Furthermore, we complemented the analysis
with relevant documentation, such as strategy documents, organization charts and press
clippings which we additionally retrieved from interview partners and web sources.

4.2.5 Framework for Mobile Government Adoption

The developed framework saturates the theoretical categories from Figure4.6with ap-
propriate concepts that play a role in m-government adoption. Figure4.7displays the
results of the qualitative analysis with total code frequencies and valuations in brackets.
The code frequency (#) can be interpreted as the relative importance of a concept,
while the valuation (+/) describes the strength of driving and inhibiting influences.
For reasons of brevity, we briefly describe the main concepts in the following.

Contextual Variables

Contextual variables, on the one hand, refer to properties of the municipality such as
size, economic profile and budgetary situation. On the other hand, we found properties
of the municipal IT organization and IT architecture to be relevant for m-government
adoption. Examples are horizontal distribution of the IT organization, the degree of

64
 4.2 Towards Transformational IT Governance

Context Drivers and Inhibitors Strategies Consequences

User level
Citizen benefits (40,++) Citizen involvement (38)
Increase city attractiveness (44,+) User orientation (35)
Business benefits and demand (23,+) Employee qualification (2)
Municipality Citizens demand (69,) Service outcomes
Size (9) Employee benefits (14,) Planned (29)
Economic profile (31) Employees technological attitude (12,--) Implemented (54)
Budgetary Failed (4)
situation (5) Economic level
Process improvements (49,++) Supra-regional funding (15)
Potential savings (13,++) Financial sustainability (8)
Cost-effectiveness (34,+)
IT organization Available funding (23,-) General outcomes
Horizontal Financial risks (8,--) Overall adoption
distribution (99) Adoption of status (18)
Governance Governmental level M-Government Perceived
mechanisms (23) Strategic intent (12,++) potential (16)
IT architecture (19) Political influence (46,) Inter-communal
Sourcing degree (18) IT capabilities (38,) cooperation (46)
Legal conditions (15,-) Internal marketing (10)
Ability to change (34,--) Regulatory changes (10)

Technological level
Technical feasibility (47,) Industrial and scientific
Security and privacy (30,) cooperation (47)
Market dependency (26,) Legacy coexistence (9)
Integration ability (27,--) Structural changes (5)

Figure 4.7:Framework for m-government adoption

external sourcing, as well as governance mechanisms which are used to integrate IT with
the functional departments. The complex influence of such variables should be elicited
by the case studies presented in Section4.2.6.

Drivers, Inhibitors and Strategies

Drivers and inhibitors are the factors that have either positive or negative influence on
the municipalitys decision on mobile government adoption. Strategies are the actions
that are taken in response to drivers and inhibitors. In the interviews, these actions have
been identified by key words such as need/should/must be done. To group drivers,
inhibitors and strategies, four self-evident levels emerged as categories: user, economic,
governmental and technological level.

User Level On user level we mainly find drivers which relate to the expected benefit of
m-government for the three target groups (citizens, businesses and governmental
employees). Some cities also aim to increase their attractiveness by offering mobile
services, while others clearly doubt such kind of impact. The highest group-specific
benefit is perceived for citizens who will be able to use a new channel of communication
and interaction with their municipality. However, some interviewees also note that there
is a considerable lack of knowledge on what the citizen demand. A similar pattern can be
observed regarding government employees. Some employee groups see a clear benefit in
the use mobile technologies in their daily work routines while others fail to adopt these
due to their technological attitude.

65
4 IT Governance and Innovation Adoption in E-Government

Consequently, the most important strategies to ensure a later user acceptance are to
involve citizens at an early state and to focus consequently on the target group during the
implementation of a new service (user orientation). Training the employees (employee
qualification) however, has only been mentioned few times. Regarding the business target
group, there seems to be a slightly different relationship. In case local businesses see a
clear benefit for a certain service, e.g. for mobile billing or parking, they will immediately
utter such demand, so that business benefits and demand represent both an (even though
weak) driver.

Economic Level Not surprising, all interviewees put large emphasis on economic as-
pects, so that we applied a semantically rather fine-grained code set at this level. In the
first place, possible process improvements, e.g. by increasing level of information and
throughput times, are mentioned to be a key driver for m-government. Such improve-
ments may further lead to improved cost-effectiveness and potential savings. However, as
some interviewees comment, not all process improvements pay off on the budget side, as
freed resources are often immediately consumed by other tasks or an increased demand
in the new service.
Surprisingly, availability of funding is only a light inhibitor for the municipalities. The
rationale behind this is that a good application will automatically pay off, so that costs
are to be seen as an investment rather than expenses. Therefore, financial risks play a
stronger yet not very prominent role as an inhibitor in such investments. Strategies to
ensure the economic success of such project are access to supra-regional funding (e.g., on
federal or EU level) and a focus on financial sustainability on the whole.

Government Level On governmental level we find strong drivers as well as strong

inhibitors. Few interviewees quote a certain strategic intent, i.e. a strategic vision and
creativity of the municipality in the use of new technology, as well as the capabilities of
their team (to a smaller extent) as important drivers. But rather than relying only on
their internal capabilities, most municipalities strongly refer to their network of inter-
communal cooperation to share experiences and drive diffusion. Such networks can be
formed by clusters of cities as well as cooperation of public service providers.
Political influence seems to be a double-edged sword. On the one hand, political
bodies need to be arduously convinced (internal marketing) to commit to technologically
enabled changes. On the other hand the interviewees report that larger transformations
can much easier e implemented once there is a political will. However, if such political
will is not present (which is mostly the case), the ability to change within the municipal
administration can become a major inhibitor. Furthermore, legal conditions concerning
data security, laws, as well as different legal standards between authorities and federal
states are an issue. This obstacle, in some cases had, in other cases still may to be
overcome by appropriate regulatory changes.

66
 4.2 Towards Transformational IT Governance

Table 4.8:M-government service outcomes

Citizens / businesses (G2C/G2B) Government (G2G)
Planned Mobile city portal (2) Traffic warden support (3)
Mobile library services (2) City council support system (2)
Mobile tourist guide (1) Civil engineering support (1)
Problem reporting service (1) Public order office support (1)
Mobile payment platform (1) Retirement homes control (1)
Environment agency support (1)
Implemented Mobile city portal (4) Traffic warden support (2)
Single service number5 (4) Food inspection support (2)
Mobile parking tickets (2) City council support system (1)
Public transportation Veterinary services support (1)
information (1) Firefighter support (1)
Public transportation payment (1) Mobile radar equipment (1)
Parking information service (1)
Appointment service (1)
Failed Live townhall meetings (1)
City Wi-Fi services (1)
Car sharing service (1)

Technological Level Regulatory changes are also related to technological influences,

especially security and privacy standards that municipalities need to comply with. How-
ever, stronger inhibitors on technological side refer to the feasibilityto implement new
technical solutions and dependency on the market. Deciding for a certain technology
may lead to a lock-in to certain vendors. The ability to integrate mobile applications
with existing procedures also fails due to a lack of open standards (e.g., web services) in
existing applications.
Several municipalities address feasibility issues through strategies of cooperation with
industrial partners or research institutions to stay informed about the state of the art. To
minimize technological risks, also a coexistence of the mobile process with the legacy system
is proposed. Yet, this strategy generates additional costs. Finally, some IT leaders
mention the need to structurally change the IT landscape in order to be able to integrate
with mobile channels.

Consequences

We divided the consequences of M-Government adoption into service outcomes and

general outcomes. Concerning the latter, most IT decision-makers generally see a high
potential in m-government possibilities and expect an increasing relevance of this topic in
the next two to three years. However, the perceived adoption status differs. While some
municipal representatives admit not having ventured any project with respect to mobile
government, others consider themselves to have a stronger expertise and experience.
We found these perceptions generally reflected in the number of service outcomes ob-
served which refer to the concrete mobile services which have been planned, implemented

67
4 IT Governance and Innovation Adoption in E-Government

or failed. Service outcomes are presented in Table4.8. Here, the numbers in brackets
correspond to the number of municipalities in the sample (n=12). We acknowledge
that these numbers serve as a rough indicator as interviewees might forget to mention a
certain service during an interview. We find that there are some common (planned or
implemented) applications across cities, e.g. the implementation of mobile city portals,
traffic warden support, single service number5 and mobile parking tickets. Other services,
by contrast, seem to be piloted largely by the initiative of individual cities.

4.2.6 Mobile Government Cases

To better understand the differences between municipalities we use the presented frame-
work to conduct multiple case studies following the methodology and selection criteria
outlined in Section4.2.4.

Case A: An M-Government Innovator

Case A is a city with more than 500,000 inhabitants in one of the economic centres in
Germany and considers itself to be a leader in e- and m-government. The IT department
is subordinate to the general office for central services and can be divided into two sub-
units: the department for IT operations and the office for organization and IT steering.
Further, in course of the EU services directive6, e-government has been given a high
priority through establishing an additional inter-disciplinary Competence Centre within
the Department for Economic Development. The interface to the departmental areas
is provided by a decentralized liaison role. Initiatives for new IT projects, such as
mobile services, may be filed by the departmental areas and evaluated by central IT
steering, which governs the IT budget. The city has outsourced data centre operations,
networks and communication infrastructure and large parts of application management
to a communal IT service provider.
The Head of Organization and IT steering sees the main benefit of mobile services
particularly on citizen and internal side. Mobile applications, such as an appointment
service or the ex-ante completion of required forms, may reduce citizens attendance
at public bodies, but also efforts for the municipality. This tendency is seen as an
important building block for increasing city attractiveness embedded in a comprehensive
strategy of urban development. Same applies to local businesses, where the facilitation
of routine interactions with the municipality may become an absolute location factor.
Economically, such applications should account for their costs, however, some applica-
tions which focus more on quality usually pay off only in the long run. Thus, almost
5
The single service number (D115) represents an exception inasmuch as it results from a Federal
initiative of the Ministry of the Interior and is gradually rolled out nationwide. Also, the classification
of this service as M-Government may be debatable.
6
The EU directive on services in the internal market (2006/123/EC) has been issued with the objective
of establishing a single market, and amongst other requires governments to provide a point of single
contact via electronic means.

68
 4.2 Towards Transformational IT Governance

everything which is technically feasible will be evaluated in the light of the IT strategy
and municipal goals. Possible internal resistance to changes can widely be mitigated
through the central empowerment of the office for organization and IT steering within
the municipality administration.
The only restrictions from this perspective are technical possibilities and certain le- gal
regulations, such as data protections laws. On the technical side, the interviewee states
that most internal procedures and applicationscontrary to a service-oriented
paradigmyet do not offer the possibility to be opened towards the (mobile) Internet.
Thus, on both issues more nation-wide standards, as occurred for example with the
introduction of the electronic ID7, would be appreciated to avoid redundancy. The mu-
nicipality actively cooperates with industry, such as a local telecommunication company,
as well as with local research institutions. As an outcome, the city offers and operates a
number of external and internal mobile services, which enjoy a high acceptance among
their users, see Table4.9.

Case B: An IT Experienced Municipality

As case B, we chose a medium-sized city in Southern Germany with a structurally

challenging but relatively dynamic economic environment. Our interviewee is the head
of an e-government staff function, which has been installed to coordinate between the
municipal demands, internal IT and a communal IT services provider. Like in case A,
the link to the departmental areas is provided by IT liaison roles that manage day-to-day
operations. On the strategic level, there is an IT steering committee meeting every 4-6
weeks and comprising members from the departmental areas, staff council, auditing and
internal IT. This committee decides on new and ongoing IT projects and only needs the
city councils approval beyond a certain investment volume.
The driver for m-government is mainly seen in the benefit for local industries. These
industries deservedly claim appropriate interfaces to simplify their own processes, e.g. for
switching from a single to a collective billing for parking services. On the contrary, the
interaction with citizens is characterized by consumption and creation of a benefit that is
hardly quantifiable. Moreover, the interviewee reports on difficulties in recognizing the
citizen demand, as the demands of single interest groups cannot necessarily be understood
as a collective need. On the economic level, the importance of a payment platform is
stressed, hence only such services will be implemented which directly pay off, e.g. an on-
demand provision of geo-data. However, implementing projects just for the sake of the
citys image was something for the 90s, rather the city puts emphasis on the
sustainability of the solution and pursues long-term partnerships with various industry
partners.
On the governmental level we find interplay of sound internal IT capabilities and
7
Germany has launched a new electronic ID from November 2010 onwards, driven by an initiative of
the German Ministry of Interior, the Federal Office for Information Security and multiple research
and industry partners.

69
4 IT Governance and Innovation Adoption in E-Government

Table 4.9:M-government cases overview

Case A Case B Case C Case D
Innovator IT experienced Efficiency-orientd. Laggard
Municipality
Sizea Very large Medium Medium Very large
Budget 2,500 mn Eur; 500 mn Eur; 300 mn Eur; 2,500 mn Eur;
and empl.b 9,000 empl. 4,000 empl. 1,000 empl. 9,000 empl.
Financial Moderate Moderatepoor Poor Moderate
situation
IT organization
Budget 20 mn Eur; 6 mn Eur; 6 mn Eur; 10 mn Eur;
and empl. 300 empl. 60 empl. 8 empl. 250 empl.
Hor. distr. Centralized Centralized Centralized Decentralized
Structural Major decision rights IT Steering Meeting Idea Mgmt process, IT Steering Meeting,
mechs. on IT Steering side IT Steering decisions IT Planning Group
Sourcing High Medium Very high Very Low
User ++ + +
level + Citizen demand + Business benefits + Business demand + Ideas from
+ Business demand Lack of Citizen demand departments
+ City knowledge on Citizen demand
attractiveness citizen demand
Economic + ++ ++
level + Effort reduction + Cost-effectiveness + Process + Goal to save costs
+ Process + Co-financing from improvements Ability to save
improvements supra-regional + Cost savings costs
+ Long-term level Financial risks
benefits
Govern- + +
mental + IT capabilities + IT capabilities Political influence Lack of alignment
level Legal barriers + Political influence Internal
Resistance to Legal barriers resistance
change Political
instability
Techno-
logical + Nation-wide + Market solutions Security and Operational
level standards Authentication privacy problems
Integration Technical
ability coexistence
Service outcomes
Planned Library services City portal Councillor n/a
Problem Payment support
reporting platform
Elderly homes Library services
control Tourist guide
Environment
insp.
Imple- City portal Parking Traffic wardens City portal
mented Appointment information Food and Radar control
service veterinary Parking payment
Parking inspection
information
Food inspection
Failed n/a n/a Live townhall n/a
meetings
General Currently: Bring online services Few focused Arbitrary
outcomes information to mobile devices applications with applications, less
provision, future: clear benefit focus on efficiency
more transactional
services
a
City size according to population (medium 100,000250,000; very large >500,000 inhabitants)
b
Figures rounded for anonymity
70
 4.2 Towards Transformational IT Governance

financing. Similar to case A, the interviewee also cites regulatory challenges on the
distribution of competences between different bodies, as well as data protection laws as
significant constraints. As a result, despite being very active in the e-government area,
the municipality has realized few mobile solutions, yet is planning to expand in this field,
see Table4.9.

Case C: An Efficiency-oriented Municipality

Case C is one of many German municipalities that must manage their resources in accor-
dance with a budget-balancing concept.8 IT operations have been entirely outsourced to
a shared communal service provider. Similar to case A, the CIO and his staff, who are in
charge of IT steering, are subordinate to the central office for personnel administration
and organization. The municipality has established an internal idea management process,
where employees from all kinds of departments may hand in ideas which are subsequently
evaluated by the CIO and the central office.
The CIO considers m-government as an important topic. In course of cost savings and
budget cuts, the administration is permanently forced to check for feasible alternatives.
This mainly refers to the internal side (G2G), as such closed user groups mitigate the
investment risk. Despite the growing popularity of internet connected devices, he sees
less potential on the citizen side, as there seems to be only a hand full of younger
citizens that would use m-government services and those are harder to influence as a
target group.
On the governmental level, stakeholders are partly affine, partly averse, leaving it to IT
to assure, that estimated costs-benefits are kept. A major concern on the technological
level is, that a redundancy of m-government with traditional e-government channels, e.g.
for mobile ticketing, will generally lead to higher costs. As an outcome, three out of four
ideas for m-government improvements have successfully been implemented, two of which
are internal applications. The fourth, a live internet transmission of town-hall meetings,
has been dismissed due to an uncertain citizen demand (see Table4.9).

Case D: An M-Government Laggard

Size and financial situation case D municipality are comparable to case A. Despite a
considerable debt level, the municipality benefits from a relatively strong economic
environment and corresponding tax incomes. Over the past years, the IT organization
has undergone several changes. The current setup is comparably fragmented with more
than half of the 250 IT employees being located in the main departmental areas. Central
ICT is operating infrastructure, telecommunications and networks as well as the city
website, and moreover generates business with external clients. There are two main
8
The German Local Government Code provides that municipalities must manage their resources in
accordance with a budget-balancing concept (Haushaltssicherungskonzept), in case expenditures ex-
ceed income in the municipal cameralistics. Such municipalities no longer have the freedom to decide
whether certain voluntary tasks should be pursued.

71
4 IT Governance and Innovation Adoption in E-Government

governance bodies for IT decisions: IT steering meeting and IT planning group. The
first comprises the IT heads of each IT department and meets 1-5 times a year to inform
about new IT projects. Yet, a joint decision-making is only required only for large volume
projects. IT planning group consists of two department managers from central IT and
two from the office for personnel administration and organization. They jointly decide
on technological and organizational guidelines of the municipality.
Central IT seems to have limited insight into user level drivers for m-government.
According to the department head, there are hardly any requests to central ICT to offer
any new services due to the decentralized structure. With the exception of the mobile
city portal, all existing initiatives and realized applications stem from the solo efforts
by departmental IT units, e.g. a parking payment application realized by the municipal
traffic department. In some cases IT steering meeting was not sufficiently able to create
alignment on such developments. This fact may even have induced operational problems,
such as bandwidth problems and missing infrastructure support, when involving central
ICT at a very late stage.
On an economic level, the objective of achieving cost savings is perceived strongly.
However, central ICT as well as IT planning group largely fail to lever efficiency improve-
ments due to a lack of empowerment and internal resistance, especially when processes
and resources are concerned. The main strategy to overcome such resistance is seen
in take an indirect way via the political level. For this reason, currently a strategic
paper including e- and m-government elements is being elaborated. Once the political
level adopts an idea, there may be a stronger momentum for renovation of the service
landscape. However, the current outcome concerning mobile services is rather seen as
the result of the departmental initiatives than of an overall strategy and efficiency goals
(see Table4.9).

Case comparison: Contingencies for M-Government Adoption

Regarding the service outcomes, we argue that municipality A has the strongest, cases
B and C medium, and case D the lowest adoption of m-government services, which is in
line with the quantitative findings provided in (Winkler and Ernst2011). Furthermore,
when exploring the target groups (citizens, business and government employees) we find
that municipality A is focusing on all three of them, while B explicitly excludes citizens due
to a perceived lack of demand from this user group. Municipality C is even more
restrictive and relates m-government primarily to closed user groups, i.e. internal staff.
For municipality D, no clear user focus could be recognized.
As a municipality context we explored size, economic profile and budgetary situation.
Concerning size we selected two comparable cases each for mid-sized (B, C) and very
large (A, D) cities. Based on the presented cases, we conclude that there is no support
for a coercive correlation between the size of a municipality and the adoption of mobile
services. In respect to the financial situation (economic profile and budgetary situation),
we find more gradual differences between the four cases. Cases A and D exhibit a
comparably good, case B a moderate-poor and case C a poor financial situation. We

72
 4.2 Towards Transformational IT Governance

Table 4.10:Contingencies for m-government adoption and target groups

IT governance
Transformational Non-transformational
Strong Target groups:
G2C, G2B, and G2G
(Case A: Innovator)
Financial Moderate Target groups: Unfocused adoption
situation G2B and G2G (Case D: Laggard)
(Case B: IT experienced)
Poor Target group:
G2G
(Case C: Efficiency-oriented)

consider this fact to be an indicator of a relationship between the financial situation of a

municipality and the outcomes of m-government adoption for cases A, B and C. However,
for case D there seem to be further organizational contingencies.
Properties of the IT organization refer to horizontal distribution, governance mecha-
nisms and sourcing degree. Regarding horizontal distribution and sourcing degree, we
find considerable differences between cases A, B, C on the one hand, and case D on
the other. We draw on IT governance theory and deduce that case D exhibits an unbal-
anced allocation of decisions rights between centralized and decentralized IT departments
(Weill and Ross2004a). Although municipality D has certain governance mechanisms in
place, such as decision committees and an investment approval process, these practices
apparently do not work effectively. Thus, decisions which enhance the efficient use of e-
and m-government technology encounter large internal resistance. Moreover, unlike in
cases A, B, C, the responsibilities for IT and organizational issues are organizationally
separated, so that central ICT is not given sufficient decision rights to promote changes
of a transformational kind, which amongst other effects, also inhibits m-government
adoption. Concluding, we propose a new dimension for the IT governance property
which will tell us, whether IT is equipped with such transformational decision rights or
not (non-transformational). The resulting contingencies are summarized in Table4.10.

4.2.7 Conclusion

This work used a multi method qualitative approach to explore the factors that influence
public sector institutions in the adoption of mobile government services. We condensed
these factors in a novel and empirically well grounded framework and demonstrated how to
apply such framework in four case examples.

73
4 IT Governance and Innovation Adoption in E-Government

Findings and Contribution

Our findings suggest that process improvements and expected citizen benefits are among the
strongest and most quoted drivers for m-government while the ability of the admin-
istration to change as well as technical integratability represent some of the strongest
inhibitors. Furthermore, municipalities choose strategies such as inter-communal coop-
eration and increased citizen invovement to foster innovation diffusion and adequately
address their target groups. The framework developed (Figure4.7) may serve as an
orientation for practitioners and academics who wish to better understand the set of
factors that are critical in m-government adoption.
The framework is generally in line with the burgeoning literature on m-government.
However, it throws a new light on the graduation between the factors, which have tradi-
tionally focused more on issues of privacy and security as well as accessibility. Privacy,
security and the pertinent legislation still remains an important issue for local administra-
tions, yet with less perceived impact on m-government adoption than literature suggests.
The latter, accessibility, did not emerge as a major factor from our analysis. This reflects
that the attention of local governments meanwhile has shifted from infrastructure to user-
related issues so that municipalities today approach m-government with increased
determination.
Examining the differences between those municipalities that lead the way and others
that follow in m-government diffusion, the results of the case comparison provide support
for the contingent influence of the contextual variables financial situation and IT gov-
ernance. Building on these findings, we argue that municipalities with an economically
challenging environment should first focus on internal m-government applications, while
those with a comparably good financial situation may have the freedom to exploit the
full range of citizen-, business- and emplyee-oriented m-government applications.
To underline our findings regarding the second contenxtual variable, we introduced the
concept of transformational IT governance. This concept helps us to demonstrate that
only those public sector agencies will succeed in transformational projectssuch as m-
governmentthat are able to effectively connect responsibilities for IT and organi- zation
so that organizational change can be managed and resistance be mitigated. We hereby
enrich literature by a currently underrepresented, but important strategic aspect in e-
government research and provide appropriate mid-range theory for m-government
adoption on municipal level.

Limitations and Future Work

This work has some limitations which should be considered when interpreting the find-
ings. First, due to the sample size of 12 municipalities, the theory developed may possess
limited generalizability or leave out further important facets of m-government adoption.
Second, the legal framework as well as the culture in Germany may be different from
other countries which may limit the applicability to other national contexts. Finally, the
voluntary participation in interviews may lead to a non-response bias of the interview

74
 4.2 Towards Transformational IT Governance

data. However, these limitations are inherent to the qualitative approach, since for case
study research a statistical sampling is generally not required (Merriam1998). In a
future work, we aim to validate the proposed theory in larger, confirmatory studies and
investigate m-government adoption across different national contexts.

4.2.8 Summary
Using mobile services for e-government opens a new way of interaction between govern-
ments, citizens and businesses as well as within the public administration. However,
governments face different drivers and inhibitors in their adoption behavior. In this
chapter we investigated the question why public sector organizations differ in their
adoption of mobile government innovations. Based on 12 in-depth interviews with
municipal IT decision makers, we applied grounded theory and content analysis to derive
a framework for drivers and inhibitors of m-government adoption. Further, we utilized
this framework to analyze four cases in detail and identified the contextual differences in
such municipalities. Finally, we isolate the effects in the way IT is governed and proposed
the concept of transformational IT governance to explain some of the key differences in
municipal m-government adoption.

75
4 IT Governance and Innovation Adoption in E-Government

4.3 Citizen Acceptance in Urban Sensing

4.3.1 Preamble

This chapter has been initially published and presented at the 20th European Conference
on Information Systems, 2012 in Barcelona, Spain (seeWinkler, Hirsch, Trouvilliez, and
Gnther2012a).

4.3.2 Introduction

Urban sensing is receiving high attention as an emerging paradigm in pervasive comput-

ing (Cuff et al.2008). This paradigm refers to understanding todays mobile devices
which are increasingly capable of capturing and transmitting image, audio, location
and other dataas well as their users as sensor nodes of large information networks
(Burke et al.2006). Such sensing data can be useful for a broad range of applications
of public interest (Johnson et al.2007, p. 5), such as traffic and pollution monitoring
(Kinkade and Verclas2008, p. 52), environmental impact assessment (Mun et al.2009),
noise control (Maisonneuve et al.2010). In particular, urban sensing may also enhance
institutionalized e-government practices by the mobile channel and enable new ways of
citizen participation.
Citizen participation is seen as an important building block for accountable and trans-
parent urban governance and has been primarily studied under the aspect of public
influence in policy making (Irvin and Stansbury2004). However, ever since citizens have also
been involved in public service delivery and helped to co-produce public value, for example
when contributing or non-contributing to the regular activities of government officials
(Whitaker1980;Alford2002). The enablement of such participation has been described as
the highest stage in e-government evolution (Moon2002). Emerging mobile information and
communication technologies (ICT), such as urban sensing, allow for new possibilities of
citizen participation and thusgiven a wide acceptancemay help to reduce costs in
both, public decision making and public service delivery (Kumar and Vragov2009).
A number of recent studies have investigated the citizens acceptance of e-government
services (e.g.,Carter and Blanger2005;Dimitrova and Chen2006;Veit et al.2010) as
well as the user acceptance of mobile services (e.g.,Wu and Wang2005;Wang et al.2006; Xu
et al.2009). However, few works have addressed the peculiarities at the intersecting domain
of mobile e-government services, in particular urban sensing. Most notably, the link
between the citizens propensity to participate in urban affairs and the use of urban sensing
is yet unclear ([Link] and Paulos2010). A significant link would be indicative
that offering participatory ICT tools is an effective means to enable more citizen
participation (while insignificance would suggest that acceptance rather depends on other
factors such as opportunistic motivations or enjoyment). Hence, we formulate our
research questions: a) How can we explain the citizens acceptance of an urban sensing
application and b) Is there a link between citizens willingness to participate

76
 4.3 Citizen Acceptance in Urban Sensing

and the use of urban sensing applications?9 These questions are of crucial importance
not only from a research perspective, but also for urban decision makers planning to
implement participatory ICT tools.
In this chapter we investigate the user acceptance of a mobile reporting service, i.e. a
participatory sensing application that enables citizens to report urban infrastructure issues
such as potholes, waste and other defects. By the use of a mobile device, such report
can be sent to the local authority directly on the spot, ideally tagged with a photo and
according location coordinates. We chose this practical example of urban sensing as
there are first mobile reporting applications (apps) available in the real-world. For
instance, in Germany currently an increasing number of cities implement such service as
an integral part of their mobile e-government offerings (Vitako2011, p. 11). Moreover,
independent private or semi-public service providers are pushing into this market and
potentially cooperate with municipalities in offering web platforms that can be accessed
by government officials as well as the citizens mobile devices.10
As a theoretical framework for our study we draw on the seminal literature on tech-
nology acceptance (Fishbein and Ajzen1975;Davis1989;Venkatesh and Davis2000;
Venkatesh et al.2003;Gefen et al.2003) and combine our research model with appro-
priate constructs from e-government and mobile commerce research. To capture the
unique characteristics of participatory urban sensing, we theoretically develop the new
antecedent constructs environmental awareness, willingness to participate as well as mo-
bile literacy. Empirical tests using data from 200 potential adopters support the validity
and predictive power of all three constructs, which allows for interesting implications
from a theoretical and a practitioners standpoint.
In the remainder of this chapter we will first review the theoretical foundations and
develop our research model (4.3.3). In Section4.3.4we explain our methodology, the
measurement instrument and sample characteristics. Section4.3.5analyzes and discusses
the empirical results. Finally, Section4.3.6concludes by outlining the implications,
limitations, and future work.

4.3.3 Theoretical Foundations and Model Development

Since urban sensing is a relatively new phenomenon, we root our research in the widely
accepted technology acceptance model (TAM) byDavis(1989). TAM has been proven to
provide robust predictions of intended use, even if a technology is not yet fully available to
its prospective users (Sheppard et al.1988;Davis et al.1989). The overall research model is
depicted in Figure4.8.
9
This work focuses on the intention to use urban sensing. Although there are numerous factors that
possibly influence individuals in translating their intentions into behavior (Morwitz1997), the in-
tention to use is still inherently linked and strongly correlated with actual use (Davis et al.1989;
Sheppard et al.1988)
10
Examples are amongst others City Sourced ([Link]) in US, FixMyStreet
([Link]) in UK, as well as Mark-a-Spot ([Link]) and WDW-Anliegen
([Link]) in Germany.

77
4 IT Governance and Innovation Adoption in E-Government

Environmental H4 (+) Perceived

Awareness Usefulness H1 (+)
H5 (+)

Willingness to Intention to Use

H3 (+)
Participate
H7 ()

H2 (+)
H6 (+) Perceived Perceived
Mobile Literacy
Ease of Use Privacy Risk

Figure 4.8:Research Model

Technology Acceptance in Urban Sensing

Urban sensing can be regarded as a class of information and communication technologies

that are embedded in a whole system of actors and artifacts. Such system would typically
consist of the citizens, their mobile devices running the urban sensing application, a
central server processing messages from the citizens, and a website presenting according
information to citizens and recipients of the urban administration. The citizens individ-
ual decision of adopting (i.e., installing the application on his or her mobile device) and
using such technology (i.e., sending reports when encountering an infrastructure issue)
can be explained by drawing on TAM.
TAM has been developed byDavis(1989) based on the theory reasoned action (Fish-
bein and Ajzen1975, TRA,) and was found to be a robust theory for explaining tech-
nology adoption on user level. With regard to technology adoption, TRA essentially
states that beliefs about information technology influence the users attitudes, which
subsequently lead to behavioral intentions and actual technology usage. According to
TAM, beliefs about information technology can be attributed to the two dimensions
perceived usefulness (PU) and perceived ease of use (PE). TAM provides that both
constructs may be influenced by further external variables that capture specific beliefs
about the respective technology (Davis1989). Numerous adaptations and extensions of
TAM have been proposed (seeLegris et al.2003;Yousafzai et al.2007, for overviews).
Particularly the users attitudes and their mediating influence have been removed from
the model in a later revision (Venkatesh and Davis2000).
Despite its origin in organizational contextswhere the users of a technology are
typically employeesTAM has also been transferred to the adoption of diverse non work-
related technologies (e.g.,Gefen et al.2003). In such voluntary settings, the influence
exerted by subjective norms has been found to be less significant (Venkatesh and Davis
2000;Venkatesh et al.2003). The scenario of an urban sensing application offered by
a local municipality represents such a voluntary setting. It can be attributed to the
intersection of mobile commerce and e-government, an emerging field that has also been
termed as mobile government (Kushchu and Kuscu2003). We will draw on the literature
in both intersecting fields to develop our research model.

78
 4.3 Citizen Acceptance in Urban Sensing

Intention to Use, Usefulness and Ease of Use in Urban Sensing

Since urban sensing is a rather new phenomenon, we focus on the intention to use (IU) as
the dependent variable of this research. According to TRA, behavioral intention can be
interpreted as the subjective probability that a person will perform a specific behaviour,
e.g., to adopt an urban sensing application (Fishbein and Ajzen1975). Although indi-
viduals tend to overestimate their intended use, it has been shown that this variable
strongly correlates with the later adoption and use of a system (Sheppard et al.1988;
Davis et al.1989).
Perceived usefulness (PU) can generally be defined as a measure of the individuals
subjective assessment of the utility of an information technology in a specific task-related
context (Gefen et al.2003). In case of a citizen and the voluntary use of urban sensing,
perceived usefulness may refer to the individual task-related context as well as to the
broader collective utility of an urban context. Individual utility relates to convenience,
efficiency and effectiveness when reporting issues directly on the spot via a mobile de-
vice, compared to the traditional ways of reporting (i.e., by phone calls or emails to
the urban administration). Consequently, collective utility is generated if the collective
usage leads to according effects on the overall quality of living, e.g. through a cleaner
environment and a higher responsiveness of the urban administration. That is, we are not
restricting perceived usefulness to an individual performance expectancy of the citizen
([Link] et al.2003), but include the idea of collective usefulness and public value
within this construct (Alford2002, p. 33). Some authors argue that inclusion of such
domain-specific characteristics in PU can provide a better understandability of the overall
model (Yousafzai et al.2007, p. 300). In line with TAM we pose
H1: Perceived usefulness (PU) will positively affect the intention to use (IU) urban
sensing.
According to Davis (1989), perceived ease of use (PE) is defined as the degree to which
a person believes that using a particular system would be free from effort. Ease of use
is also one of the central goals in human computer interaction and usability engineering
(e.g.,Nielsen1994). Usability can be achieved by creating systems that are easy to under-
stand, easy to remember, efficient to use, and subjectively pleasing (Nielsen1994, p. 26).
Regarding a mobile application for urban sensing, this may refer to user interface design as
well as the usability of the mobile device itself. It has been shown that users largely base
their usability assessment on general beliefs and previous experience, especially for
technologies that are new to them (Venkatesh et al.2003). Therefore system-independent
factors are at least equally important to objective usability criteria for the ease of use
perception. According to TAM, perceived ease of use not only influences the intention
to use (IU), but also perceived usefulness (PU). That is, the easier an urban sensing
application is to use, the more useful it can be. We pose:
H2: Perceived ease of use (PE) will positively affect the intention to use (IU) urban
sensing.
H3: Perceived ease of use (PE) will positively affect the perceived usefulness (PU) of

79
4 IT Governance and Innovation Adoption in E-Government

urban sensing.

Antecedents of Perceived Usefulness in Urban Sensing

TAM provides that the core constructs (PE, PU) may be explained by external variables
which are more specific to the technology (Davis1989). Urban sensing applications can be
regarded as large information networks (Burke et al.2006). Social network theory
provides that the behavior of each of the nodes within this network can be characterized
by opportunistic or altruistic behavior (Hui et al.2009). We build on this dichotomy
and propose two antecedents for perceived usefulness: environmental awareness (EA) and
willingness to participate (WP).
Environmental awareness (EA) refers to the opportunistic motivation for using urban
sensing. The fundamental idea of urban sensing is to capture and counteract environ-
mental issues such as traffic, noise, or infrastructure defects (Burke et al.2006). Yet,
citizens presumably differ in their perception of the urban environment. Therefore we
define environmental awareness (EA) as the degree to which an individual is concerned
about the physical state of his/her direct urban environment. Those citizens who are
highly aware about their environment will possibly find an urban sensing application
more useful, because they have an inherent interest that irregular environmental issues be
corrected. In a wider sense this construct can also be related to performance expectancy
in an organizational context (Venkatesh et al.2003;Gefen et al.2003). That is, the
urban sensing application helps environmental-aware citizens to better perform their
context-related task of reporting an issue. We pose
H4: Environmental awareness (EA) will positively affect perceived usefulness (PU) of
urban sensing.
Willingness to participate (WP) is the central construct of this research and captures
a more altruistic motivation for using urban [Link](2002) identifies willingness as
the principal reason for citizens to participate and co-produce public value. Based on
this, we define the willingness to participate as the degree to which an individual wants
to participate in public affairs and urban decision making (no matter by which means).
In Political Science, participation is regarded as one of the core aspect of civic engagement
for society (Dimitrova and Chen2006, p. 177). Yet, in the Information Systems field,
few works have explicitly considered this construct to explain e-government adoptiona fact
that may as well call for more interdisciplinary approaches, as some authors claim
(Carter2006, p. 9).Carter(2006) explains the adoption of Internet voting through political
motivations such as political interest, efficacy and mobilization, besides other technological
and demographic [Link] and Chen(2006) emphasize that civic mindness is
one of four major factors that influence the intention to use e- government [Link]
et al.(2010) add political motivation as new construct to
a TAM model for e-participation and social media tools. Their findings suggest that
political motivation does not have a moderating influence on the relationship between
EU and IU, but significantly influences adoption of e-participation tools as an antecedent

80
 4.3 Citizen Acceptance in Urban Sensing

of IU. However, these works mostly refer to (a) stationary Internet tools and (b) to the
policy making level. We assume that (a) the use of mobile Internet tools (b) on the
service delivery level is also likely to be an extension of the citizens political involvement
via traditional channels (cp.Whitaker1980). Therefore we pose
H5: Willingness to participate (WP) will positively affect perceived usefulness (PU) of
urban sensing.

Mobile Literacy and Perceived Privacy Risks in Urban Sensing

Since todays mobile devices are becoming increasingly smart, i.e. powerful in function-
ality and connectivity, they set new requirements regarding basic user skills and trust
([Link] et al.2007). We include mobile literacy of the user (ML) and the perceived
privacy risks towards mobile service providers (PR) as two further constructs of our
research model.
Other authors have suggested that the concept of literacy could be an influential
underlying mechanism that is currently underrepresented in technology adoption research
(Venkatesh et al.2003, p. 469). Based on the concept of computer literacy (Winter et al.
1997) we define mobile literacy (ML) as the perceived ability to use smart mobile devices
efficiently and effectively. This conceptualization can ultimately be regarded as a special
form of computer self-efficacy (Davis1989, p. 321). Social cognitive theory provides
that users strongly anchor their ease of use perceptions about an information system to
their computer self-efficacy, i.e. their perceived ability to use this system (Venkatesh
et al.2003, p. 455). Applied to the context of mobile technology, self-efficacy may refer
to perceived skills for browsing the mobile Internet, installing a mobile application and
handling touch-screens. Such general beliefs can be derived from own experience (i.e.,
the use of mobile applications other than urban sensing) or from observations of others
(Venkatesh and Davis2000, p. 192). Accordingly we pose:
H6: Mobile literacy (ML) will positively affect the perceived ease of use (PE) of urban
sensing.
Trust and trustworthiness play an important role in user acceptance of both, e-
commerce and e-government systems (Gefen et al.2003;Carter and Blanger2005).
Users assess different types of risks when engaging in online activities or transactions
(Wu and Wang2005, p. 722). This effect becomes even more prevalent with the use
of mobile services, where location and personalization information may unintentionally
be disclosed to a third party (Xu et al.2009). Inverting the definition for perceived
credibility fromWang et al.(2006), we define perceived privacy risk (PR) as the extent
to which a person believes that using a mobile service will not be free from privacy
threats. Literature suggests that individuals perform a privacy calculus in balancing the
risks with the outcome they receive as a return for providing personal information (Xu et
al.2009;Krasnova and Veltri2010)in case of urban sensing for example location, photo,
recording and potentially identifying contact information (Johnson et al.2007). Therefore,
constructs like perceived risk are commonly hypothesized to directly affect

81
4 IT Governance and Innovation Adoption in E-Government

the intention to use (Gefen et al.2003;Wu and Wang2005;Wang et al.2006, e.g., in).
Accordingly we pose
H7: Perceived privacy risk (PR) will negatively affect the intention to use (IU) urban
sensing.

4.3.4 Methodology
A field survey was conducted to test the proposed research model. Since urban sensing
represents a class of information systems rather than a concrete application, we chose a
mobile reporting service, i.e. an urban sensing application to report urban infrastructure
issue (e.g., potholes, waste, and other defects) as a concrete scenario for our survey. In
the following we describe our approach to develop the measurement instrument and to
acquire a sample of citizen respondents.

Instrument Development

Model constructs were derived based on existing literature and operationalized on 5-point
Likert scales. New items were developed where necessary, especially for the proposed
constructs environmental awareness (EA), willingness to participate (WP) and mobile
literacy (ML). Also, established constructs such as perceived usefulness (PU) were com-
plemented by appropriate new items in order to adapt them to the specific context of a
mobile reporting service. We targeted at 5 items for each of these new or adapted
constructs, 4 items for the more established variables, and 3 for intention to use.
Based on our conceptualization as perceived attributes, all constructs are modelled as
reflective latent variables. That means that the causal direction is assumed to point from
the construct to the item (i.e., the item reflects the true value of the latent variable).
In order to allow for later consistency and response bias checks, we formulated one
reverse-coded item for each of the constructs (except for IU). The final questionnaire
was pretested with fellow researchers as well as potential respondents and revised in a
number of iterations. The resulting measurement instrument and originating literature
is presented in Table4.11. The online survey in German language is presented in
Appendix2

Sample Characteristics

The survey was administered online and distributed across the personal networks of the
authors. Apart from the model items, it contained a brief introduction to the topic
(including a series of images to explain the functionality of a mobile reporting service),
questions regarding mobile phone usage and demographic data, as well as the possibility
to leave an email address for receiving the later survey results. No further incentives for
participation were offered. Special attention was paid to the demographic distribution
of the recipients in order to achieve a balanced panel of different occupational and age
groups (i.e., not only students).

82
 4.3 Citizen Acceptance in Urban Sensing

100 others not stated

Age retired 16 5
75 24
studying
50 homemaking 64
7
25
working
0 84
<= 20 21 - 30 31 - 40 41 - 50 51 - 60 61 - 70 >70

Figure 4.9:Respondent age and occupation ( n=200)

The response period was three weeks during July 2011. Out of 320 participants who
opened the survey, 215 completed all relevant survey questions. Data cleaning was
performed carefully to filter for unrealistic or inaccurate answers. Apart from the reverse
items, we had included two redundant questions (Do you use a smartphone? no/ yes;
Are you planning to use a smartphone? I already use/ I am planning to use/ I might be
planning to use/ etc.). 15 respondents did not answer these questions consistently and
thus were rigorously removed from the sample. (Note that the term smartphone had
been defined earlier in the survey, however, using a smartphone was not a prerequisite
to participate in the survey.) The resulting distribution of survey respondents age and
occupational status (53% female, 45% male, 2% not stated; 45% users of smartphones,
55% non-users) is depicted in Figure4.9.

4.3.5 Model Analysis and Discussion

We employ partial least squares (PLS) to assess the psychometric adequacy of the
measurement model and test the hypothesized structural model. The choice of PLS
is motivated by the rather explorative character of this study including three new devel-
oped constructs. AsReinartz et al.(2009, p. 341) note, the variance-based PLS is less
likely to overestimate relationships between constructs that have potentially not been
well operationalized, compared to covariance-based approaches to structural equations
modeling such as LISREL. Calculations were performed using the software tools SPSS
version 17.0 and SmartPLS version 2.0.M3 (Ringle et al.2005).

Measurement Model and Common-Method Bias Assessment

As the first step in the model assessment we evaluate the indicator reliability. Indicator
loadings are listed in Table4.11. Values should be above 0.7 so that the reflective latent
variable explains at least 0.5 (0.72) of the variance of each indicator (Chin1998b). This is
not the case at least for some of the items adapted from literature (PR3 and PR4). We
ascribe the weak loading of PR3 to a confounding phrasing which omits the words I think,
thus asking for an agreement to an act that is actually legally prohibited (i.e., I will
typically strongly disagree that a mobile service can/ ought to divulge my data, while I
may still see high privacy risks in using it). Consequently, item PR3 is removed

83
4 IT Governance and Innovation Adoption in E-Government

Table 4.11:Measurement instrument, descriptive statistics and reliabilites

Item Text Load Source
IU1 I can imagine to use the mobile reporting service 0.90 Davis1989;
IU2 I think about using the mobile reporting service 0.83 Venkatesh et al.
IU3 I intend to use the mobile reporting service (MRS) 0.91 2003
PU1 A MRS would save time for reporting infrastructure issues 0.86 Venkatesh et al.
PU2 A MRS allows one to report problems on the spot 0.82 2003;Gefen
PU3 A MRS service allows people to report more infrastructure issues 0.83 et al.2003;
PU4 Overall, I find it useful to have a MRS for my city 0.93 Alford2002;
PU5r I think a mobile reporting service would be ineffective for my city 0.75 self-developed
PE1 The interaction with this service would be easy for me to understand 0.81 Davis1989;
PE2 Using a mobile reporting service would not be much effort for me 0.90 Venkatesh et al.
PE3 Overall, I think that a mobile reporting service would be easy to use 0.99 2003
PE4r I find it cumbersome to use a mobile reporting service 0.84
EA1 I am concerned about infrastructure issues in my environment 0.65 Self-developed;
EA2 I want any infrastructure issues in my environment to be removed 0.85 Burke et al.
EA3 I am interested in keeping my neighborhood clean 0.75 2006;Gefen
EA4 I appreciate if the properties in my city are clean and tidy 0.68 et al.2003
EA5r I dont care about infrastructure issues in my environment 0.79
WP1 I like to have an influence in my city 0.73 Self-developed;
WP2 I like to call the attention of my city on certain grievances 0.85 Dimitrova and
WP3 I like to take part in decision-making in my city 0.89 Chen2006;
WP4 I communicate issues to the municipality that I find important 0.76 Carter2006;Veit
WP5r I dont need any influence in my city 0.68 et al.2010
ML1 For me it is easy to use internet services on a mobile phone 0.94 Self-developed;
ML2 I am well versed in using internet services on a mobile phone 0.90 Venkatesh et al.
ML3 I could often use internet services on a mobile phone 0.84 2003;Winter
ML4r I need help with using internet services on a mobile phone 0.69 et al.1997
PR1 I think that service providers can abuse user data 0.81 ; 0.87 Wang
PR2 I am reluctant to provide personal information to a mobile service 0.83 ; 0.87 et al.
PR3 A mobile service can divulge my personal data 0.12 ; 2006
PR4r I think that one can trust the providers of mobile internet services 0.39 ; 0.54

r reverse-coded item; significance levels: p < 0.1, p < 0.05, p < 0.01

from the analysis.

A recalculation in PLS leads to a significant loading of 0.54 in PR4 (p<0.1). Also,
we note that reverse-coded items largely have lower loadings than their forward-coded
counterparts. However, we abstain from removing further items from the measurement
model, since all criteria for convergent validity (especially for PR) yield in acceptable
values. (For robustness, we also calculated models with strict removal of items PR4, EA1,
WP5, and ML4 which did neither change the principal outcomes of the hypotheses tests,
nor the relative strengths between the paths of the structural model.) Convergent validity
criteria of this revised model are above the recommended thresholds for Cronbachs alpha
(Alpha>0.7), composite reliability (CR>0.6) and average variance extracted (AVE>0.5),
see Table 2. This indicates that the remaining items sufficiently reflect the properties of
their respective constructs (Chin1998b).
Discriminant validity was assessed by three types of analyses: (1) an evaluation of cross-
loadings (Chin1998b, p. 321), (2) the Fornell-Larcker (1981, p. 46) criterion as well as
(3) an exploratory factor analysis (EFA). (1) The mean of absolute cross-loadings is 0.20
with a maximum at 0.59 (ML1 on PE), thus below the variance-explaining value of

84
 4.3 Citizen Acceptance in Urban Sensing

Table 4.12:Convergent and discriminant validity criteria

Convergent validity Discriminant validity (correlations and AV E)
Construct Alpha CR AVE IU PU PE EA WP ML PR
IU 0.86 0.91 0.78 0.88
PU 0.89 0.92 0.70 0.52 0.83
PE 0.89 0.92 0.75 0.56 0.56 0.87
EA 0.82 0.86 0.55 0.42 0.30 0.09 0.74
WP 0.83 0.88 0.59 0.28 0.17 -0.07 0.40 0.77
ML 0.86 0.90 0.70 0.29 0.25 0.57 -0.09 -0.06 0.84
PR 0.78 0.81 0.59 -0.13 -0.07 0.00 -0.08 -0.08 -0.16 0.77
(incl. PR3) (0.76) (0.64) (0.37)

0.7 for direct factor loadings. (2) The Fornell-Larcker criterion, which states that square
root of AVE (represented as diagonal elements) should exceed the off-diagonal elements
in the construct correlation matrix, is also fulfilled, see Table4.12. (3) The results of an
EFA (Kaiser-Meyer-Olkin criterion=0.84; eigenvalues>1) support discriminant validity,
since the principal components analysis produces seven factors that can be clearly distin-
guished after varimax rotation. Overall, the assessment of convergent and discriminant
validity support the psychometric adequacy of the revised measurement model.
As for all self-reported data, there is a threat for a common method bias (CMB)
due to the subjects motif to give socially desirable and cognitively consistent answers
(Podsakoff and Organ1986). We assessed CMB by a Harmans one-factor test as well
as a latent method factor approach (Liang et al.2007). The first factor from the EFA
accounts for 0.23 of the total variance (and predominantly loads on the indicators of
PU), thus contradicting the existence of a single dominant factor according to Harman.
Following the procedure described byLiang et al.(2007, p. 85), we included a common
method factor in the PLS model comprising all model indicators, and calculated the
influence on each principal indicator by its substantive construct and by the method
factor. The analysis shows that the average substantively explained variance is 0.69 while
the average method-based variance is 0.012 (ratio 44:1). Additionally, after bootstrapping
all substantive loadings remain significant (p<0.01) while most path coefficients from
method factor are not significant. Altogether these results indicate the method bias is
not a serious concern for this study.

Structural Model Assessment and Discussion

The results of the structural model assessment are depicted in Figure4.10. Statistic
significance of the parameter estimates (i.e., path coefficients in Figure4.10, as well as
factor loadings in Table4.11) was assessed through T-tests based on a bootstrapping
procedure using 1,000 resamples. For the purpose of hypothesis testing, the path coef-
ficients (c) and explained variances (R2) can be interpreted similar to parameters in a
simple regression.

85
4 IT Governance and Innovation Adoption in E-Government

0.19*** (H4) Perceived Hypothesis tests

Environmental
Usefulness
Awareness 0.29*** (H1)
R2=0.39 H1 Supported
0.14**
(H5) H2 Supported
Intention to Use
Willingness to 0.56*** R2=0.39 H3 Supported
Participate (H3)
H4 Supported
-0.11ns (H7)
H5 Supported
Perceived 0.40*** (H2)
0.57*** (H6) Perceived
Mobile Literacy Ease of Use H6 Supported
Privacy Risk
R2=0.32
H7 Not sup.
*p<0.10; **p<0.05; ***p<0.01; ns=not significant; n=200

Figure 4.10:Structural model results

H1, H1: Perceived usefulness (PU), perceived ease of use (PE) and (with certain
limitations) perceived privacy risks (PR) jointly explain R2 IU =0.39 of the variance in
the intention to use a mobile reporting service. This level of determination is largely
consistent and fully in the range of previous TAM studies (Legris et al.2003, p. 200;
Venkatesh et al.2003, p. 441). Yet, in our model the influence of perceived useful-
ness (cPU IU =0.29) is clearly weaker than the influence of perceived ease of use
(cPEIU =0.40). This is remarkable since most TAM studies suggest that even in
voluntary settings the influence of PU outweights PE (Venkatesh et al.2003, p. 441;
Yousafzai et al.2007, p. 299). We attribute this difference to the newness of the tech-
nology and a lack of concrete user experiences with urban sensing. Several authors have
demonstrated that PE is relatively more important in student samples and laboratory
settings (Yousafzai et al.2007, p. 299), while it becomes non-significant with increasing
experience of the users (Davis et al.1989;Venkatesh et al.2003, p. 433).
H3: In line with previous research, the results also demonstrate a strong significant
relationship between ease of use and usefulness (cPEPU =0.56). The explained vari-
ance in perceived usefulness R2P U=0.39 reduces to 0.10 when removing this link from the
model, which shows the strong influence of PE on PU (effect size 2 f
P EP U =0.48). Thus,
users not only anchor the intention to use, but also their perceived usefulness in their
ease-of-use perceptions of new urban sensing technologies (and similar experiences, as
we will see further below).
H4, H5: Regarding the antecedents of perceived usefulness (PU), we find that en-
vironmental awareness (cEAPU =0.19) and willingness to participate (cWP PU =
0.14) are both significant predictors in urban sensing adoption. However, the path
strengths indicate thatin the given case of a mobile reporting serviceopportunistic
goals connected to environmental awareness (i.e., wanting infrastructure issues to be
corrected) seem to be more important than the altruistic motivation of political partic-
ipation (i.e., to have an influence in public affairs and urban decision making). This is
not surprising inasmuch as a reporting service (and the promise that occurring issues
will be taken care of) largely targets the level of service delivery. Nevertheless, other
urban sensing applications (e.g., for long-term traffic monitoring or noise controlling),
may gather high-level information that targets more at the decision-making level. This
underlines the importance of considering willingness to participate (WP) as a dedicated

86
 4.3 Citizen Acceptance in Urban Sensing

construct within this adoption model of urban sensing.

H6: Mobile literacy is found to be a strong predictor of perceived ease of use (cMLPE =
0.57). This is in line with the concept of computer self-efficacy derived from social
cognitive theory (Venkatesh et al.2003) and confirms that users largely anchor their ease
of use perception in previous experience or observations of others. The strong path co-
efficient demonstrates that we were able to provide a well-performing operationalization
of this construct specifically for the context of mobile technology.
ns
H7: Perceived privacy risks (cPRIU = -0.11 ) are negatively correlated to intention
to use, however, do not significantly affect this construct (effect size 2
f P RIU =0.02).
Statistical reasons for this non-significant path have been ruled out during the assessment
of item reliability (see Section 4.1). Thus, we aim to provide a context-related explana-
tion for thisat first viewcounter-intuitive results: Perceived privacy risks have been
conceptualized as the extent to which a person believes that a mobile service will not be
free from privacy threats. Thus, similar to mobile literacy, this construct referred to the
general perceptions about (public or private) mobile services. In this survey, we described
the concrete scenario of a public mobile reporting service, i.e. a sensing application that
is offered by a local authority. Thus, although the citizens reporting information may
pass a number of third parties (e.g., network providers, platform providers) the recipient
remains a public entity. Factor analyzes performed byCarter and Blanger(2005, p 10)
indicate that citizen do not clearly distinguish between trust of the Internet and trust
of the government when using e-government tools. We argue, that a generally more
positive trust attitude towards the local government influences in this relationship, so
that privacy risks are not a severe issue for this type of sensing application.

4.3.6 Conclusion

In this work we investigated the citizens acceptance of urban sensing applications,

based on the example of a mobile reporting service. For this purpose, we concretized
a technology acceptance model (TAM) to our specific context and extended it by the
three external variables environmental awareness, willingness to participate, and mobile
literacy. Empirical tests employing the partial least squares method (PLS) and data
from 200 potential adopters support the psychometric validity and significance of all
three constructs. Most prominently, mobile literacy emerges as an important anchor
not only for the ease of use perception, but ultimately also for perceived usefulness and
intention to use novel mobile services such as urban sensing.
Regarding our focal question, we contend that indeed there seems to be a significant
link between the citizens willingness to participate in public affairs and the use of urban
sensing. However, this link is slightly outweighed by the individuals environmental
awareness, suggesting that opportunistic and utilitarian motivations will still prevail in
the use of urban sensing. Besides these results, we were unable to replicate the findings
of previous studies stating that trust, more specifically perceived privacy risk, plays an
important role in the adoption of mobile services. We attribute this finding to the given

87
4 IT Governance and Innovation Adoption in E-Government

scenario of a municipal reporting service and conclude that privacy risks seem not to be
a significant barrier to adoption of public mobile offerings.
These findings provide important implications for practice, foremost for public au-
thorities that are faced with decisions on their mobile e-government strategy. First, the
significant link between willingness to participate and perceived usefulness suggests that
mobile applications such as urban sensing are not just toys, but one out of a range of
possible tools to enhance citizen participation ([Link] and Vragov2009). To some
extent, this challenges the opinion still prevailing in many in urban authorities that
mobile government is just something for a hand full of younger citizens (Winkler et al.
2011b, p. 10). Second, concerns regarding privacy and data security, which are widely
discussed in public sector and often act as a barrier already on decision-making level
(Winkler et al.2011b, p. 2), can be largely dispelled based on the given findings. For
example, in case a specific municipal mobile service requires entering the users contact
details to ensure proper functionality, this is not likely to inhibit the citizens adoption.
Ultimately, attention in the development of mobile government offerings should rather
be drawn on the ease of use of these applications, in order to match usability criteria of
comparable (commercial) mobile services and thus reach a broad user base.
This work also aims to make a contribution to the research community. Originating
from the field of pervasive computing, the comparably young stream of urban sensing
has largely been driven by technical considerations and design-oriented works. To the
knowledge of the authors, this is one of the first works to investigate user acceptance in
urban sensing. We contend that behavioral research on end-user level can complement
urban sensing research to allow for a better transition to practice. We concretized
the robust and widely applied technology acceptance model (TAM) to the context of
urban sensing and demonstrated its use in a specific mobile government scenario. Other
researchers may use this model in different urban sensing scenarios and thus facilitate a
better understanding of user characteristics and their motivations in this emerging field.
A few limitations of this study merit consideration, foremost related to sample char-
acteristics and generalizability. Although representativeness to a basic population is
not a general requirement for correlation-based survey approaches, and variances from
voluntary participation were ruled out by common-method bias tests, we still acknowl-
edge that our sample exhibits a notable emphasis on students between 21-30 years.
Further, technology acceptance, especially privacy concerns, may be subject to cultural
influences. Most of the survey responses were acquired in Germany, a national context
with comparably great sensitivity to privacy issues (Krasnova and Veltri2010). However,
the results may still differ in other national contexts, for example due to greater distrust
in public governments (Carter and Blanger2005, p. 9). Finally, since the focus of this
work was on theory development and instrument validation, the model does not explicitly
take into account potentially moderating variables such as age, experience, and gender
(Venkatesh et al.2003). In a future work, we hope to conduct a larger study in different
national contexts and evaluate the moderating influence of demographic parameters on
the adoption of further (public and private) urban sensing applications.

88
 4.3 Citizen Acceptance in Urban Sensing

4.3.7 Summary
Urban sensing describes the use of todays mobile devices to collectively gather informa-
tion about environmental issues of public interest. Such information and communication
technology tools can enhance current e-government practices by enabling citizens to
actively participate in urban decision making and service delivery. Yet, it is widely
unclear whether there is a link between the citizens propensity to participate and the use
of urban sensing technology. In this chapter we drew on the seminal adoption literature
to propose a model for the acceptance of a mobile reporting service, i.e. a sensing
tool for reporting urban infrastructure issues to a municipality. The model explains
perceived usefulness of urban sensing by the citizens degree of environmental awareness
and his/her willingness to participate in public affairs. Furthermore, we conceptualized
mobile literacy as an important antecedent of perceived ease of use. Empirical tests using
data from 200 potential service adopters support these ideas. The findings also suggest
that for mobile e-government offerings, perceived privacy risks are not a significant barrier to
adoption. We outlined important implications for theory and practice.

89
4 IT Governance and Innovation Adoption in E-Government

4.4 Municipal Benefits of Urban Sensing

4.4.1 Preamble

The work presented in this chapter emanates from a joint cooperation with the State
Capital of Saarbrcken. We gratefully acknowledge the permission of the Saarbrcken
municipal administration and the support of the involved persons for conducting and
publishing this research. At the time of publishing this dissertation, this chapter has
been accepted for the Journal for Applied Electronic Commerce Research (JTAER) and
will appear in a special issue on Smart Applications for Smart Cities: New Approaches
to Innovation (seeWinkler, Ziekow, and Weinberg2012b).

4.4.2 Introduction

Local governments increasingly understand citizen participation as an important building

block of modern public administration. New public management poses that through par-
ticipation, citizens and public agencies can co-create public value which also strengthens
democratic authorization, legitimacy, and trust (Moore1995). While participation has
been primarily studied under the aspect of influence in public policy and decision making,
citizens have ever since also participated in public service delivery and thus co-produced
public value in the day to day routines of public administrations (Alford2002;Whitaker
1980).
Information and communication technology (ICT) plays an important role not only
to enable participation, but also to dramatically reduce efforts and increase the speed of
interaction compared to traditional ways of citizen involvement (Kumar and Sinha2007).
Today, the pervasiveness of mobile internet-connected devices, such as smartphones and
tablet PCs, makes it possible to gather opinions and environmental data from citizens
directly on the spot (Cuff et al.2008). Currently, an expanding research community
is developing participatory sensing systems for a broad range of applications, such as
environmental impact assessment (Mun et al.2009), air quality monitoring (Aoki et al.
2008), traffic mitigation (Hull et al.2006), and noise control (Maisonneuve et al.2010).
However, most of these initiatives are designed in a grassroots fashion as individual
projects, separate from the realm of integrated E-Government initiatives (Burke et al.
2006). One reason behind this may be that public institutions are generally thought of as
too slow to keep pace with the rapid developments in urban sensing and therefore place
more emphasis on mobile applications for equipping internal staff (Kamal2006). Some
laggard municipalities may even have the conception that mobile citizen participation is
more a nice-to-have than an absolute necessity to sustain the efficiency of public service
delivery (Winkler and Ernst2011;Winkler et al.2011b).
This study argues in contra, that it is worth to bridge the gap between participatory
sensing approaches and municipal E-Government strategies to embrace data from citizens
in urban management. We show that not only soft, but also hard quantifiable benefits can
be achieved for the municipality related to information quality and cost. As a scenario,

90
 4.4 Municipal Benefits of Urban Sensing

we consider a mobile reporting service, i.e. a participatory sensing application where

citizens can report urban infrastructure issues, such as potholes, waste and other defects to
the local authority, ideally tagged with a photo and according location coordinates. Such
services are becoming increasingly popular as part of the integrated mobile offerings of a
number of cities (e.g., a prominent example is NYC31111).
We take an action design research perspective and develop a simulation model as an
organization-dominant artifact (Sein et al.2011) that facilitates decision-making and
fosters the acceptance of such services among municipal stakeholders. The simulation
estimates the citizen adoption and benefits from a municipal perspective in terms of issue
awareness and required street inspections. The model itself and its usefulness in decision-
making contexts are demonstrated by the case of the State Capital of Saarbrcken, a
large German municipality. As currently various possibilities of urban sensing emerge,
the result of our work can aid local governments and service operators equally to engage
in public-private partnerships and jointly foster more urban citizen participation.
In the remainder of this chapter we will first review related work regarding urban sens-
ing and the simulation method. Then, in Section4.4.4we describe the methodological
approach of this research. Section4.4.5presents the proposed simulation model and
its underlying assumptions. Section4.4.6describes the results of the case validation
separated into the stages of process analysis, model instantiation, scenario simulation,
qualitative analysis and parameter estimation. Finally, Section4.4.7concludes by eval-
uating our methodology and outlining the contributions, limitations and future work.

4.4.3 Related Work

This section explains the underlying concepts of this research from a thematic and a
methodological perspective. This work can be thematically attributed to the intersecting
field of urban sensing and E-Government research. Methodologically we employ System
Dynamics (SD) as a method to design our simulation model.

Participatory Urban Sensing

Urban sensing describes the leap in pervasive computing from embedded networked
sensing in the laboratory to the real-world environment, e.g. in form of mobile phones
(Cuff et al.2008). As these smart mobile devices are increasingly capable of capturing,
classifying and transmitting image, acoustic, location, acceleration and other data, they
can be perceived as sensor nodes and location-aware data collection instruments (Burke
et al.2006). Thus, urban sensing and its numerous applications that emerge can be seen as
a first manifestation of the vision of ubiquitous computing (Weiser1991).
Depending on the role of the user in urban sensing, two fundamental interaction
patterns are differentiated. Participatory sensing refers to applications where conscious
human interaction is required to decide which data is shared, while in opportunistic
11
Non-emergency information service New York City, USA ([Link]/apps/311/[Link])

91
4 IT Governance and Innovation Adoption in E-Government

sensing the device acts autonomously and acquires data whenever its state matches
defined context requirements (Lane et al.2008). The term participatory sensing itself
has been coined byBurke et al.(2006) and was primarily motivated by the idea to enable
grassroots campaigns that gather data on issues of public interest.
Over the last few years, some municipalities have embarked on the concepts of partici-
patory sensing. Regarding the given scenario, an example for a participatory application is
the service NYC311 from New York City11 where users can report various local
problems via their mobile device. In contrast, other applications (e.g., Street Bump
piloted in Boston12) aim to report potholes and other street defects based on location
and acceleration profiles of a driving vehicle, thus taking an opportunistic approach to
sensing. In addition, there are a number of research projects that make tremendous
technical advances for supporting these and further application scenarios.
An example provided by a team at University of California Los Angeles13 is PEIR , the
personal environmental impact report, a system that uses GPS data from mobile phones
and classification algorithms to determine the users mode of transportation and travel
routes. From this information the service deducts information on the environmental
impact, e.g. the carbon footprint and pollution impact, and gives the user feedback
(Mun et al.2009). The project Common Sense at Intel Labs Berkeley 14 focuses on the
development of mobile sensors for collecting pollution and air quality data (such as carbon
monoxide, ozone, nitrous gases, temperature and humidity) as well as the visualization
of data in sample applications. The device is designed such that it can be easily carried
along by a person and transmits environmental data to via in-build GPRS or through
the users mobile phone (Aoki et al.2008).
Further examples are the CarTel projects at Massachusetts Institute of Technology15,
which address road transportation issues including traffic mitigation and pothole detec-
tion. The traffic mitigation application uses location data from mobile phones as input to
algorithms for traffic analysis, prediction, and according traffic-routing suggestions. The
pothole detection uses dedicated sensing devices mounted on vehicles, which measure 3-
axis acceleration and location data (Hull et al.2006). A solution that addresses noise
pollution has been presented by the Sony Labs Paris and is now maintained at Vrije
Universiteit Brussel16. The system NoiseTube uses mobile phones as acoustic sensors
and records noise levels along with location data. These are combined in a noise map
to raise the awareness for urban noise pollution. In addition, labels reflecting the noise
sources can be assigned by the users or inferred by the system (Maisonneuve et al.2010).
While many of these research projects have yet outgrown the realm of grassroots
initiatives, only few have been incorporated in the official services of local governments

12
Street Bump, New Urban Mechanics, Boston, USA ([Link])
13
Center for Embedded Networked Systems (CENS), University of California Los Angeles,
USA([Link]
14
Common Sense, Intel Labs Berkeley, California, USA ([Link])
15
CarTel, Computer Science and Artificial Intelligence Laboratory, Massachusetts Institute of Technol-
ogy, USA ([Link]
16
NoiseTube, BrusSense group, Vrije Universiteit Brussel, Belgium[Link]

92
 4.4 Municipal Benefits of Urban Sensing

Urban Sensing E-Government

Opportunistic M Government
M-G
External Internal
Participatory

Mobile reporting service

Figure 4.11:Field of research

and reached a broader user base. One of the reasons for this may lie in a low acceptance
of urban sensing applications among administrative stakeholders and a lack of insights on
the potential benefits. Nevertheless, when a governmental body acts as the recipient of
sensing information, similar to the two examples mentioned above (NYC311 and Street
Bump) we can instantly relate this stream of research to the emerging field of Mobile
Government.

Mobile Government

Mobile Government (M-Government) can be defined as an extension of E-Government in-

volving the utilization of all kinds of wireless and mobile technology, services, applications
and devices for improving benefits to the parties involved in E-Government (Kushchu
and Kuscu2003). Depending on the target group, we may further differentiate between
internal and external M-Government. The former is mostly concerned with equipping
governmental staff (e.g., traffic wardens, food and veterinary inspectors, fire-fighters and
police) with mobile devices to improve internal processes. External M-Government refers
to applications that offer mobile services to the customers of government agencies (i.e.,
citizens and businesses) and expand the current service offering to this new channel of
communication. Participatory urban sensing promises new applications particularly for
external, citizen-centric M-Government (Rossel et al.2006). Figure4.11illustrates the
relationship between the mentioned fields and positions the service that his study
addresses.
Challenges and success factors of M-Government (i.e., internally and externally) are
addressed by a number of authors (seeNapoleon and Bhuiyan2010for a more structured
overview). The most critical issues on user level are particularly seen in accessibility,
user friendliness, quality, convenience, privacy as well as user education (Al-khamayseh
et al.2006;El-Kiki and Lawrence2007;Kumar and Sinha2007;Kushchu2007;Sandy
and McMillan2005). From a technological perspective, authors also emphasize the
availability of communication networks, payment infrastructures, security requirements
and compatibility between different online and mobile E-Government channels (Kushchu
and Kuscu2003;Kushchu2007;Winkler et al.2011b). Economically, most works find
that governments view investments in M-Government technologies as a means to achieve

93
4 IT Governance and Innovation Adoption in E-Government

potential savings (Sandy and McMillan2005;Winkler et al.2011b), rather than simply

a cost factor. Organizational challenges can be that sometimes centralized governance,
political support and process re-engineering are necessary to leverage these cost advan-
tages (Borucki et al.2005;Sandy and McMillan2005;Winkler et al.2011b). However,
other authors find that M-Government initiatives primarily occur at the local level and
rarely cause structural changes (Borucki et al.2005).
Prior work on M-Government also suggests that due to these drivers and inhibitors
of adoption, agencies place more emphasis on internal than on external M-Government
projects (Winkler et al.2011b). As many local governments are facing continual pres-
sure to save costs, the benefits in internal processes appear more tangible and hence
receive a higher priority. Moreover, government employees can be regarded of as a more
controllable group of users, so that possible user-level acceptance issues can be better
mitigated (Winkler et al.2011b). In contrast, citizen-centric mobile applications are
often perceived as a dispensable feature reserved for those municipalities that can still
afford to experiment with new innovations (Winkler and Ernst2011). This reluctance
and risk aversion is consistent with the broader literature, which generally views the
public sector to lag behind the private economy in innovation adoption (Kamal2006)
This is particularly lamentable from a citizen participation standpoint inasmuch as a
recent study indicates that urban sensing tools actually do provide significant means for
citizens to participate in urban management (Winkler et al.2012a).
Altogether, given that one of the main barriers of external M-Government adoption
particularly consists in the perceived uncertainty of process and cost improvements,
researchers may take a closer look at evaluating the benefits of urban sensing as a means
of citizen-centric Mobile Government.

Benefits Case Appraisal and System Dynamics

The appraisal of information systems (IS) benefits, or more broadly, the question how
information systems create value, has ever been a center of gravity in the IS literature
(Schryen2010). The literature provides a number of measures and methods to evaluate IT
investments prior to implementation (Ballantine and Stray1998), also specifically for an E-
Government context (Prakash et al.2001). Besides financial and operational indicators,
these often include qualitative aspects such as benefits and risks (Ballantine and
Stray1998;Farbey et al.1999;Irani et al.2002;Prakash et al.2001). However,
more recently authors also emphasize that the success of an evaluation (i.e., whether
the expected benefits have been appropriately estimated) is not solely dependent on the
appraisal method. These authors demand to view the investment decision as a staged
process that primarily targets at aligning the interests of involved stakeholder groups
(Irani2002;Ward et al.2008). In this context, a benefit case refers to a high-level
appraisal at the earliest stage of a project to justify the commencement before a detailed
business case can be conducted (Wallace2007).
During the appraisal process, stakeholders face different information requirements
(e.g., for quantifying relevant benefits), knowledge requirements (e.g., in understanding

94
 4.4 Municipal Benefits of Urban Sensing

Notation

t
Formal Level
rate (t) rate (t) dt Level(t )
t0
1 2 0 auxiliary(t) f (i1(t),..., in (t))
Script Level = INTEG(rate1rate2, L_t0) auxiliary = ...

Figure 4.12:System Dynamics syntax

the appraisal technique) and organizational problems (e.g., a lack of time and interest)
(Ballantine and Stray1998). For this reason we find rather simple methods among the
most common methods for evaluating risks and uncertainty, such as brain storming and
scenario planning, while simulation and other sophisticated techniques rank on the lower
places (Ballantine and Stray1998). However, these simple appraisal methods are often not
able to sufficiently capture the complexity of a target domain, such as dynamic and
mutually dependent behavior of variables (Farbey et al.1999). Especially regarding
the benefits of external M-Government services, where many variables of adoption lie
outside the scope of the organization and only limited prior knowledge is available, there
is high uncertainty in the investment evaluation. Therefore the use of an appropriate
appraisal technique, that is both comprehensible to stakeholders and still able to capture
the variable behavior of the target domain, appears to be of even greater importance.
System Dynamics (SD) is a method developed byForrester(1961) and his colleagues,
originally employed for explaining managerial and industrial problems. It essentially
combines a notation for modeling complex systems with a mathematical representation to
simulate a model behavior discretely over time. The main components of a model are level
and rate variables, as well as inputs (i.e., constants, parameters or lookups) and
auxiliaries. Mathematically, a model corresponds to a set of coupled first-order differ-
ential equations. The main feature of a dynamic system is presented by loops between
variables, leading to reinforcing or balancing feedback behavior. In this sense, SD has
the advantage to make complex systems comprehensible by disaggregating them into
chains of cause-effect relationships and thus facilitates the solution of intricate decision
problems. Prior authors have argued that this can be particularly helpful to improve
communication in public decision making (Stave2002). A summary of the basic SD
syntax is presented in Figure4.12.
Despite its origin in the field of management, SD has been used for a wide range of
hard and soft modeling problems in physics, biology, psychology and other domains. SD has
also been employed in strategic project management contexts to evaluate investment
decisions and project benefits (Lyneis et al.2001). In the field of Information Systems,
for example, SD has been used to evaluate investments into IT security and the effect on
the number of attacks (Behara et al.2007) as well as to model the benefit realization in the
adoption of enterprise resource planning systems (Heijkoop and Cunningham2007).
Therefore we contend that SD is also a useful method to evaluate the diffusion and

95
4 IT Governance and Innovation Adoption in E-Government

municipal impacts of participatory urban sensing. In our work, we use a SD simulation in

a real case from this problem domain and thereby obtain further insights on its practical
applicability.

4.4.4 Research Method

Our approach for developing and validating the proposed simulation model took place
in three main stages and has been oriented in an action design research methodology.

Methodological Perspective

Design research primarily seeks to address practical needs in a generalizable way through
creating new and innovative artifacts (Hevner et al.2004). Artifacts in this sense have
been classified as constructs, models, methods and instantiations (March and Smith
1995). Our simulation model can be regarded as an artifact of the type model inasmuch
as it aims to aid problem and solution understanding and the exploration of the
effects of design decisions and changes in the real world in the domain of external M-
Government services (Hevner et al.2004, p. 79). However, design research assumes that
building and evaluating an artifact follows a clearly sequenced approach and thus
implicitly separates the design process from its organizational application (Sein et al.
2011).
Our case organization has been studied and accompanied for a period of over 1.5
years, counting from the first impulses for implementing a mobile reporting service
until its rollout and incremental use by the citizens. For this reason, the development
process was inevitably characterized by several iterations as well as interventions of the
researchers, which have potentially influenced both the design of the artifact as well as the
design and implementation of the mobile reporting service itself (Baskerville et al.2007).
Action design research (ADR) approaches have emerged as an alternative methodological
lens that explicitly accounts for the iterativeness in the design process as well as the
interventions by the researchers. More than that, ADR has been positioned as a design
approach that can reflect the theoretical precursors and intent of the researchers, but
also the influence of users and ongoing use in context (Sein et al.2011, p. 40).
Depending on the source of innovation in the designed artifact, ADR differentiates
between organization-dominant and IT-dominant research approaches (Sein et al.2011).
Since the primary focus of the presented simulation models is on facilitating organiza-
tional decision-making (and not on developing the reporting service itself), our research
can be clearly classified as an organization-dominant approach. ADR posits that the
methodological approach should comply with several principles in order to ensure that
artifacts are both practically relevant and theoretically rigorous. These principles relate
to: (1) inspiration through practice, (2) ingraining through theory, (3) reciprocal shaping,
(4) mutually influential roles, (5) concurrent evaluation, (6) guided emergence, and
(7) generalized outcomes (Sein et al.2011). Therefore at its core, ADR foresees an
interwoven design process of building, intervention and evaluation (BIE) activities.

96
 4.4 Municipal Benefits of Urban Sensing

Researchers

Problem Raw Simulated Validated model Estimated model

ADR formulation Model scenarios and decision and case learnings
Team
Practitioners
(municipal Validation Process analysis
administration interviews and design Post-implementation
professionals) workshop interviews

End-users
Use
(citizen)

time
H2/2010 H1/2011 H2/2011 H1/2012

Figure 4.13:Build, evaluation, intervention cycles (adapted fromSein et al.2011)

Research Stages

Although the process for designing and evaluating the proposed decision model was
potentially not as highly interwoven as ADR would posit, it was clearly characterized
by three distinguishable BIE cycles. These stages and evaluation cycles are depicted in
Figure4.13. We will use a chronological order to describe the major activities, outcomes
and the interactions between researchers and practitioners (i.e., the ADR team,Sein
et al.2011) in each stage.
The contact to the City of Saarbrcken (south-western Germany) originated from prior
research on M-Government adoption. The municipality (represented by the Head of IT
Coordination) had expressed a major interest in evaluating the benefits of external E-
Government services. In a fist interaction regarding this research in late 2010, the Head of
IT Coordination stated that lack of cross-functional thinking and uncertainty regarding the
potential benefits were some of the reasons why this topic had been given little
attention in the past (problem formulation). The researchers developed a first version
of the presented simulation model (raw model ). In addition they prepared a catalog of
questions to acquire more information (e.g., internal instructions and documentation)
and to determine the first set of parameter values (e.g., for those related to street
inspectors). The validation interviews with the Head of IT Coordination and the Chief
of the Department for Street Maintenance took place in February 2011.
The raw model was refined in a second iteration. The researchers conducted a survey
and extensively drew on secondary data to make additional parameter estimations (espe-
cially for variables related to adoption and diffusion). Different scenarios were simulated
based on the given assumptions. In May 2011, the ADR team organized a workshop
that embraced a total of seven stakeholders from different departments and entities (IT

97
4 IT Governance and Innovation Adoption in E-Government

Coordination, Street Maintenance, Public Relations, Office for Data Privacy, Internal IT
department, Regional Association of Communal IT Providers) which allowed for a broad
assessment of related benefits and concerns. The goals of the workshop were
(a) to validate the model through a detailed process analysis and determine missing
data (especially the costs for processing reports) and (b) to present results of the model
scenarios and using them as quantitative basis for discussing solution design options.
(a) Process modeling can be regarded as a valuable validation method as it enables a
common understanding of a domain of interest (Aguilar-Savn2004). Compared to the
SD model, where flows are aggregated to continuous variables, process models provide
a more detailed representation of a process by assuming a discrete task and event logic
(Melo and Pidd2000). The validation facilitated the researchers with a structured
method to acquire the missing information from practitioners and their perceptions about
the relevant activities. Furthermore, practitioners themselves, who pertain to diverse de-
partments, got to a comprehensive understanding of the entire process which may likewise
have improved their ability to understand the domain of interest, in our case the defect
management procedures. (b) Two model scenarios (optimistic and pessimistic) were
presented to the workshop participants and enabled a discussion on further qualitative
aspects and design options for a mobile reporting service. For instance, the simulation
supported the reasoning about appropriate means for handling the anticipated amount
of citizen reports. Comprehensive workshop documentation was subsequently provided
to the participants (see Appendix4).
As a consequence of this joint workshop, the municipality (in responsibility of the Head
of Public Relations) decided to include a mobile reporting service in a planned mobile
offering, which was originally intended to incorporate merely non-transactional services
(i.e., no urban sensing or similar functionality). Application development and server
operation was organized through a public-private partnership between the municipality
and a local information portal operator. In September 2011, the city launched the first
version of the citys official mobile application for one popular mobile platform 17. The
application includes an event calendar, city walks, information on the zoo, as well as a
mobile reporting service called Mngelreporter (defect reporter). Application down- loads
as well as incoming reports from citizens grew steadily.
After about half a year of operation, the researchers conducted post-implementation
interviews with the Heads of IT Coordination, Street Maintenance and Public Relations
to evaluate the first learnings and obtain detailed usage statistics. Worth noting, in April
2012 the Saarbrcken mobile application has been awarded the prize as best municipal
app in a national competition at a major IT trade show, which was according to the
Head of Public Relations also due to the Mngelreporter. The Head of IT-Coordination
emphasizes that the initiative from researchers in the area of urban sensing was crucial
for the decision of whether integrating a mobile reporting service into this offering.

17
[Link]/de/rathaus/medien-_und_buergerkommunikation/saarbruecken-app

98
 4.4 Municipal Benefits of Urban Sensing

4.4.5 Simulation Model

The simulation model used in this research consists of four parts: (1) a diffusion model for
the adoption of a mobile reporting service among a citys inhabitants, (2) a deterioration
model that simulates the occurrence of new infrastructure issues, (3) a reporting model
that estimates the number of reported issues, and (4) a cost model. We will describe
each of these submodels separately by highlighting the required inputs, the underlying
assumptions, and the characteristic functional relationships. The modeling and simu-
lation was performed with Vensim PLE software. An overview of the entire model is
provided in Figure4.14. A full list of equations can be found in Appendix3.

Cost
model

Diffusion
model

Reporting
model

Deterioration
model

Figure 4.14:System dynamics representation

99
4 IT Governance and Innovation Adoption in E-Government

Diffusion Model

The diffusion model estimates the number of adopters of a mobile reporting service in the
municipality of interest. One of the most famous approaches for predicting adoption is the
model introduced byBass(1969), who extended innovation diffusion theory byRogers (1962)
by a mathematical representation. The Bass model is based on the behavioral
assumption that the probability for a new product being adopted in period t is a linear
function to the number of previous adopters At1.
At1
P (at) = p + q (4.1)
M
This probability is composed of two influences. The first is the influence represented by
the rate of innovators p who adopt a product independent from the number of previous
adopters. The second is given by the rate of imitators q whose adoption is dependent on
the share of past adopters by the total market size M (i.e., the relative past adoption).
This share has also been interpreted as a probability of contagion or word-of-mouth from
adopters to non-adopters. The Bass model has been transferred to a SD representation
bySterman(2000, p. 332). This model accordingly contains a balancing loop that
determines the rate of innovation, and a reinforcing loop which adds the imitators to
this to obtain the number of new adopters per period (at)
We embed the Bass model into our context of a mobile reporting service by defining
the following input parameters:

City population: the number of inhabitants of the city [citizen]

Mobile broadband development: the share of users of smart mobile devices on
country-level [lookup]
Potential user share: the expected share of users who would install the mobile
reporting application [dmnl]
Implementation time: the time ti after t0 when the mobile reporting service is
released [months]
Innovators rate: the share of potential adopters that would install the new appli-
cation [1/month]
Imitators rate: the pressure on potential adopters to imitate the adopters [1/month]

The level of potential service adopters at t0 is initialized with the percentage of mobile
users among the city population, multiplied with the potential user share. Obviously, only
users of a smart device can potentially adopt the service and install a mobile reporting
application on their device. The share of new potential adopters in a city is moderately
increasing with the overall mobile broadband development. When the implementation
time ti has been reached, the innovators among potential adopters incrementally become
new adopters. Once the number of adopters of the service increases, the number of
imitators increases and adds up to the new adopters. However, the probability for this
word-of-mouth advertising decreases with the decreasing number of potential adopters.

10
0
 4.4 Municipal Benefits of Urban Sensing

The equations of the diffusion model are listed in Appendix3.

Deterioration Model

The deterioration model simulates the occurrence of issues in the urban street infras-
tructure. The following inputs are required:

Street network size: determines the total urban space where infrastructure issues
occur [km]
Max spots per km: limits and discretizes the potential spots where issues occur
[spots/km]
New issues per km: describes the constant occurrence of new issues on these spots; in
other words the natural deterioration of urban infrastructure [issues/km/month]

This process assumes that a street kilometer is segmented into a number of spots
(for example 1,000 spots with each 10 square meters of space, assuming that the average
street width is 10 meters). Issues (e.g., potholes, waste, broken trees etc.) occur randomly
within these spots. Since several new issues may fall on the same spot, the number of
new spots with issues is less or equal than total new issues. (This also implies that spots
can infinitely get worse.) Logically, the probability for such overlap increases the more
spots already have problems, i.e. the higher the issue density is. Thus, overlapping
issues still represent a single new spot with an issue, since it would still require a single
report to the municipality. Before being reported, issue spots account up to the level of
undetected spots. The equations of this submodel are given in Appendix3.

Reporting Model

The reporting model connects the deterioration model with the diffusion model. Munic-
ipalities typically employ dedicated workforce for inspecting urban infrastructure. The
model calculates the number of reported issues based on the number of reports from
these street inspectors, plus those from citizens that have adopted the mobile reporting
service. The following inputs are required to parameterize this submodel:

Traditional reporting rate: number of infrastructure issues reported via conven-

tional means such as email and telephone [report/month]
Reporting rate of service users: number of reports per adopter of the service
[report/month]
Number of inspectors: permanent workforce for inspecting streets [inspector]
Walking km per inspector : the average length of the street network that an inspector
controls per month [km/inspector/month]
Repair duration: average duration for removing issues that have been detected
[month]

10
1
4 IT Governance and Innovation Adoption in E-Government

The model assumes that reports via traditional means are issued by non-adopters
whereas reports from the adopters replace traditional channels of communication. The
total number of reports from citizens, as well as the level of undetected spots are used to
calculate the actual issue spots found by citizens. Similar to the issue generation model,
reports from citizens are likely to overlap, e.g. if several citizens report the same spot
with a pothole. In this case, the problem can be referred back to an occupancy problem,
where citizens randomly distribute x balls (x=number of reports) on N urns (N =number
of undetected spots). According toGrottke and Rssler(2003, p. 6), the expected value
for the number of urns occupied with at least 1 ball (Ux=issue spots found by citizens)
is given by the following equation

1
E (Ux) = N 1 x
(4.2)
1
N
At the same time, issues are detected by the citys street inspectors. Street inspectors
accomplish a certain amount of total walking km per month. We assume that undetected
issues are geographically equally distributed over the street network and that street
inspectors find all the issues on their way, so that the share of issue spots found by
inspectors on undetected spots is presented by the fraction of the total street network
size covered. The effectively new reported spots (sCI) are given by the sum of spots
reported by citizens (sC ) and by inspectors (sI ), subtracting the spots sCI that overlap
between the two sets. We assume that the citizens reports are equally distributed over
the street network. Therefore, the expected share of the overlapping spots by sI is equal
to the share of sC by all undetected spots sU . We can write

sCI sC
in s
CI sI sC
sI = sU = sCI = sC + sU (4.3)
sU
This variable sCI reduces the level of undetected spots and adds up to the level of
reported spots. That is, citizens and street inspectors would not report these infrastruc-
ture issue anymore, as they have been logged and appear in the municipalities reporting
system with an according status of processing. The issue detection rate can be regarded as
a performance indicator for the overall effectiveness of this process. The further removal or
repairing of an issue takes an average repair duration and takes place subsequent to issue
reporting. The equations of the reporting model are given in Appendix3.

Cost Model

The cost model sums up the monthly costs related to infrastructure reporting, i.e. costs
for the mobile reporting service, for the processing of reports and for street inspectors.
For this purpose, a few input variables are required. The detailed equations employ
simple arithmetic, see Appendix3.
Implementation costs: the onetime costs for implementing the service in ti, e.g. for
application development or purchasing licenses [EUR]

10
2
 4.4 Municipal Benefits of Urban Sensing

Ongoing costs: the running expenses of the mobile reporting service, e.g. for
perpetual licenses or application maintenance [EUR/month]
Costs per traditional report: average variable costs for reviewing, classifying and
processing issues reported via phone or email [EUR/report]
Costs per mobile report: average variable costs for reviewing, classifying and pro-
cessing issues reported via the mobile service [EUR/report]
Costs per inspector per month: total labor cost per street inspector [EUR/inspector/
month]

4.4.6 Case Validation

We validated the correctness and usefulness of the simulation model (our artifact) in
course of the different research stages and interventions at the case site of the City of
Saarbrcken. This validation includes a detailed process analysis, a model instantiation
with case data, the simulation of different scenarios, the assessment of further qualitative
aspects and the estimation of model parameters. This section describes each of these
steps.

Process Analysis

Process modeling was performed in a workshop with different municipal practitioners and
stakeholders. The urban defect management process including defect identifica- tion,
processing and removal was collaboratively discussed and incrementally drawn on a
brown paper using process mapping methods (Biazzo2002). For reasons of quick
comprehension, we used a simple flow chart notation which was later documented by
the commonly known business process modeling notation BPMN 2.018 and provided to
the participants. This documentation consists of (a) an overview of the overall process
and (b) detailed documentation of its sub-processes. For reasons of brevity, we will
summarize the overall process and highlight two relevant subprocesses to explain, how
implementing a mobile reporting service affects these processes.
The overall complaint and defect management process in Saarbrcken is depicted in
Figure4.15. It starts either with a report from a street inspector or with a complaint
from a citizen. Citizen complaints are directed to the Complaints Office via telephone,
regular mail, email, a website form oras a new mediumas a message from a mobile
reporting service. The Complaints Office generally processes these issues and routes
them to the department in charge (e.g., to public order, traffic authority, public utilities,
or in the frequent case of infrastructure defects to the Head of Street Maintenance).
The same applies to those defects regularly reported by the street inspectors, who are
subordinate to the Office of Street Maintenance. Road workers responsible for repairing
minor defects (e.g., potholes) are situated locally at the Saarbrcken districts. In some
cases, district road workers are requested to perform an additional pre-inspection of an
18
Business Process Model and Notation, Object Management Group ([Link])

10
3
4 IT Governance and Innovation Adoption in E-Government

Figure 4.15:Complaint and defect management process and selected subprocesses

issue spot before (instantly or later) starting corrective maintenance (e.g., when size and
condition of a pothole are unclear). In cases of larger defects, these are typically bundled
and scheduled for a tender to external street construction firms (e.g., for asphalt coating
of an entire street). Both repair works from the wards as well as those by external
contractors are approved by the Head of Street Maintenance when completed. In case
the repair was due to a complaint, usually a feedback to the citizen is provided.
Based on the process analysis, we find that the principal flow of the defect management
activities is not affected by the existence (or non-existence) of a mobile reporting service.
That is, the Complaints Office remains the main interface to the citizens so that the
reporting service simply represents an additional communication channel. Nevertheless, the
analysis reveals two main levers in the complaints and defect handling subprocesses where
the choice of the medium may strongly affect throughput times, see Figure4.15 (right
side).
Regarding the complaints handling subprocess, the Head of Public Relations states
that many incoming messages are invalid or do not provide sufficient level of information
(e.g., regarding the precise location of a defect). Such messages cause additional time
and effort for inquiring the missing details with the citizen (if possible) instead of being
instantly forwarded to the department in charge. (This is considered later as costs per
mobile report for the model instantiation in Section4.4.6). Arguably, with the use of a
mobile reporting service, which includes on-the-spot location coordinates and possible
photos, the need for further inquiries and hence processing times can be reduced.
A similar effect is expected for defects handling. The Head of Street Maintenance states
that defects reported by citizens are usually no support for their work at all, since they
caused more effort for sending road workers and performing pre-inspections than they
actually save in preventing damages. Even in some cases, the pre-inspectors are unable
to locate the reported spot so that additional feedback loops with the Head Office and

10
4
 4.4 Municipal Benefits of Urban Sensing

citizens are required. This would certainly change if the report featured precise location
coordinates and a photo, as the interviewee affirms.
Altogether the process analysis indicate (a) that the assumptions made in the pre-
sented system dynamics model match the concrete process instance at the Saarbrcken
municipality, and (b) that there are certain levers associated with the use of a mobile
reporting service that would facilitate the process flow and reduce processing times.

Model Instantiation

Based on the case analysis, the simulation model has been instantiated with values of the
Saarbrcken municipality, survey data and information from further data sources. We
use the submodel logic to explain the parameters used and how they have been estimated.
Regarding the diffusion model, we retrieved the percentages of the mobile broadband
development in Germany from the annual report of the International Telecommunication
Union (ITU2010). These figures have been extrapolated by a time dependent lookup
function from the years 2011 to 2016 in the system dynamics model, growing from 35%
to 68% of the population. (Arguably, there would be a good argument to simulate this
growth as a Bass model as well; however, we opted for a static function due to the good
availability of past information on this variable).
The values for innovator rate p and imitator rate q determine the speed of adopting the
mobile reporting service (i.e., the slope and skewness of the new adopters curve). Higher
values generally signify that a product is more contagious and q is generally greater
than p (usually for a factor greater than ten, (Mahajan et al.1995, p. 82). We reviewed
several diffusion studies of technology products, for example, on mobile phones in Spain
(Cronrath and Zock2007), and on an automatic telephone enrollment service at a New
Zealand university (Wright et al.1997). Plausibly, the closed user group of an enrollment
service exhibits a much faster adoption compared to the macroeconomic diffusion of
mobile phones, which explains why the p, q-parameter estimates of both studies are
similar despite being measured on different time scales (annual versus fortnightly). (Note
that normally innovation diffusion parameters behave approximately proportional to
the measurement scale, e.g. annual parameters are about twelve times the monthly
parameter estimates (Putsis1996). We assume that a mobile reporting service on urban
level ranks between those two cases and choose (monthly) parameters for our model which
are slightly less than half of the fortnightly parameter estimates of the New Zealand
example (p=0.02, q=0.2). While we acknowledge that these are rough estimations, we
consider them as sufficient for the purpose of scenario simulation since they only influence
the speed, but not the overall level of adoption and incoming reports. Implementation
time is set to 5 months for illustrative purposes (which equals the time from the workshop
to the implementation in the case example).
The most influential variables of the diffusion model, however, are the potential share
and the estimated reporting rate of service users since they determine the total amount
of reports that the municipality expects to receive from citizens. Ex-ante, these variables

10
5
4 IT Governance and Innovation Adoption in E-Government

Would you use a mobile How often would you use a mobile reporting service per year?
40
reporting service?
30
No /
undecided 20
45% Yes
10
55%
0
0 1 2 3 4 5 6-10 11-20 21-30 31-40 41-50 51-100 >100
Number of intended reports per year

Figure 4.16:Survey results

are hard to determine and therefore subject to uncertainty. Nevertheless, we draw on

technology adoption theory (Davis et al.1989) and assess the intended reporting rates as
a proxy. For this purpose, we conducted a survey with potential service users across all
age groups and professions. The survey was administered online and distributed via email
across the personal networks of the authors. It contained an introduction to the topic and,
amongst other items, a question on each of these parameters of diffusion. We effectively
received n=213 usable responses, 48% from male and 52% from female respondents.
Regarding occupation, 28% stated to be students, 37% working, 10% retired, and 11%
others. The distribution of the considered survey variables are depicted in Figure4.16.
According to the survey, 55% of mobile phone users would use a reporting service. The
median (and also the mode) of intended reports per year is 5. We view these unexpectedly
positive results with caution. While intended and actual technology adoption are clearly
related (with typical correlation values in the range of 0.20 - 0.30, (Davis et al.1989),
the literature provides numerous influences that moderate and distort this relationship.
Such influences may result from (1) factors that affect the accurate expression of current
intentions and (2) factors that affect the accurate prediction how these intentions will
change over time (Morwitz1997). Regarding the former, one bias in our setting may a
particularly result from the survey setting and the respondents tendency to give socially
desirable answers (Richman et al.1999). For example, studies in the field of health
behavior found that people translated their good intention into action in about 50%, i.e.
half of the cases (Sheeran2002). However, while the literature demonstrates a significant
correlation between intention and use, concrete rules of thumb how to translate intended
into expected usage frequencies are widely absent. To account for this uncertainty, we
assume two scenarios: an optimistic scenario with a 22.5% (half of the stated 55%)
potential user share and 2.5 (half of the mean 5) reports per user per year in line with
the survey data and the references from behavioral studies, as well as a pessimistic
scenario with a 5% potential user share and an average 1 report per user per year. For
comparison, the average traditional reporting rate in the city Saarbrcken with 180,000
inhabitants is 15 reports per month.
Regarding the deterioration model, internal documentation informed that the street
network size of Saarbrcken amounts to 665 km. We discretize a street km into 1,000
max spots per km. Streets are classified by priorities according to the intervals in which
they need to be checked in order to comply with the citys due diligence. We learned

10
6
 4.4 Municipal Benefits of Urban Sensing

that the city employs a number of 6 inspectors in the Department of Street Maintenance
to do this job fulltime. We calculate that per month, each of the inspectors controls an
average 40 km of the overall network. On his route, an inspector encounters in average
8 issues per km; that is, if streets were checked monthly, this occurrence would account
for approximately 3 new issues per km per month.
Finally, we instantiate the cost model. In our case setting, a municipal inspector incurs
about 50,000 Euro of total labor cost per year. Costs per report can then be calculated
as differential process costs (Drury2007). That is, we account only for the times for
handling and processing citizen reports from receipt to repair (excluding material cost and
overhead). Based on the process analysis, the workshop participants estimate that a
defect report via traditional channels takes an average 20 min and a mobile report 10
min (due to less citizen inquiries) for classifying and routing within the municipal
administration. Additionally, traditional reports are more likely to cause pre-inspections.
The average (conservatively) estimated surplus for pre-inspections is 10 min (i.e., about
2 hours for one out of twelve defects). For demonstrative purposes, implementation costs
are set to 10,000 EUR. The input variables and their values are summarized in Table4.13.

Scenario Simulation

We instantiate and simulate the model according to the two different scenarios (optimistic
and pessimistic scenario) and compare these with a baseline scenario, i.e. the case if no
mobile reporting service was implemented at t=5. For an additional comparison, we
run a pessimistic scenario where one inspector less is employed by the municipality
(i.e., number of inspectors=5). Four selected variables from the model are depicted in
Figure4.17.
Regarding the diffusion model, the results show that the number of new adopters
describes the typical bell-curved shape postulated by innovation diffusion theory (Rogers
1962). In month 17, i.e. 12 months after the service is taken in use, the curve peaks at
1072 new adopters for the optimistic, and 238 for the pessimistic scenario. The curve
is notably skewed whereas its long tail is superimposed by the overall mobile broadband
development and an ongoing increase in new mobile phone users.
In terms of deterioration, it is obvious that the model first needs to get into equilibrium,
since we start with a city free from defects (0 undetected spots in t0). As we can see from
the baseline scenario, this equilibrium is at an issue density of 0.82% within the urban
space. However, in the optimistic scenario, this measure for the deterioration goes down
after the adoption due on the large number of reports that the municipality receives and
converges to 0.03%. This is also reflected in the percentage of issues that are detected. In
the optimistic scenario, the detection rate increases to about 96% (in t=60). That means
that practically every uprising issue would be immediately reported by the collective
monitoring of the citizens. In the pessimistic scenario, the detection rate climbs up from
36.3% in the baseline scenario to about 42.3% in t=60, reducing the issue density to
0.7%.

10
7
4 IT Governance and Innovation Adoption in E-Government

Table 4.13:Simulation model parameters

Parameter Value Unita Sourceb
Opt. Pess.

City population 180,000 Citizen Case

Mobile broadband dev. 35% - 68% (2016) Dmnl ITU2010
Implementation time 5 Nonth Assumption
Potential user share 0.225 0.05 Dmnl Survey,Sheeran2002
Innovators rate (p) 0.02 1/mth Mahajan et al.1995;
Putsis1996;Wright
Imitators rate (q) 0.2 1/mth et al.1997

Street network size 665 Km Case

Max spots per km 1000 Spots/km Assumption
New issues per km 3 Problem/km/mth Case

Traditional reporting rate 15/180,000 Reports/citizen/mth Case

Reporting rate of users 2.5/12 1/12 Reports/citizen/mth Survey,Sheeran2002
Number of inspectors 6 Inspector Case
Walking km per inspector 40 Km/inspector/mth Case
Repair duration 2 Mth Assumption

Implementation costs 10,000 EUR Assumption

Ongoing costs 100 EUR/mth Assumption
Costs per mobile report 1.68 EUR/report Case
Costs per traditional report 4.91 EUR/report Case
Costs per inspector 50,000/12 EUR/mth Case
a
Dmnl = dimensionless; mth = month
b
Simulation results are invariant to variables for which assumptions have been made

With regard to process costs, we note that in the pessimistic scenario, after month 5
(and the initial investment peak for service implementation), monthly costs increase by
+3.4% compared to baseline scenario (25,934 vs. 25,073 EUR/month until t=60) due to
the increased efforts for processing reports from citizen. However, if we reduce fix cost
by setting the number of inspectors to 5, then monthly costs are even below baseline
by -12.9% (21,823 EUR/month). This is remarkable in view of the fact that the issue
detection rate in this scenario almost equals the baseline level (36.2% in t=60). In other
words, the results indicate that participatory sensing can provide a comparable level of
information about the urban infrastructure at lower cost than single street inspectors,
given there is a moderate participation by the citizens.
However, the chart on monthly costs also reveals one of the major risks of implementing
such service: If the level of citizen participation exceeds the expectations, such as in the
optimistic scenario, high costs and efforts accrue for processing the citizen reports without

10
8
 4.4 Municipal Benefits of Urban Sensing

new adopters issue density

2,000 0.01 pessimistic (-1 insp.)
baseline
1,500 0.0075
optimistic pessimistic
1,000 0.005
optimistic
500 0.0025
pessimistic
0 0
1 16 31 45 60 1 16 31 45 60
Time (Month) Time (Month)

issue detection rate monthly costs

1 40,000
optimistic
0.75 35,000 optimistic

0.5 pessimistic 30,000

basel ine
pessimistic (-1 in sp.) pessimistic
0.25 25,000
baseline pessimistic (-1 insp.)
0 20,000
1 16 31 45 60 1 16 31 45 60
Time (Month) Time (Month)

Figure 4.17:Simulation results (selected variables)

achieving any additional benefit in issue detection. Both scenarios have been presented
and discussed to the stakeholders in the Saarbrcken case.

Qualitative Analysis

The presentation of the simulation results, besides providing indicative evidence of the
potential benefits and risks, also fostered the discussion on further qualitative aspects
connected to the implementation of a mobile reporting service. These aspects related
to both expected benefits and potential concerns from which we can derive (functional
and non-functional) design requirements that need to be addressed during a system
implementation (Maciaszek2007). For example, the quantification of expected adopters
and incoming messages helped to balance between privacy requirements and performance
goals when deciding on the question of whether citizens would need to enter personal
data with every report or not. Qualitative aspects can be broadly classified into four
categories: user, process and organization, privacy and legal, as well as technical aspects.
We will briefly summarize each of these points, see Table4.14.
Workshop participants stressed the importance that the client application of a mobile
reporting service should possess high usability and enable easy interaction to outmatch
traditional channels of communication. An additional feature refers to the tracking of
reported issues, i.e. the mobile reporting service could communicate the status of issue
processing and communicate this to the citizens. However, this may also raise the citizens
expectation that issues are taken care of at a faster pace (which can not necessarily
be guaranteed). Therefore the issue status should only be presented to the sender.

10
9
4 IT Governance and Innovation Adoption in E-Government

Table 4.14:Qualitative aspects of introducing mobile reporting

User aspects Process and Privacy and legal Technical aspects
organization aspects
Easy interaction Increased report Needed citizen data Message filtering
Issue tracking quality Defined terms of Broadband
Improved Reduced use coverage
communication throughput times Deletion after Less media
Exceeded Accident closing disruptions
expectations prevention Transparency issue Integration with
Increased Analytics Photo third party ticketing system
satisfaction Routing across rights
municipalities Liability in
imminent danger

Altogether it was found that a mobile reporting service can improve the communication
between the citizen and municipal administration (i.e., to make it more efficient and
more effective), which is ultimately also expected to have a positive effect on citizen
satisfaction.
Regarding process and organization, the process analysis revealed that due to the stan-
dardized form of a mobile report (including location coordinates and potential photos)
the quality of citizens reports is expected to increase. This can have positive effects on
throughput times for complaint and defect handling, since additional loops for inquiries
to the citizen and for visiting the issue spots are avoided. Furthermore, some workshop
participants argued that a mobile reporting service could particularly help to detect
hazardous spots faster and thus to prevent from traffic accidents in the urban area. In
a more long-term view, the sensing data from citizens could also be used to perform
analytics and thus gain additional knowledge on problem areas and general trends. A
concern was raised regarding the need for routing reports to other municipalities in case
citizens use the service out of other municipal districts.
Several privacy-related and legal aspects need to be considered when implementing a
mobile reporting service. First, the data needed from citizens (e.g., for doing inquiries
and giving feedback) was a subject to discussion. However, there was a consensus that
data needs (especially for the location) should be stated clearly in the terms of use of
the service and deleted after closing of a case. Publishing open issues (including photos)
on the client application or a website and thus creating full transparency was viewed
rather critically not only for raising false citizen expectations but also to avoid abuse
and protect the rights of potential third parties (e.g., passers-by pictured on a photo).
Finally, dealing with urgent issues and the liability in imminent danger (e.g., due to a
missing street drain cover) was perceived as a concern inasmuch as in these cases the
administration would still prefer to receive a phone call.
One important technical aspect refers to implementing appropriate mechanism for
filtering redundant or invalid reports in order to prevent from information overflow (as
shown in the simulation of the optimistic scenario). Further concerns relate to a lack

11
0
 4.4 Municipal Benefits of Urban Sensing

new adopters reports from adopters

2000 250 optimistic
1750 pessimistic
200
1500 actual
1250 150 fitted
1000
750 100

500
50
250
0 0
Sep 11 Oct 11 Nov 11 Dec 11 Jan 12 Feb 12 Mar 12 Apr 12 Sep 11 Oct 11 Nov 11 Dec 11 Jan 12 Feb 12 Mar 12 Apr 12

Figure 4.18:Estimated and actual adoption curves

of broadband coverage in rural areas as well as indoors (e.g., in subways). Regarding

internal process flow and the support through IT systems, it was argued that a mobile
reporting service would help to overcome media disruptions at the start process (e.g.,
compared to phone and email). This could be achieved through implementing an entire
workflow management system which is integrated with the ticketing system used in the
complaints office. Altogether the discussion unfolds a broad set of aspects that are
correlated with the implementation of a mobile reporting service and thus need to be
considered for the design of such service.

Empirical Estimation and Case Learnings

As a direct consequence of the design workshop, the Saarbrcken mobile application

Saarbrcken App (including the mobile reporting service) has been implemented and
released in mid September 2011. Seven months later, we conducted post-implementation
interviews to inquire the major learnings from this project and to obtain first usage
figures. The rate of new adopters can be expressed by the monthly number of users who
download the Saarbrcken mobile application from the platform provider (i.e., from the
app store) and install it on their mobile device. The number of incoming reports via
this new medium is logged in the municipal Complaints Office, see Figure4.18. We use
these figures to estimate the actual parameters of diffusion and compare them to our
previous model assumptions.
The innovator, imitator rates as well as the number of potential adopters (i.e., pa-
rameters p, q and M of the Bass model) can be estimated by performing a quadratic
regression of the number of new adopters on the number of cumulated adopters in tn1
(Bass1969, p. 219). As fitted model parameters we obtain p=0.019, q=0.156 and
M =42,833 (R2=99%). The number of potential adopters also allows us to derive an
estimate for the potential user share, which would be around 63% according to the fitted
model (calculating with 68,400 users of a smart phone in Saarbrcken at ti, based on the
numbers fromITU(2010).
The actual and fitted curves of new adopters are depicted in Figure4.18. As we note,
the actual adoption rates are clearly above the simulated scenarios, just as the
calculated potential user share. We attribute this to the fact that the Saarbrcken

11
1
4 IT Governance and Innovation Adoption in E-Government

app is as well downloaded for a number of other purposes than mobile reporting and
potentially also by other people than local citizens, e.g. tourists. In our last interview,
the Head of IT Coordination tells that the most popular part of the application is actually
the information page on local movie show times. However, the actual adoption would
potentially be even higher if the application had been made available for several mobile
platforms. According to our interviewee, many citizens criticized that the application
developed for a single operating system discriminates other potential usersa fact that
he now considers a major weakness to be addressed in the near future.
In contrast to the application downloads, the incoming reports from citizens are at the
lower end of the expected range. A linear regression of the cumulated adopters on the
incoming reports shows that the empirical reporting rate of service users is 0.0066, i.e.
much below the expected rate (presumably due to the high number of downloads not
related to mobile reporting). Using this average reporting rate allows us to also define a
fitted reporting curve, which approximately coincides with the number of reports in the
pessimistic scenario, see Figure4.18.
Although this short time series regression only allows for a first estimation (Van den
Bulte and Lilien1997), this indicates that the benefits outlined for the pessimistic
scenario actually appear quite realistic. In our follow-up interview, the Head of the
Complaints office emphasizes that about 90% of incoming mobile reports are valid and
usable. Concerns related to application abuse could be largely dispelled since the mu-
nicipality decided to request senders to authenticate themselves with a name and phone
number (analogously to traditional phone and email complaints). The Head of Street
Maintenance affirms that the existence of a photo in citizens reports facilitates their
work, although their main source of information for urban street maintenance (i.e., the
work of the street inspectors) remains unchanged.

4.4.7 Conclusion

In this work we provided insights on the benefits that urban sensing entails for local
governments to simultaneously foster citizen participation and improve urban service
delivery. We chose the example of a mobile reporting service and developed a simulation
model that connects the adoption and diffusion of such sensing application with urban
deterioration, municipality-internal defect processing and cost-related variables. The
model has been validated based on a case study of a large German municipal admin-
istration over 1.5 years. The evaluation produced a detailed process analysis, a model
instantiation, a simulation of different scenarios, a discussion of further implementation-
related aspects, and post-implementation learnings as well as an estimation of the real
model parameters.
The results of this research are multifaceted. The detailed process analysis shows
that the proposed quantitative simulation model generally fits to the activities and
responsibilities in urban complaints and defects management. Thus the model captures
the main characteristics of a real-world case well. Furthermore, it reveals that two

11
2
 4.4 Municipal Benefits of Urban Sensing

potential levers of process improvement, namely reduced loops for citizen inquiries and
fewer pre-inspections, can arise from the increased information capabilities (i.e., photo
and location data) of mobile reports. In course of the model instantiation, we provide
reasonable parameters for the proposed simulation model and demonstrate how it can be
customized to the context of a specific case municipality. The subsequent simulation of
different scenarios suggests that the use of mobile reporting can have substantial
benefits in terms of available information (here: detected infrastructure issues) and does
not necessarily lead to greater cost compared to internal defect reporting procedures.
Nevertheless, the main purpose of the proposed simulation model is not to provide an
argument for introducing urban sensing services based on purely economic considerations
and neither it was the motivation in the given case. Conceptualized as an organization-
dominant artifact (Sein et al.2011), the major strength of this model is to facilitate a
comprehensive understanding of the benefits and concerns at the intersecting domain of
external M-Government services, which can help to improve communication and foster
decision-making in various E-Government implementation contexts. We demonstrated
this strength by summarizing the key qualitative aspects that emerged from the discus-
sion with municipal stakeholders and ultimately led to an implementation of the mobile
reporting service. These qualitative aspects related to user, processes and organization,
privacy and legal, as well as technical categories. Finally, our post-implementation
evaluation indicates that the scenario parameters are within the bounds of possibility.
Based on the case learnings we provided insights on why the observed deviations occur.

Methodological Evaluation

This research was oriented in an action design research (ADR) paradigm. In the following
we use the seven principles proposed bySein et al.(2011) to evaluate our approach
methodologically: 1) Our research has been inspired through practice inasmuch as the
idea to develop a simulation model was motivated by prior studies indicating that munic-
ipalities place less emphasis on external M-Government services (Winkler et al.2011b)
and by the municipalitys interest in evaluating these opportunities. (2) The proposed
artifact is especially ingrained through theory by drawing on innovation diffusion theory
(Rogers1962) andin quantitative termson theBass(1969) model for new product
adoption.
Reciprocal shaping took place in three major cycles (validation interviews, process
analysis and design workshop, post-implementation interviews) which allowed us to flesh
out the details of the model and determine its underlying parameters. Conversely, the
quantitative simulation results allowed municipal stakeholders to reason about design
requirements for example for handling the anticipated amount of citizen reports. This
way, prior uncertainties and a lack of cross-functional thinking (see problem statement)
have been addressed effectively. (4) During the design cycles, we observed several in-
fluences and effects of mutual learning among the project participants, not limited to
the insights between researchers and practitioners. For example, after our workshop
all participants confirmed that the conducted process analysis has provided them with

11
3
4 IT Governance and Innovation Adoption in E-Government

greater understanding about the different ends of the overall complaints and defects
management process. (5) Although we present the core artifact and its validation in
separate sections, the development and evaluation took place in a largely authentic and
concurrent manner. This may appear evident as with any new parameter assumption we
could instantly test how this change would affect the overall model behavior.
Finally, (6) guided emergence denotes that the created artifact reflects not only the
preliminary design created by the researchers (i.e., in our case the model itself) but
also its shaping by the use in an organization (Sein et al.2011). We believe that this
principle expresses the core finding of this research, in which the use of the model has
first led to a broader discussion of implementation-related issues and then ultimately
guided organizational decision-making. (7) We provide more generalized outcomes as a
contribution of this research in the following section.

Contribution

In its nature as a design-oriented approach, the main contribution of this research

is represented by the generated artifact, i.e. the simulation model for urban sensing
adoption and benefits. Although we instantiated and evaluated this model in a specific
case example (Saarbrcken), we argue that this model and our approach can be useful
on different levels of problem classes. The narrowest of these classes is certainly the
decision on implementing a similar mobile reporting service in a different municipal
context. Besides numerous qualitative considerations, we presented a validated model and
provided reasonable parameter estimations, so that government practitioners can simply
adjust the specific case variables in order to obtain realistic and usable results for their
respective city.
However, we put forward that our model and the demonstrated approach are also
applicable in a much wider class of problems related to external E-Government services
and organizational decision-making. That is, our model might as well be used to estimate
the use and impact of other urban sensing applications, such as environmental impact
assessment (Mun et al.2009), air quality monitoring (Aoki et al.2008), traffic mitigation
(Hull et al.2006), and noise control (Maisonneuve et al.2010). As we motivated at the
outset, many public agencies today struggle with introducing innovative M-Government
services and opening this new channel of interaction to their customers (i.e., the citizens)
partly due to their inability to moderate the internal decision process (Winkler et al.
2011b). The principal logic of the model (stating that citizen information adds up to the
level of organization-internal information and accounting for overlapping information)
is likely to remain unchanged across different urban sensing scenarios. Therefore, this
model can provide a valuable starting point to assess the potential impact of urban
sensing in various emerging application scenarios.
Finally, our findings also contribute to the broader literature on investment appraisal
and decision-making (e.g.,Ballantine and Stray1998;Farbey et al.1999;Irani et al.2002;
Irani2002;Prakash et al.2001;Stave2002;Ward et al.2008). Our case specifically illus-
trates how different types of models (more precisely, a conceptual SD model, a detailed

11
4
 4.4 Municipal Benefits of Urban Sensing

process model, and a simulated SD model) can support organizational decision processes
at different stages. Our case findings suggests that the soft modeling approaches (such
as the SD notation and the process model) facilitate a mutual understanding of the
problem domain among the stakeholders, while the hard modeling approaches (such as the
mathematical SD simulation) can be used to demonstrate different scenarios and thus
(intentionally) enable a discussion on appropriate design requirements. In this sense, this
research reconfirms the strengths of system dynamics as a valuable method to address
both hard and soft modeling problems in organizational decision-making, especially in a
public sector context (Stave2002).

Limitations and Future Research

Limitations of this study are primarily related to the content of the proposed model, the
estimated parameters as well as the selection of a single case site. Since any modeling
approach inherently represents a simplification of reality, one may find further potentially
relevant variables that did not enter into our model. Also, the time period considered
for the empirical estimation of the model parameters (especially for p, q, and M ) limits
the accuracy of the prognosis on the expected adoption and market sizes. However, as
the focus of the proposed SD model was on demonstrating the interrelationships and
dependencies of the presented variables, rather than delivering econometric parameter
estimations, we consider the parameters as a sufficient approximation for the purpose of
scenario simulation in our case of a decision-making context. As more empirical data
accumulates, future researchers may have the opportunity to derive more precise esti-
mations and extend the model by adding further potentially relevant variables (Van den
Bulte and Lilien1997).
The selection of the case site represents a limitation inasmuch as the case municipality had
already signaled interest in evaluating external M-Government services at the time the
contact was established. Thus, a positive bias may be present. Regardless, the
selected municipality represents as a critical case for our research since it was one of
the first municipalities in Germany to embrace urban sensing within its municipal M-
Government offering. Therefore we are confident that our research, including quantitative
and qualitative case findings, will represent a valuable source of information for other
municipalities that aim to advance in this field. One promising approach to achieve this,
as we learned from our case, is to work with external parties and engage in public-private
partnerships to operate such novel M-Government services. Future research should ex-
pand the focus of investigation to multiple case sites and investigate how implementation
decisions and operating models of such services vary across different municipal contexts.

4.4.8 Summary
Involving citizen in public affairs through the use of participatory sensing applications
is an emerging theme in mobile E-Government (M-Government). Prior work suggested
that agencies place more emphasis on internal than on external M-Government projects.

11
5
4 IT Governance and Innovation Adoption in E-Government

This chapter took an action design research perspective to provide insight into the often
overlooked potential of citizen-centric, external M-Government services. We considered
the scenario of a sensing application for reporting urban infrastructure issues to the
municipality and presented a System Dynamics model to estimate the diffusion, use,
and municipal impacts of such service. The model was validated based on the case of a
large German city, a dedicated survey and further data sources. The simulation results
indicate that, compared to internal information acquisition procedures, the use of urban
sensing can improve availability of environmental information at a comparable level of
cost. Furthermore, we discussed a number of aspects and learnings related to an urban
sensing implementation and provided an empirical estimation of the diffusion model. Our
results provide an impetus for researchers and government practitioners to reconsider the
benefits of urban sensing applications in E-Government endeavors.

11
6
5 Innovation Adoption and IT Governance
in Enterprise Information Systems

5.1 The Impact of Software as a Service on Information

Systems Authority

5.1.1 Preamble

This chapter has been initially published and presented at the 2011 International Confer-
ence on Information Systems (ICIS) in Shanghai, China (seeWinkler, Goebel, Benlian,
Bidault, and Gnther2011a).

5.1.2 Introduction

Software as a Service (SaaS) is changing the enterprise application landscape. In Central

Europe, currently about 40 percent of companies use SaaS in at least one area of business
applications (Benlian et al.2009), in the US even more than 75 percent (Landmark2011).
Market researchers predict that in 2011, 25 percent of new business software purchases
will be procured via SaaS and that this share will increase to approximately 34 percent
by 2014 (Landmark2011).
SaaS refers to applications delivered as services over the Internet and is commonly
seen as the highest layer of the Cloud Computing stack (Armbrust et al.2009). SaaS has
evolved from earlier web-based deployment models such as application service providing
(ASP) and is typically characterized by a highly scalable multi-tenant architecture on
the provider side which allows for corresponding economies scale (Cusumano2010).
Compared to traditional on-premises installations, SaaS usage consequently leads to
a reduction in a broad range of operational IT tasks in the user organization, such as
infrastructure management, application management and application development (Lee
et al.2003).
Despite such reductions, however, there are a number of decisions and activities taking
place at the interface between IT and the business, that clearly have to remain within
the adopting organization. In this context, the concept of IT governance has received
high attention as a major complementary resource to achieve IT performance through
aligning business and IT (Brown and Magill1994;Sambamurthy and Zmud1999;Weill
and Ross2004a;Gu et al.2008). An important part of IT governance is determined
by the distribution of decision authority between the business and the IT function.

117
5 Innovation Adoption and IT Governance in Enterprise Information Systems

It is commonly recognized that there are three major archetypes of such governance
arrangements: centralized, decentralized and federal modes (Brown and Magill1994;
Brown1997;Sambamurthy and Zmud1999). Thus, the overall question arises whether
the deployment of SaaS and the reduction of original IT tasks also leads to a shift in the
distribution of IS authority for SaaS applications?
Regarding this question, there exists some anecdotal evidence that with the advent
of SaaS the business side takes over stronger IS authority for original IT activities,
such as change management and application support (Yanosky2008), which suggests a
rather decentralized governance mode. Yet, other authors take the view, that the IT
departments themselves will transform into a central mediator and integrator of external
SaaS services in the long run (Gardner2008), thus adopting (or preserving) a rather
centralized SaaS governance mode.
While IS literature provides a considerable body of research that examines factors of
SaaS adoption (e.g.,Xin and Levina2008;Benlian et al.2009;Benlian and Hess2010b;
Benlian et al.2010), few works have been devoted to explore the impact of SaaS adoption
on internal IS governance and organization. Therefore, we aim to advance in this domain
by formulating the following three research questions:
RQ1: How do organizations allocate IS authority for SaaS applications?
RQ2: What are the factors that possibly influence this allocation?
RQ3: How should organizations allocate IS authority for SaaS applications?
To address these questions, this works combines a contingency perspective (Gresov
1989) with a qualitative approach. Based on rich interview data, we apply Grounded
Theory to extend an existing contingency model (Sambamurthy and Zmud1999) by a
set of factors for governance arrangements specifically on the application level. Further,
we examine the explanatory power of these factors in four comparative cases of SaaS
adoption.
Extrapolating the current trends in the SaaS market, our findings may help to better
understand the long-term impacts of SaaS adoption on user organizations, which is
of interest for an academic as well as a practitioner audience. The remainder of the
chapter is structured as follows: In the next section (5.1.3) we review the theoretical
foundations regarding SaaS adoption, IT governance and contingency theory. Then,
in Section5.1.4we explain our research approach. Section5.1.5describes the derived
contingency model, which is then applied in four comparative case studies (Section5.1.6).
Finally, Section5.1.7concludes by pointing out the limitations and future work.

5.1.3 Theoretical Foundations

In the following, we briefly review existing literature that informs the relationship be-
tween SaaS adoption and IS organization, and motivate why the concept of IT governance
and a contingency approach are particularly suitable to further explore this relationship.

118
 5.1 The Impact of Software as a Service on Information Systems Authority

SaaS Adoption and IS Organization

Software as a Service (SaaS) commonly refers to a delivery model where standard software
is hosted at an external provider and is used by multiple tenants over the Internet
(Armbrust et al.2009). It is contrary to traditional IT delivery where the software is
typically either operated on the companys own computing infrastructure (on-premises)
or hosted on dedicated instances at provider side (Lee et al.2003). Thus, the dis-
tinguishing criterion between the two delivery models is the multi-tenancy capability
of the software, rather than the question of internal or external delivery (Cusumano
2010). Economically, SaaS is often characterized by subscription-based pricing models,
as opposed to perpetual-use licensing for on-premises software (e.g,Choudhary2007a).
However, empirical works show that in practice licensing models for SaaS are almost as
diverse as for traditional software offerings (Lehmann et al.2010).
Several works investigate the decision of firms for adopting SaaS-based enterprise
software (e.g.,Xin and Levina2008;Benlian et al.2009;2010). These studies confirm
that, inter alia, lower application specificity and smaller firm size are among the main
drivers for SaaS deployment (Benlian et al.2009), while security concerns represent a
major barrier (Benlian et al.2009;Benlian and Hess2010b). Research on SaaS adoption
mostly draws on Transaction Cost Theory (TCT) and the Resource Based View (RBV)
as theoretical frameworks to explain the phenomenon (Xin and Levina2008;Benlian
et al.2009). Apart from SaaS adoption, these theoretical frameworks may also provide
appropriate constructs regarding the impact of SaaS on IS organization.
With respect to SaaS, TCT essentially states that companies will rather source software as
a service if application specificity and uncertainty regarding the outsourcing provider is
low. Application specificity in this context has been operationalized by other works as
degrees of modularity, customization and integration into the application landscape
(Benlian et al.2009). Thus, from an organizational standpoint, TCT implies that the
border of the IS organization using SaaS is defined by the overall degree to which SaaS-
based sourcing is performed.
According to the RBV, only resources that are not source of a competitive advantage
should be outsourced (Xin and Levina2008). Resources in this sense may be physical,
such as hardware or human capital, or intangible, such has processes, knowledge, or
software. Thus, with regard to IS organization, RBV informs that resources related to
SaaS-based sourcing remaining in-house, will be those that possess a high strategic value,
such as certain knowledge or capabilities. In this context, absorptive capacities have been
identified as an important organization-level construct explaining the capability of the
employees to assimilate and make use of external information to the advantage of the
firm (Cohen and Levinthal1990;Roberts et al.2011).
While providing important constructs to explain the border between the SaaS provider
and the consumer organization, the theoretical frameworks presented hardly shed light
on the internal borders, i.e. the arrangements between business and the internal IS
organization. In this context, IT governance theory has emerged as an important concept.

119
5 Innovation Adoption and IT Governance in Enterprise Information Systems

IT Governance and Contingency Theory

IT governance is commonly understood as a subset of corporate governance that has

evolved from IS strategy (Webb et al.2006). The need for IT governance is motivated
by Agency Theory (AT), which considers two or more actors (here: business and the
IT organization) interacting in an asymmetric relationship. The business (principal)
contracts the IT organization (agent) to implement and operate certain IT services.
AT provides that certain governance mechanisms (norms) can be established to avoid
opportunistic behavior of the agent and thus improve the effectiveness of this relationship.
One of these governance mechanisms refers to establishing the accountability frame-
[Link] and Zmud(1999) define this as the patterns of authority for key IT
activities. These can be allocated either in a centralized or decentralized manner. In a
corporate setting, centralization typically correlates with allocating decision rights to the
IT department, while decentralization refers to the lines of business. Several approaches
have been taken to operationalize IS authority (Brown and Magill1994;Sambamurthy
and Zmud1999;Weill and Ross2004a).Brown and Magill(1994) discuss the use of a 7-
point versus a 5-point scale for assessing the locus of decision [Link] and Ross
(2004a) present a more sophisticated operationalization by six governance patters, which
essentially combine the horizontal (i.e., business versus IT) with the vertical dimension
(i.e., executive versus employee level). In addition, they define five decision domains (IT
principles, IT architecture, IT infrastructure strategies, Business application needs, IT
investment and prioritization) that these governance patters can be assigned to.
The distribution of IT decision rights has been found to be a major complementary
organizational resource for achieving IT and firm performance (Weill and Ross2004a;
Gu et al.2008). Yet, research shows that there is no single IT governance mode that
fits all firms and domains of IT. Rather, the optimal governance mode is determined
by a rich set of factors (Brown and Magill1994;Sambamurthy and Zmud1999). For
example, in terms of IT infrastructure governance a more centralized mode should be
adopted by firms with cost or revenue synergies while a more decentralized governance
mode fits better to those firms that require local agility (Gu et al.2008). Therefore a
contingency approach appears also particularly useful to explain the mode of governance
regarding the domain of SaaS applications.
Contingency theories originally stem from the field of organization theory. They rep-
resent a class of management research which recognizes that an organization needs to fit
within its environment and contextFiedler and Others(1964). Thus, there are different
decisions that are optimal depending on the salient forces in each of these situations. The
theory of multiple contingencies for IT governance bySambamurthy and Zmud(1999)
is based on Gresovs (1989) conceptualization of fit and misfit in organizational design.
It identifies a number of organizational-level contingencies for the overall IT governance
mode which may also inform the question of the right mode of governance for SaaS
applications.

120
 5.1 The Impact of Software as a Service on Information Systems Authority

5.1.4 Research Methodology

Data Acquisition

To explore the factors that determine SaaS authority on an application level, we followed
a multiple case study approach and collected data from four companies using a SaaS.
The replication logic of this study combined elements of a theoretical replication (i.e.,
selecting cases expected to yield in different results) with those of a literal replication
strategy (i.e., cases producing similar results, seeYin2003, pp. 46-53).
First, to ensure comparability of the cases and rule out influences presented by the
type of application, we decided to focus on one popular SaaS application. We chose the
customer relationship management (CRM) solution from [Link] (SF), as CRM
is among the most popular application types for SaaS (Benlian et al.2009) and SF is one
of the market leaders in this segmentLandmark(2011). Then, potential case companies
were drawn from a SF customer references site and contacted formally. In this selection
we were seeking for differences in context variables such as industry and size to increase the
generalizability (Yin2003). Naturally, the selection also had to follow opportunistic criteria,
as not all companies were willing to disclose information on their case. As we had found
two cases for each manifestation of IS authority (allocation to the business, respectively to
the IT side), we regarded the number of cases as sufficient to deepen into the analysis of
contingent factors.
Interviews took place between May and July 2010 and followed a common semi-
structured guideline with four main sections: a) introduction, b) company context,
c) SaaS adoption, d) organizational impact (see Appendix5). To identify peculiarities
of SaaS usage, we replicated the questions in part c) and d) regarding an exemplary on-
premises solution. In those companies that allocate SF authority mainly on the
business side, we interviewed representatives from IT and the business. The overall 6
interviews amounted to more than 6 hours, respectively 67 pages of transcription which
were subsequently reviewed by the interviewees. We complemented the interview mate-
rial with company information from web resources and press clippings. The respective
characteristics of the companies, SF usage and interviewee roles are listed in Table5.1.

Data Analysis

We analyzed the interviews using grounded theory, a qualitative research method increas-
ingly employed in IS studies (Matavire and Brown2008). Grounded theory is contrary to
other research methods as it systematically seeks to develop theory rather than verifying
or testing it (Glaser and Strauss1967). We opted for Glaserian grounded theory as it
offers a more abstract conceptualization compared to the Straussian approach, which is
thought of to be more prescriptive (van Niekerk and Roode2009). Grounded theory
originally postulates to conduct research without a priori knowledge. However, given the
large theoretical body regarding our research, we took a more analytical approach which
allowed us to integrate previous theories during the coding process. This is in line with

121
5 Innovation Adoption and IT Governance in Enterprise Information Systems

Table 5.1:Case companies and key figures

Case A Case B Case C Case D
Education Manufacturing High-Tech Pharma
Revenue / 8 m EUR / 70 m EUR / 150 m EUR / 650 m EUR /
Employeesa 120 600 1,700 10,000
IT employeesa 10 7 40 400

No. of SF 30 60 150 860

CRM usersa
Implementation n/a 1 week 3.5 months 9 months
time
IT interviewee Head of IT (A) IT Application Head of Compe- Application
Manager SAP tence Center Manager (D1)
(B1) CRM (C)
Business - Sales Organizer - CRM Associate
interviewee and SF Key User Manager (D2)
(B2)
a
Figures rounded for anonymity

the most common pattern of using grounded theory in the IS field (Matavire and Brown
2008).
We first analyzed the interview material by performing an open coding using an
adapted Glaserian (1992) C-family paradigm with four main categories (context, contin-
gency, covariance and consequences). The resulting 165 codes with total 507 quotations
covered a calculatory 47% of the total interview material. As the core phenomenon,
the arrangements for SaaS application governance, was clear from the beginning of the
study, these text fragments could be directly assigned to the main categories in function
of their relationship with the core phenomenon, see Figure5.1.
As a second step in open coding, we aggregated these codes iteratively to appropriate
mid-level concepts. Codes describing similar activities conducted by different actors
(e.g., business or IT) were merged in this step. During this procedure, we also inten-
sively drew on previous theoretical foundations and related work. For example, some
categories describing the properties of the mentioned application emerged describing the
complexity to integrate the SaaS application, the ease to customize it and required
training needs during its introduction. Based on TCT and SaaS adoption literature,
these three themes were clustered as proxies for overall application specificity. In the
sense of a theoretical sampling (Glaser1992), we also compared the variables of SaaS
usage with those regarding typical on-premises applications.
Finally with regard to selective coding, we grouped these mid-level concepts around
the core phenomenon and analyzed the type of relationships. Concepts that did not
exhibit a logical link to the theme of application governance were discarded from the
analysis, for example codes related to costs for SaaS/On-premise (31), privacy and
security concerns (13) and properties of the SaaS provider (11). The resulting condensed

122
 5.1 The Impact of Software as a Service on Information Systems Authority

Context Contingencies Covariances Consequences

The location of Moderating variable or Variables which correlate Outcomes in
events or incidents influences that have a Phenomenon with the phenomenon, response to the
pertaining to a bearing on other variables, here application here in effect occurrences phenomenon, here
phenomenon, here here, properties related to governance of application authority variables related to
the organization the application itself success

Figure 5.1:Adapted coding paradigm from Glaser (1992)

model comprises 17 variables related to application governance, 5 of which refer to the

organization and 12 to the application level, see Figure5.2.

Comparative Case Studies

After deriving the contingency model, we dimensionalized each of the variables on appro-
priate three-point scales (i.e., high, medium, low) to prepare the case comparison. For
the variables of IS authority however, we chose a five-point scale as we perceive that the
complex arrangements between business and IT need to be represented with sufficient
detail (scale: business, business with IT involvement, Federal/Both, IT with business
involvement, IT). Based on the relevant text passages, three authors separately assessed the
17 variables for each of the four cases in a comparative manner. The resulting inter-
rater reliability of this assessment accounted for =0.35 measured by a weighted Cohens
kappa for ordinal measures, which can be regarded as a fair result (Landis and
Koch1977).
After the assessment, the authors discussed the contingency factors that were salient
for each case. According to contingency theory (Gresov1989;Sambamurthy and Zmud
1999), salient forces may either induce an IT- or business-oriented governance mode
while weaker forces are not perceived to be decisive for determining the governance
allocation. To validate our choice between salient and weaker forces, about half a year
after the interviews, we made a second inquiry to the interviewees asking for the precise
motivation for installing the respective governance mode. Finally, the case study findings
presented here have been cross-checked with our interviewees and passed minor revisions.

5.1.5 Results A Contingency Model for Application Governance

The derived model saturates the theoretical categories from Figure 1 with variables
that describe governance for SaaS applications. Figure 2 depicts the model with code
frequencies for each variable. In the following, we will illuminate these variables for each
category highlighting theoretical and practical references as well as relevant quotations.

123
5 Innovation Adoption and IT Governance in Enterprise Information Systems

Business IT department

Organization Application Properties IS Authority Success

(context) (contingencies) (covariances) (consequences)

Usage Decision authority

SaaS On-premise
Corporate Governance
Firm Size (10) Scope of usage (23) Change decisions (25)
Managerial autonomy (10) Financial decisions (28)
Specificity Architecture decisions (8) Governance
Strategic IS goals (4)
Ease of customization (28) Application effectiveness (42)
Gov ernance
Training needs (26) Task responsibility
Absorptive capacities
Integration complexity (16) Change implementation (15)
Line IT knowledge (1)
IT business knowledge (6) Initiation First level support (15)
Origin of the initiative (43) Second level support (25)

Figure 5.2:Contingency model for application governance (incl. code frequencies)

IS Authority

The variables defining IS authority are central to the phenomenon as they are expected to
correlate with the actual mode of governance (i.e., centralized, federal, or decentralized),
in other words they operationalize the concept of application governance. These variables
refer to activities which are typically not outsourced (Lee et al.2003), but remain
within the company, either on business or IT side. Based on the common distinction
between decision and execution, we divided this category into decision authorities and
task responsibilities.

Decision Authorities The major decision areas regarding SaaS refer to application
changes, financials and architecture, which is in line with the general domains for IT
decisions (Weill and Ross2004a) as well as with standard IT process models such as the
IT Infrastructure Library (TSO2011) .
Change decision authority refers to the responsibility of choosing which requirements
should be implemented by the SaaS application. Such requirements typically refer to
changes in data fields or workflows. ITIL suggests that such decisions are to be taken
by a Change Advisory Board which represents a central governance body within the
Change Management process (TSO2011). One of our interviewees (A) states that for on-
premise it is typically the IT role of the Technical System Owner [who] initiates the
change process, while regarding SaaS, in another case (D), there is an Operational
Committee [which] is made up of a representative from each of the business units and
the groups in the business unit to take such decisions, which shows the variance in this
variable.
Financial decision authority is a central governance domain in any IT organization
(Weill and Ross2004a). Financial decisions regarding an application, such as annual
spend for licenses and maintenance as well as singular investments and their prioritiza-
tion, are usually taken by the organizational unit that owns the application budget. ITIL
describes this decision domain in the Financial Management for IT Services process area

124
 5.1 The Impact of Software as a Service on Information Systems Authority

(TSO2011). One participant (case A) emphasizes that SaaS usage leads to a very high
cost transparency to the business while for in-house applications a number of hard-
and software, support-, upgrade, and backup-related costs accrue to IT costs which are
not visible for the organization. Thus, even if the application budget is owned by a
certain side of the organization (here IT), business may claim stronger decision authority
for financial decisions regarding SaaS, compared to traditional on-premises applications.
Architecture decision authority deals with technical choices including data, technology,
and applications to plot a path for migration and integration (Weill and Ross2004a).
Regarding a SaaS application, where most of data, technology and application logic
is provided externally, this decision domain largely refers to questions of internet con-
nectivity, security, and foremost integration with other applications. One interviewee
(B2) states that through the use of SaaS, fewer capabilities were required for technical
integration on business side compared to on-premise applications since we outsourced
this to the SaaS provider. In contrast, for the on-premise solution, in this case Enter-
prise Resource Planning, it is always IT to decide on integration and maintain the
interfaces. For comprehensiveness, it may be noted that ITIL disregards architecture
management as a dedicated process (TSO2011).

Task Responsibilities Apart from decision authorities, there are a number of key op-
erational activities regarding SaaS applications mainly referring to change management
and end-user support. In some cases the SaaS provider carries out these tasks. However,
the ultimate responsibility for these activities still needs to be allocated within the client
organization.
Change implementation responsibility refers to the operational handling of changes,
e.g. implementing customizations or new workflows and is defined in ITIL as part of the
ITIL Change Management process (TSO2011). Despite the multi-tenant design and an
inherently high degree of standardization, SaaS applications typically offer the possibility
to perform changes to the system in a web-based manner by authorized administrators
(Sun et al.2008). Thus, administrators may more likely be located in business. This is
at least the case for interviewee D2 who states: Im more part of the business, but I just
manage Salesforce and [. . . ] there are a lot of changes that we need to do. However,
more common appears to be an allocation to IT as it is usually the case for traditional
IT applications (Weill and Ross2004a). Interviewee B 2 states: I would be fine if I
could implement minor customizations in the [on-premise ERP] without doing this with
IT. This wouldnt be a disadvantage for the company. Finally, even if there are non-
trivial changes to the SaaS system, the example demonstrates that this can be handled
on the business side by contracting an external partner [. . . ] who worked without any
involvement of the IT department.
Another important domain of IS authority on task level is the responsibility for end-
user support (Brown and Magill1994). We further differentiate between first and second
level support. First level support refers to the responsibility to provide a central point
of contact for users that have an issue with the respective SaaS application. Typical

125
5 Innovation Adoption and IT Governance in Enterprise Information Systems

first level support activities include responding to user questions, or to grant access to
a specific functionality. In ITIL, this corresponds to the Incident Management, Request
Fulfillment and Access Management processes (TSO2011). Interviewee C states that
the user sends a ticket to IT support which is routed to the respective Competence
Center [. . . ] and then we implement that and close the ticket. In Case D there are
three levels of support, first level which is myself [interviewee D2] and the other system
administrators and theres the second level which is basically IT, thus in this case tickets
are first routed to support staff in business.
Second level support is understood as the responsibility for handling issues which
cannot be resolved at first level. For a SaaS application, second level requests typically
refer to any non-routine technical disturbance. This largely correlates with the Problem
Management process defined in ITIL (TSO2011). For SaaS, second level support may
involve the IT department for example when we have someone that cant access the
system and on the [providers] side everything looks absolutely fine. However, we
recognize that for SaaS, internal IT may often be bypassed and problems are directed to
the provider. Due to a higher degree of standardization, it is easier for the SaaS provider
to support heterogeneous client organizations. Interviewee D1 confirms that by stating
obviously with a SaaS solution, the vendor is much more aware of the integration aspect
than he might be with a regular ERP [i.e. on-premises] solution.

Contingency Factors

In order to identify and structure the possible contingencies for the mode of SaaS gover-
nance, we distinguish between factors that refer to the organizational context of the firm,
and such that relate to properties of the SaaS application itself. Since organization-level
influences can be largely derived from IT governance theory (Sambamurthy and Zmud
1999;Weill and Ross2004a), we did not focus on them during the interviews. This is
also reflected in the low numbers of related quotations in the interviews (see Figure5.2).

Organization-Level Contingencies Sambamurthy and Zmud(1999) define multiple

contingent influences on overall IT governance, out of which two categories are found
to be relevant for application governance as well: corporate governance andabsorptive
capacities. Regarding corporate governance, IT governance theory informs that firm size
influences the mode of IT governance (Sambamurthy and Zmud1999). Smaller firms
tend to establish centralized IT departments to better coordinate the interdependencies
between business functions. As firms grow, they develop more divisional structures, also
calling for a more decentralized IS organization to maintain responsiveness to different
lines-of-business (Ein-Dor and Segev1982). We propose that such influence on IS gov-
ernance will analogously apply to the narrower context of SaaS governance and include
firm size into our model, i.e. larger firms will be more likely to allocate decision authority
for SaaS applications to the business.
Following this argument, the general degree of managerial autonomy (Sambamurthy

126
 5.1 The Impact of Software as a Service on Information Systems Authority

and Zmud1999) of the firm is equally expected to explain the mode of SaaS governance.
Suppose there are two firms of the same size, then rather the one where business-line
managers have greater autonomy would also happen to decentralize IT governance for a
SaaS application. Several studies emphasize that such correlation between overall firm
governance and IT governance exists (e.g.,Ein-Dor and Segev1982;Brown and Magill
1994). Therefore we propose that managerial autonomy also influences in the allocation
of SaaS authority.
As a third contingency in this category, we aggregate strategic IS goals. Strategic
goals for the IS organization follow from business goals and have been conceptualized
in different ways, e.g. as a trichotomy of IS efficiency, IS comprehensiveness and IS
flexibility (Sabherwal and Chan2001).Weill and Ross(2004a) demonstrate that compa-
nies defining low costs and standardized business processes as their IT principles would
adopt a more centralized IS governance mode, while IS innovation- and growth-oriented
companies rather install decentralized decision rights for the IS organization. Thus, we
replicate this contingent influence to explain the mode of SaaS governance.
The second category on the organization level captures absorptive capacities. Absorp-
tive capacities play an important role from a resource based perspective. Generally,
they refer to the knowledge of a firms employees that facilitates assimilation of external
information and its application to desirable ends (Cohen and Levinthal1990). Related
to the question of SaaS governance, absorptive capacities may be particularly relevant
regarding the IT knowledge in business-line organizations (Brown and Magill1994;Bas-
sellier et al.2003). We propose that business managers and users who are knowledgeable
about IT will also be more likely to take over decisions and activities for managing a
SaaS application.
Conversely, absorptive capacities can also concern the business knowledge within the
IT department. We find this particular theme of absorptive capacities largely underrep-
resented in IT governance research (Roberts et al.2011). An exception may be seen in
the work byArmstrong and Sambamurthy(1999) who analyze the influence of the CIOs
business knowledge on IT assimilation. The more knowledgeable CIO and IT staff are
about business processes, the more responsive may the IT organization react to business
demands. Thus we propose that higher business knowledge within the IT department
may also serve as an explanatory factor for a more centralized SaaS governance.

Application-Level Contingencies Based on the interview data, we contend that fac-

tors on organization level are not sufficient to explain the phenomenon of application
governance for SaaS applications. Therefore, we extend the contingency theory on
organization-level (Sambamurthy and Zmud1999) by factors on application level and
validate these in the light of related literature. These factors refer to the properties of
the SaaS application itself and its implementation. They are divided in the categories
scope, application specificity, and initiative.
The scope of application usage may depend on the type of application (Benlian et al.
2009). The SaaS CRM application subject to the investigation is typically used by

127
5 Innovation Adoption and IT Governance in Enterprise Information Systems

employees in the marketing and sales departments, representing only a fraction of the
firm. Interviewee A reports the more departments use this application to manage data
about potential customers, the more user requests come up and the more complex [the
system] gets, therefore, a central process was required to control data quality. Thus,
if the application is used only by a few employees (e.g., for Engineering or Business
Intelligence), it is more likely that these will also be more involved in decisions and
activities. In contrast, a company-wide used SaaS application (e.g., for Office applications
and Collaboration) would possibly induce a more central mode of application governance.
The following category, application specificity, stems from Transaction Cost Theory
and has also been identified as one of the major (inverse) drivers of SaaS adoption
(Benlian et al.2009). We propose that specificity may also have a major influence on
the induced mode of governance and conceptualize it by the three variables, ease of cus-
tomization, training needs andintegration complexity. Regarding the first, applications
differ in their degree of customizability, i.e. the possibility to adapt user interface, data and
processes to specific needs (Sun et al.2008;Xin and Levina2008). All interviewed subjects
agree that SF CRM is very easy to handle and fairly easy to implement due to the fact
that everything is web-based (here participant C). In contrast, to change something on
the example on-premise application, interviewee B2 reports, he first need[s] to go to IT,
then it will be changed in the test system, then its carried over to quality assurance and
finally with the next release cycle it goes to the production system, which makes the process
clearly more time-consuming. Therefore we propose that a lower ease of customization
may also explain the degree of IT staff involvement in SaaS-related decisions and
activities.
The variable training needs reflects application specificity from the end user perspec-
tive. End user training has been commonly identified as a major success factor in software
roll-out projects (Nah et al.2001). Yet, with the increasing popularity of web-based
applications such as SaaS, which satisfy new standards of software ergonomics, trainings
needs might reduce. Interviewee B1 states that trainings for introducing the SaaS solution
went faster, easier and it was less manpower needed than in other projects. B2 adds
congruently from a business perspective that [he] took over trainings for the colleagues
of the foreign subsidiaries and did things, which were not in the job description of a
salesperson. Thus, we propose that lower training needs may be a motivation for stronger
business involvement in application-related responsibilities and vice versa.
As a third proxy for application specificity we identified the complexity to integrate
the SaaS application. There are different patterns for integrating a SaaS solution within the
enterprise application landscape, e.g. for integrating user accounts, user interface, data
and processes (Sun et al.2007). Interviewee C reports that his company integrates the
SaaS solution via a special buffering database which has been programmed by the [ERP]
team, and therefore they have quite some effort with this. In contrast to this, company
B uses no bi-directional, but a one-way interface from [ERP] to SF CRM so that internal
IT says they have almost nothing to do with running SF (interviewee B1). Thus, we
propose that the degree of integration complexity of the SaaS application may help to
explain how much central IT involvement is required in managing and deciding

128
 5.1 The Impact of Software as a Service on Information Systems Authority

about that application.

Finally, another category emerged from the interviews capturing the origin of the
initiative. IT projects such as introducing a SaaS application may be triggered from line
organization or the IT department (Weill and Ross2004a). Interviewee D states that
the business decided that they wanted to go with a hosted system contrary to case C
where the participant opposes I think the last step came from IT to say, lets do it. The
concept of project initiation is also used in other domains of IS literature, for example
in green IT adoption (Zarnekow et al.2010). As such motivation may also be sustained
towards the later allocation of IS authority, we propose the origin of the initiative as the
last factor to explain the mode of application governance.

Table 5.2:Contingency factors and proposed influence on application governance

Category Contingency Induced mode of IS authority Supporting literature
factor IT dept. Business org.
Organization level
Corporate [Link] size Small Large Ein-Dor and Segev1982;
governance Sambamurthy and Zmud
1999
[Link] Centralized Decentralized Ein-Dor and Segev1982;
autonomy Brown and Magill1994
[Link] IS Efficiency Growth Sabherwal and Chan2001;
goals Weill and Ross2004a
Absorptive [Link] IT Low High Brown and Magill1994;
capacities knowledge Bassellier et al.2003
[Link] business High Low Armstrong and
knowledge Sambamurthy1999
Application level
Scope [Link] Company-wide Single unit Benlian et al.2009
usage
Specificity [Link] High Low Sun et al.2007;Benlian
complexity et al.2009
[Link] of Low High Sun et al.2008;Xin and
customization Levina2008

[Link] High Low Nah et al.2001

needs
Initiative [Link] of IT dept. Business org. Weill and Ross2004a;
initiative Zarnekow et al.2010

Governance Effectiveness

Several consequences regarding the arrangements for application-level governance emerged

from the data. Based on literature we reduced these to a single success variable reflecting
different facets of the governance effectiveness for the SaaS application. The relatively
high number of quotations of this concept (42) reflects a broad conceptualization as well

129
5 Innovation Adoption and IT Governance in Enterprise Information Systems

as a high relevance of this variable.

One aspect of governance effectiveness refers to the alignment of business and IT
(Brown and Magill1994;Sabherwal and Chan2001). For example, Interviewee D 2 states
regarding the impact of SaaS: I thought that there was quite a gap between business
and IT rather than thinking we work the same company, [but] were all on the same
side. And now [after implementing SaaS] business and the IT work together rather than
it seems that one is holding over the other. Another aspect refers to empowerment
regarding the roles of business and IT, which results from an improved alignment. In the
words of D2: I think that [the usage of SaaS] has empowered IT even more. It has given
quite a positive outlook for IT from a business perspective. Independently from this, the
IT representative D1 confirms from his perspective: One clear thing is that by seeing
[. . . ] the SaaS solution, the business is more empowered to use and make sense with [the]
application in what obviously the application allows. I think the SaaS application has
empowered them to make the changes and give them more freedom. So apparently in
this case, the organization has found a governance arrangement that enables both sides,
business and IT, to work together in a mutually satisfactory way.
Nevertheless, we contend that there are also more moderate examples regarding gov-
ernance effectiveness. Regarding the interaction between business and IT, interviewee
B2 states soberly that the only contact we have with IT is the interface from the [ERP]
to SF. Congruently, the representative from IT (B1) does not perceive any improvement
in the relationship neither in the role of IT nor of business. Quite the contrary, the
interviewee negates any empowerment of the business side trough the usage of SaaS.
We conclude that the satisfaction with the governance arrangements regarding the SaaS
application can be regarded as rather moderate in this example.
Drawing on contingency theory, we propose that governance effectiveness describes a
positive outcome if there is a fit between the SaaS governance arrangements, the organi-
zational context and the properties of the SaaS application itself. Thus, it may be used as
a criterion variable to evaluate the current mode of governance. The contingency factors,
their operationalization and the possible influence on the mode of SaaS governance are
summarized in Table5.2.

5.1.6 Comparative Case Studies

In the following we will present the four cases of SaaS adoption and illustrate occurrences
of contingent influences as well as the resulting mode of SaaS governance for each case.
We focus on a brief description and outline the contingent forces that were salient in
each case.

Case A: Private Educational Institution

The first case is a privately owned educational institution which offers executive and
degree programs as well as consulting services to the industry. Central IT, a team of

130
 5.1 The Impact of Software as a Service on Information Systems Authority

10 people, understands itself as a service provider to the various business units that are
subordinate to the management team. The strategic goals of IT in this institution are
rather inclined to support growth and attract new students than to standardize business
processes or save costs, which was also one of the motivations to introduce a CRM system.
SF CRM has been introduced in 2006 largely by the initiative of the former Head of
IT. Since then it has been used by almost one fourth of staff in multiple business areas
mostly for managing communication to clients, students and sponsors. Technically, it is
operated as a stand-alone solution, yet a closer integration with the schools website is
currently being considered. Due to the inherent properties of SF as an ergonomic and
web-based product, ease of customization is considered to be quite higha fact, which
does not differ between the all the observed cases.
Regarding governance mode, SF fees are paid from IT budget, so that most of the
application-related decisions are also ultimately taken by IT. Ideas for changes and
customization of the system are often placed by business through a SF Business System
Owner role. User support as well as most of the changes and customizations to the
system are provided by the SF Technical System Owner, the counterpart IS role to the
Business System Owner. Altogether, application governance in case A is quite IT centric,
see Table5.3. However, the initial case observation does not provide any indication,
whether the chosen mode of SaaS governance is specifically effective or not.

Case B: Medium-sized Manufacturing Firm

Case company B is a medium-sized tool manufacturer with headquarters in Germany

and a worldwide sales and distribution network. During the past years of economic
crisis, IT budgets and workforce have been drastically cut, so that the remaining 7 IT
employees serve almost 100 times more internal clients. One of the major motivations
to install responsibility for the CRM system on business side was to disburden the IT
organization. This efficiency-orientation is an overall pattern in the IS strategy.
SF CRM is used and governed within the marketing and sales department. It is
integrated with the ERP system by a simple one-way interface, which is technically
managed by central IT, yet creates few problems. The Sales Organizer and (only) SF
Key User reports that his unit has been driving the CRM roll-out when it became clear
that ERP- and spreadsheet-based customer databases just did not serve their purpose
anymore. System customization, roll-out and training only took a minor time period.
Most SF-related decisions are now taken by the SF Key User himself, who also per-
forms most changes and customizations to the system, and contracts external support if
necessary. Yet, user issues regarding SF would first end up as usual incidents at the IT
help desk, before eventually being routed to the key user for second level processing, see
Table5.3. The case analysis does not convey the impression that the installed governance
mode has particularly contributed to the alignment or empowerment of business and IT
(see preceding section).

131
5 Innovation Adoption and IT Governance in Enterprise Information Systems

Table 5.3:Case comparison and contingent forces

Case A Case B Case C Case D
Organization-level contingenciesa
[Link] size Small Medium Medium Large

[Link] Rather Rather Rather Rather

autonomy centralized centralized centralized decentralized
[Link] IS Growth Efficiency Mixed Growth
goals
[Link] IT Medium Medium Medium High
knowledge
[Link] business Medium Low High Medium
knowledge
Application level contingenciesa
[Link] of Multiple Units Single units Single units Single units
usage
[Link] Low Medium High High (Integration
complexity (Stand-alone) (One-way) (Two-way) layer)
[Link] of High High High High
customization

[Link] Low Low High Med

needs
[Link] of IT dept. Business org. IT dept. Business org.
initiative

a
Salient contingencies in bold

Case C: Large High-Tech Company

The high-tech manufacturing and services company in case C operates in several Euro-
pean countries and has introduced SF CRM at first with 150 users in one pilot region.
For this purpose, a dedicated Competence Center CRM has been created under IT lead-
ership and hired high qualified staff, such as former consultants and university graduates,
i.e. people with a comparably high level of absorptive capacities and business process
knowledge.
Years before, several unsuccessful attempts to introduce CRM had been undertaken by
the business line, until finally the topic was taken care of by central IT. According to the
Head of the Competence Center CRM, there was a considerable effort during the roll-out
for training users and convincing them of the benefits of such system. The fact of having
capabilities on IT side to support the SF roll-out and consult the dispersed business
organization was certainly a determining factor to also allocate the responsibility for the
ongoing management of SF to this team.
The current governance structure for the CRM solution exhibits a comparably strong
decision authority of the IT department, i.e. the CRM Competence Center team. Oper-

132
 5.1 The Impact of Software as a Service on Information Systems Authority

ational activities such as first and second level support, as well as change implementation
are carried out by this team. Naturally, for main financial decisions, such as the budget
process, as well as for decisions on change requests, business management and users are
also strongly involved, see Table5.3. The Head of Competence Center CRM reports on
a clear appreciation of his team as a consultant and business enabler since implementing
this specific SaaS CRM solution and the according governance model.

Case D: Global Pharma Services Company

Case D is a globally operating firm with five strategic business units (SBUs) according
to different segments of services offered to the biotechnology and pharmaceutical sector.
The use of information systems and management of large amounts e.g. of clinical data
generally plays a crucial role in pharma services, therefore we classify IS goals as relatively
innovation- and growth-oriented compared to the other cases.
This growth-orientation was also one of the major motivations to make a CRM tool
available to the business. SF CRM has been introduced as a central solution displac- ing
two prior systems of different SBUs in order to increase transparency across SBUs and
leverage cross-sales potential. System introduction and customizing to the specific
company requirements took overall about nine months. To manage SF, former CRM
system administrators as well as qualified sales staff have been joined to form a new
CRM administration team within the business organization. Given the large corporate
setting, this relocation allowed to shift CRM authority from IT to business, while still
allocating it to a centralized function within the organization.
This team is in charge of most decisions regarding SF, performs changes to the system
and provides a single point of contact for SF user requests. If such a request cannot be
resolved at the first level (ca. 40% of the cases), it will be passed to the IT application
architecture team for second and third level processing. Furthermore, the SF solution is
highly integrated with a number of systems (e.g., financial, costing, and clinical systems)
via a special integration layer managed by IT. Consequently, architectural decisions and
changes to the integration layer require high involvement of central IT, see Table5.3.
Both, business and IT representative share the view that through the new governance
arrangements related to SaaS, overall IT alignment has improved and both parts of the
organization are empowered to perform their new tasks regarding the application.

Case Comparison and Discussion

In the following we will compare the four cases of SaaS adoption, starting with the largest
firm. We interpret the four cases by identifying patterns of reinforcing and conflicting
contingencies (Gresov1989;Sambamurthy and Zmud1999). Therefore we first take a
look at the salient contingencies, i.e. those that emerged as most determining forces
in each case. According to contingency theory, multiple reinforcing forces, i.e. those
influences which induce the same mode of application governance, would lead to a clear
shape of IS authority (either centralized to IT or decentralized to business). In contrast,

133
5 Innovation Adoption and IT Governance in Enterprise Information Systems

Table 5.4:Comparison of induced and present governance mode

Case A Case B Case C Case D
(Education) (Manufacturing) (High-Tech) (Pharma)
Induced gov. IT with business Federal IT Business
mode (conflicting)
Present gov. IT Business IT Business
mode
Governance Moderate Moderate High High
effectiveness

multiple conflicting contingencies, i.e. those that induce a different outcome, would
lead to a rather federal mode of application governance. Furthermore, we aggregate the
present mode of application governance to a single variable (business, federal, IT). It is
worth noting that there is clearly no case of a federal archetype for SaaS governance.
Finally, we may compare this as-is characteristics as well as the theoretically induced
mode of IT governance with the observed governance effectiveness for each case, see
Table 4. The results of this comparison are discussed in the following.
For Case D (pharma), the results exhibit multiple reinforcing salient contingencies
which clearly induce SaaS authority to be allocated in business. According to the pre-
sented contingency model, the comparably large firm size and rather growth-oriented IS
goals are both indicative for a business-oriented mode of application governance. The fact
that the company possesses comparably high IT knowledge within the SF administration
team, which has been achieved by staffing people who eventually had worked in IT
before, supports this perception. The present rather business-oriented governance mode
exhibits a strong fit to these contingencies. This may justify the relatively high perceived
governance effectiveness in this case.
For case C (high-tech), there are two reinforcing contingencies that militate in favor
of allocating SaaS authority to IT. These contingencies are given by the comparably
high business knowledge of IT staff and the fact that the initiative has been largely
driven by the IT department. We assume that these dominant forces have overruled
weaker influences such as managerial autonomy and integration complexity, thus leading
to a present allocation of SaaS authority which is quite IT-centric. The fit between
the present arrangement and the theoretically induced is in line with the satisfactory
governance effectiveness we experienced in this case. However, in the response to our
second inquiry, interviewee C stressed the current necessity to also develop stronger key
users in the regions and business units and thus to permit more business-orientation
even in the present governance mode.
In case B (manufacturing) we find a pair of apparently conflicting contingencies which
refer to the strategic IS focus on efficiency on the one hand and the strong initiative of
the business organization on the other. According to governance theory, such conflict
induces a federal mode of IS authority (Sambamurthy and Zmud1999), as a strategic
focus on IS efficiency is generally not compatible with a decentralized governance mode

134
 5.1 The Impact of Software as a Service on Information Systems Authority

(Weill and Ross2004a). However, we may also question the validity of the contingent
influence of strategic IS goals on the SaaS governance. As SaaS applications are per se
standardized and offer according economies of scale on the provider side, there are less
reasons to centralize their governance as opposed to on-premises software. A statement
by the business respresentative in case B underlines this: We have to consider the tight
personnel situation, existing positions have not been staffed [. . . ], so that inevitably I
have to take over things which are usually not part of my job description. Thus, here
our initial proposition regarding strategic IS goals fails to sufficiently explain the present
mode of governance. Quite the contrary, efficiency-orientation may rather correlate with
a more decentralized governance mode.
Case A (education) exhibits a single dominant contingency related to the origin of
the initiative from IT. Thus, it is worth interpreting the weaker influences, even if they
were not salient from the case. We found a number of contingencies primarily related to
application specificity as well as rather growth-oriented IT goals speaking for a governance
mode that involves more business into SaaS/SF-related decisions. However, we find
a quite IS centric present mode of governance. According to contingency theory, the
effectiveness of this arrangement could be increased if the fit to these contingencies
was improved. In fact, this finding was supported by the interviewees response to
our second inquiry. The Head of IT explains that the solution has been introduced
due to the strong engagement of the former Head of IT who did not foresee a proper
handover to business, a fact that he now considers as a weakness in the overall roll-
out procedure of the project. Meanwhile, he has divided application authority into
functional and technical responsibilities and passed large part of change decision and
implementation responsibility to the department for corporate communications, i.e. the
business organization.

5.1.7 Conclusion
In this chapter we investigated the impact of using Software as a Service (SaaS) on IT
governance, defined as the locus of authority for key decisions and task responsibilities
related to a SaaS application. Therefore we combined a contingency perspective with a
multiple-case approach to explore the factors which influence the mode of SaaS gover-
nance. Based on the analysis of rich qualitative data and the review of existing theory, we
derived a contingency model for governance arrangements on application level. Finally,
we examined the explanatory power of this model in four comparative cases of SaaS
adoption.

Findings and Contribution

The results show that for two companies (B, D) there are dominant or even reinforcing
contingencies which induce the SaaS application to be governed by business side, while
in the other two cases (A, C) there are contingencies that promote a rather IS-centric
SaaS governance. Furthermore, we were able to provide narrative evidence that the

135
5 Innovation Adoption and IT Governance in Enterprise Information Systems

prevailing governance mode for one case (A) resulted in a certain misfit of SaaS authority
which had subsequently been rectified. Getting back to our research question how firms
(should) allocate authority for SaaS applications, we contend that there is no appro-
priate governance form for all cases. The mode of SaaS governance rather depends on
certain organizational and application-specific contingencies. In this context, absorptive
capacities and the origin of the initiative have emerged as most determining forces in the
four given cases, while a contingent influence of strategic IS goals could not be supported.
Aligning the mode of governance with organization- and application-related contingencies
is expected to positively affect success variables such as governance effectiveness.
Our contribution is threefold. First, we propose a number of factors that extend
existing contingency theories (Sambamurthy and Zmud1999) towards application-level
governance. This appears particularly useful, as in todays networked organizations there
is no rationale to establish the same governance mode across all business applications
(Marks2008). Second, we illustrate that the often claimed correlation of IS efficiency
and a centralized authority does not necessarily hold for SaaS-based applications. This
underlines the call for more research on the governance of cloud-based solutions. Finally,
we propose a contingency model that can be used by both academics and practitioners
to better understand the reasons why a certain mode of governance is chosen for a given
SaaS application. We thereby hope to provide impetus and give directions for further
research.

Limitations and Future Work

Due to the qualitative case study approach taken in this chapter, the obtained results
may possess limited generalizability. However, the applied method allowed us to reveal
important details about SaaS governance in the considered cases which was required
from a theory generation and adaptation standpoint. Further, to ensure comparability
of the cases, we only considered SaaS users of the application Salesforce CRM. Not
surprisingly, this resulted in a low variance in some of the observed variables such
as application usage and specificity. Also, we acknowledge a lack of the longitudinal
dimension of the data. Focusing on a single point of time may limit the explanatory
power of the derived contingency factors, as the narrative evidence regarding case A
suggests. SaaS governance is a dynamic construct, so that future studies should consider
a longer timeframe to address this issue. Methodologically, the contingency approach
assumes exogenous factors and thus largely neglects possible interaction between gover-
nance arrangements and the organization. Conceivably, the chosen mode of governance
may reversely affect certain organizational level variables such as absorptive capacities.
Thus, other approaches such as process theories might seem more appropriate to capture
these temporal interaction effects. Finally, given the qualitative nature of the criterion-
variable we were unable make any safe assertion about the degree to which a certain
misfit is good or bad (Umanath2003). In a future work, we plan to conduct a large
sample investigation with firms using different types of SaaS applications to validate
our propositions and measure the effect of SaaS governance fit on appropriate criterion

136
 5.1 The Impact of Software as a Service on Information Systems Authority

variables.

5.1.8 Summary
For some business applications, using Software as a Service (SaaS) is becoming increas-
ingly popular. However, it is largely unknown how SaaS adoption affects the arrange-
ments between business and IT departments. In this chapter, we took a contingency
approach to investigate how firms allocate authority for SaaS applications. Based on semi-
structured interviews with business and IT representatives of companies that have
adopted the wide-spread SaaS solution Salesforce CRM, we extended existing contingency
theory to propose a set of factors for governance arrangements on the application level.
These factors are used in a comparative case study of 4 cases of SaaS adoption to explain
why application authority is allocated either to the business or the IT side. The results
suggest that in most cases there exist dominant and reinforcing contingencies determining a
definite mode of SaaS governance.

137
5 Innovation Adoption and IT Governance in Enterprise Information Systems

5.2 Comparing Authority for On-Premises Applications and

Software as a Service
5.2.1 Preamble
At the time of publishing this dissertation, this chapter represents a working paper which
is currently under review for a major IS journal, see also (Winkler and Brown2013a).

5.2.2 Introduction
Sharing information technology (IT) authority between business and IT departments
and thus balancing the interests of different stakeholder groups is an ongoing issue in
corporate management. Business units often claim higher authority over their share of
the application landscape in order to fulfill their specific business needs, while central IT
units typically need to cut down such initiatives and enforce corporate-wide standards
in order to cope with scarce resources and successfully translate the corporate mission
of economies of scale. The dilemma of balancing between economies of scale versus
responsiveness has been widely recognized as a key challenge of IT governance (e.g.,
Tavakolian1989;Brown and Magill1994;Brown1997;Sambamurthy and Zmud1999;
Weill and Ross2004a), which is according to some authors the single most important
predictor of how much value firms derive form IT investments (Weill and Ross2004a).
The allocation of IT decision rights has been investigated at the level of the overall
IT function (Tavakolian1989;Brown and Magill1994;Brown1997;Sambamurthy and
Zmud1999;Peterson et al.2000;Weill and Ross2004a), as well as for certain subdomains
of IT governance, such as data governance (Khatri and Brown2010), infrastructure
governance (Gu et al.2008;Xue et al.2011) and governance in system development
projects (Tiwana2009). Few studies, however, have investigated internal governance
arrangements between business and IT departments in an outsourcing context, e.g. when
a third party is taking over IT service delivery for a specific business application.
One of the major trend shaping todays software industry is the provisioning of en-
terprise applications and application components to multiple user organizations as a
service over the Internet, i.e. Software as a service (SaaS). SaaS has evolved from earlier
service delivery models and is now commonly regarded as the highest layer and most
widely adopted form of cloud computing (Armbrust et al.2009;Cusumano2010). In
contrast to traditional, on-premises installed software, SaaS applications are character-
ized by subscription-based pricing models (Choudhary2007b), highly scalable multi-
tenant architectures (Chong and Carraro2006;Cusumano2010), and web-based mass
customization (Chong and Carraro2006;Xin and Levina2008). Previous studies on the
SaaS model have examined adoption on the client side (Xin and Levina2008;Benlian
et al.2009;Benlian and Hess2010b), as well as governance issues between client and the
provider, such as pricing (Choudhary2007a), coordinationDemirkan et al.(2010) and
contractual arrangements (Susarla et al.2009;2010).
We argue that SaaS is also likely to defy the conventional rationales behind internal

138
 5.2 Comparing Authority for On-Premises Applications and SaaS

governance arrangements on the client side. That is, SaaS facilitates the deployment of
new enterprise applications directly by business units. First, the reduction of up- front
capital investment and implementation efforts for SaaS makes it easier for business units to
fund and deploy required IT systems and thus circumvent centralized investment processes
(Xin and Levina2008;Xue et al.2008). Second, through considering scala- bility
requirements inherently in the software architecture, user-based customizations for SaaS
can be performed largely without coding, either directly through the web interface or
by deploying modular components (Sun et al.2008;Xin and Levina2008).
This may drastically reduce the need of specialized IT capabilities for managing and
maintaining these solutions on user side. Moreover, many SaaS solutions typically apply
state-of-the-art techniquesalso referred to as rich internet applicationsin their web
frontends, making them capable to replace an increasing number of centrally managed
desktop solutions, such as collaboration and productivity applicationsFraternali et al.
(2010);Cusumano(2010). Some radical proponents even view SaaS as a manifestation
of computing as a utility and thus prognosticate a general shift of IT authority from
IT units to business departmentsYanosky(2008) and consequently the disappearance
of IT as a dedicated corporate function (Carr2005). Some researchers, in contrast,
provide exploratory case evidence that companies allocate the horizontal authority, i.e. the
distribution of application related decision rights between business and IT units, for SaaS
applications in different, but clearly distinguishable ways, dependent on certain
organizational as well as application-specific contingencies by (Winkler et al.2011a).
In this work, we build on the contingency model proposed byWinkler et al.(2011a)
and adopt a multi-theoretic approach that combines agency theoretic, transaction cost
theoretic, and knowledge based lenses to explain the horizontal distribution of IT au-
thority on the application level. A research model is derived that conceptualizes relevant
constructs for each of these theoretical perspectives. Thereby we aim to address the
following research questions:
RQ1: How can we explain the horizontal allocation of application authority in general?
RQ2: (How) does the allocation of authority differ between SaaS and on-premises ap-
plications?
The research model was validated and tested in a survey with 207 large-sized com-
panies, 76 of which provided information about an SaaS and 131 regarding an on-
premise application in use. Our findings support the initial idea that application au-
thority depends on certain organizational as well as application-specific contingencies.
Furthermore, authority for SaaS-based applications is more frequently allocated to the
business side and largely independent from overall IT governance arrangements. The
chapters central contribution is to demonstrate that it is useful to view IT governance
as a symbiosis of modular arrangements on the application level in order to capture
the changes induced by new technological architectures such as SaaS. For example,
we demonstrate that for SaaS applications the business knowledge of IT staff plays a
significantly more important role to retain authority in the IT department than for on-
premises solutions. Our findings provide important implications from a theoretical and

139
5 Innovation Adoption and IT Governance in Enterprise Information Systems

a practical standpoint.
The remainder of the chapter proceeds as follows. The next section describes the
three theoretical lenses (5.2.3), followed by the research model and hypotheses (5.2.4),
the research methodology (5.2.5), the statistical analysis (5.2.6), and a discussion of the
results including theoretical and practical implications (5.2.7).

5.2.3 Three Theoretical Lenses on Application Authority

Agency-Theoretic View

Agency theory (AT) is concerned with problems that arise in asymmetric relationships
where one party (the principal) delegates work to another (the agent) (Eisenhardt1989a).
Based on the behavioral assumptions of new institutional economics, i.e. self-interest,
bounded rationality, and risk aversion (Williamson2000), problems particularly occur
when principal and agent possess conflicting goals, incomplete information and different
risk preferences. Given a lack of information of the principal about the detailed activities
of the agent, such incomplete contractual relationships can lead to a moral hazard
of the agent to behave opportunistically, and thus cause suboptimal overall outcomes
(Eisenhardt1989a).
In terms of enterprise IT governance and the horizontal allocation of application
authority, the business units of an organization can be understood as the principal
who is engaged in a permanent contractual relationship with the IT unit (Posthumus
and Solms2008), that is, business units delegate IT delivery to IT units which are
more specialized to perform this task. AT posits that certain mechanisms are required
to prevent IT units from moral hazard, or, to put it more pragmatically, to deliver
the expected value to business (Peterson et al.2000;Weill and Ross2004a). When
IT authority is exclusively allocated to IT, there is a substantial threat that busines
interests are not sufficiently considered, leading inter alia to a lack of business alignment
(De Haes and Van Grembergen2009;Tiwana and Konsynski2010). Conversely, if IT
authority is allocated to business units, considerations from a technical (e.g., integration
with other IT solutions) as well as from an enterprise-wide perspective (e.g., the use of
standards) are not sufficiently considered. Therefore, appropriate accountabilty schemes
need to be established as a mechanism to overcome the problems posed in assymmetric
relationships.
Modular systems theory suggests that the technological as well as organizational ar-
chitecture of the IT function in the firm can be viewed as a bundle of more or less
strongly interacting subsystems (Schilling2000). Thus, there are likely to be business
and IT actors (i.e., certain business departments and IT teams) who act as principals and
agents in managing each respective enterprise application. AT suggests that application
authority needs to be well defined and shared between these actors.

140
 5.2 Comparing Authority for On-Premises Applications and SaaS

Transaction Cost Theoretic View

IS scholars have primarily drawn on transaction cost theory (TCT) to explain contractual
arrangements and other governance phenomena in outsourcing relationships (seeNagpal
2004), including in an SaaS context (Susarla et al.2009). TCT puts transactions, defined
as the ultimate unit of economic activity (Commons1931), and the cost for planning,
adapting, and monitoring this activity, into the focus of interest to explain the boundary
of an organization (or an organizational subunit) (Williamson1985). Analogously, we
apply this view to internal IT governance arrangements: Business departments enter into
a contract with IT departments in order to handle application transactions efficiently
i.e. to economize on costs for planning, adapting, and monitoring application operation
or decide to coordinate application operations hierarchically.
TCT posits two important variables that determine transaction costs for managing
applications. First, if the application is very specific to a business department and its
processes, the transaction costs for contracting out to the IT department increase, e.g.
through higher efforts for communicating requirements, requesting system changes as
well as for monitoring the IT units application operations (Susarla et al.2010). Second,
the costs for planning, adapting and monitoring application operation are subject to
economies of scale with an increasing frequency of use. Thus, the relevant question
from a TCT-perspective is which coordination form best leverages these relative cost
advantages. Regarding intraorganizational arrangements, corporate units are inherently
concerned with aggregating delivery for other business functions. Thus, in order to
economize on transaction cost, frequently used applications are more efficiently managed
by the corporate IT function. Altogether, TCT provides a framework for explaining
horizontal allocations of authority by the specificity and usage characteristics of the
respective application.

Knowledge based View

The knowledge-based theory (KBT) of the firm has evolved from the resource-based view
and focuses on knowledge as the most unique and strategically important resource of the
firm. (Grant1996). The primary purpose of the firm is to integrate the specialized
knowledge, which resides in the employees and is costly to transfer (Jensen and Meckling
1998). According to this view, centralized decision making is feasible, when knowledge
resides in a single point in the organization, otherwise, bi- and multilateral decision
making would be more suitable (Grant1996).
For managing business applications, two complementing types of knowledge are most
relevant: IT-related knowledge and knowledge of the business domain (Tiwana2009).
Due to the logic of specialized resources, the former, i.e. knowledge about technologies,
applications and IT management, typically resides in the IT department, while the latter
is inherently associated with business departments. Under this premise, KBT informs
that joint decision making is required to facilitate effective decision making and, in case
there is an overweight of any complementary knowledge on either side, this party is likely

141
5 Innovation Adoption and IT Governance in Enterprise Information Systems

to have higher authority in for application-related issues. Thus, the knowledge-based

view posits that horizontal authority for business applications is likely to be allocated
towards the department where most pertinent knowledge is present (Jensen and Meckling
1998).

5.2.4 Research Model and Hypotheses

The research model depicted in Figure5.3follows from the agency-, transaction cost-
and knowledge based theoretic considerations that help to explain horizontal applica-
tion authority. In the following we conceptualize its constructs and develop underlying
hypotheses regarding the potential difference between SaaS and on-premises applications.

Agency Theory Transaction Cost Theory Knowledge Based View

Origin of Scope of Business IT IT Business

IT Governance Specificity
Initiative Use Knowledge Knowledge

H1 H2 H3 H4 H5 H6

Delivery Model
A Hx
(SaaS/On-premises)
Hxa Hxb

Application
Decision Task
Governance
Authority Responsibility
(Business/IT)

Figure 5.3:Research model

Application Governance

A business applicatione.g. a system for enterprise resource planning (ERP), cus-

tomer relationship management (CRM), business intelligence (BI), or collaboration and
communicationcan be regarded as an IT artifact that is a subset of the entire techno-
logical architecture of the IT function, runs on a corresponding infrastructure, and helps
to deliver certain business functionalities (OBrien2005).
A business applicatione.g., a system for enterprise resource planning (ERP), cus-
tomer relationship management (CRM), business intelligence (BI), or collaboration and
communicationcan be regarded as an IT artifact that is a subset of the entire techno-
logical architecture of the IT function, runs on a corresponding infrastructure, and helps
to deliver certain business functionalities (Fowler2003;Orlikowski and Iacono2001;
Schilling2000). Decision making regarding the overall IT function is regarded as core
element of IT governance (Brown1997;Sambamurthy and Zmud1999;Weill and Ross
2004a). Prior researchers commonly emphasize the variability in the allocation of IT
decision rights, which may be decentralized to business units or centralized to corporate
IT units (Brown and Magill1994;Brown1997;Sambamurthy and Zmud1999). Some

142
 5.2 Comparing Authority for On-Premises Applications and SaaS

later works also use a more sophisticated conceptualization by six governance patterns,
which combine the horizontal (i.e., business versus IT) with the vertical distribution
(i.e., executive versus employee level) (Weill and Ross2004a). In this study, we focus
on the horizontal dimension and the relationship between business and IT units, instead
of stressing the common dichotomy of centralized and decentralized units, in order to
delineate the concept of governance from questions of organizational structure.1
Depending on the domain of IT governance, there are different classes of decision
rights. Prior research regarding governance in system development projects suggests to
distiguish between two broad classes: decision control and decision management rights
(Tiwana2009). Applied to the context of application management, this implies that
there are:
(1) Decision control rights which approve application related decisions, e.g. decisions
on which functional changes or enhancements are to be implemented according to the
business needs, decisions on investment for licenses and application maintenance as
well as on general architectural choices. We may term this dimension of application
governance as the general decision authority over an application.
(2) Decision management rights that implement application related decisions, e.g.
decisions on how to implement functional or technical changes, decisions on how to
support users and decisions how to solve application related problems. As these decision
rights are directly associated with performing application related activities we may refer
to this dimension as task responsibility.
The separation between decision control and decision management has been primarily
motivated by agency theoretic considerations (Fama and Jensen1983). That is, appli-
cation authority and task responsibility should not be allocated to the same actor in
order to establish effective control and prevent efficiency losses (Tiwana2009). In a
wider sense, the practitioner literature and common work practices widely recognize this
principle as a segregation of duties. However, decision authorities and task respon-
sibilitiesare are typically shared between business and IT roles: Decision authority for
application investment could be allocated to a business unit head, decision authority
for application changes to a business process owner, decision authority for architectural
choices to the CIO, task responsibility for functional changes to a business analyst, and
task responsibility for technical changes to an application expert. For this reason, some
researchers have conceptualized IT governance as a continuum with greater or lesser
ownership rather than a dichotomous allocation (Brown and Magill1994;Gu et al.2008;
Tiwana2009). According to this, in the aforementioned example we might find a greater
decision authority from business side, and equal sharing of task responsibility between
business and IT roles.
1
While the dimensions decentralization/centralization and business/IT are certainly correlated, the
former denotation may lack precision, since centralization may refer to various dimensions, e.g. lines
of business, business functions, or geographies. Reversely, decisions may be allocated to IT and still be
organizationally decentralized, for example if the IT organization itself is very decentralized (Brown
1999). For similar motivations, other authors have as well suggested to emancipate the concept of IT
governance from organization structure (Sambamurthy and Zmud2000).

143
5 Innovation Adoption and IT Governance in Enterprise Information Systems

Further, when an application is outsourcedas it is the case for SaaSlarge parts

of task responsibility are delegated to an external party. For example, a task structure
proposed for SaaS CRM services consists of seven tasks, out of which only two (mapping
SaaS to business requirements, streamlining customer-facing activities) are performed
by the user organization (Susarla et al.2010). However, the client organization still
needs to retain, or even build up, certain capabilities to coordinate and control the
activities of the external provider (Willcocks et al.2006). Thus, from a agency theoretic
perspective, we may say that outsourcing converts internal decision management rights
into according decision control rights over external tasks. Nevertheless, the ultimate
locus of responsibility for the providers tasks still remains within the client organization.
Therefore, in our conceptualization we include the notion of controlling external tasks
within the concept of task responsibility and limit this dimension to the continuum
between business and (internal) IT departments.
Especially for SaaS-based sourcing, where costs are more transparent compared to
complex IT investment and accounting schemes (Choudhary2007a;Xue et al.2008),
business units may claim more decision authority over application related issues. As
stated earlier, some tasks, which would traditionally require technical IT capabilities,
such as performing small changes to an application, e.g. implementing a new workflow,
might shift to the responsibility of business users, since SaaS solution are inherently de-
signed for user based customization (Sun et al.2008;Xin and Levina2008). Accordingly,
we pose our principal differential hypothesis that motivates research:
H0: In a sample of pairs of organizations and business applications, the horizontal
allocation of decision authority (H0a) and task responsibility (H0b) will differ between
applications with SaaS and on-premises delivery models.

IT Governance (organization level)

Having stated that applications are a subset of the technical IT architecture, and ap-
plication governance arrangements are a subset of overall governance arrangements,
these arrangements on application level are likely to be dependent on the overall IT
governance framework (Schilling2000;Tiwana and Konsynski2010). Congruently with
the application governance dimensions, we focus on the horizontal allocation of overall
application related decision rights between business and IT units and delieate from the
centralizatoin/decentralization dichotomy. Different domains of overall IT decision rights
have been proposed by the literature, for example a widely recognized framework of five
IT domains: IT principles, IT infrastructure strategies, business application needs, IT
investment and prioritization, IT architecture (Weill and Ross2004a). Arguably, the last
three of these categories are more likely to influence individual governance arrangements
on the application level, while the other two (IT principles and IT infrastructure) are
only indirectly linked with application-related decision making. For example, IT infras-
tructure is frequently governed in a centralized mode while application-related decision
making is more shared between business and (decentralized) IT unitsa pattern that has
also been recognized as a federal (or hybrid) archetype (Brown1999;1997;Sambamurthy

144
 5.2 Comparing Authority for On-Premises Applications and SaaS

and Zmud1999). Thus, for the purpose of this study we define IT governance more
narrowly as the horizontal distribution of IT decision rights on the organization level
between business and IT units regarding business application needs, IT investment, and IT
architecture. We may hypothesize
H1: Overall IT governance is positively associated with application level decision au-
thority (H1a) and task responsibility (H1b).
In this research we conjecture that SaaS and related delivery models, which exhibit
scalable outsourcing possibilities, are likely to unleash a new dynamism in IT organization
and governance phenomena (Peak and Azadmanesh1997;Winkler et al.2011a). Thus,
given that the use of SaaS applications in business organizations today still represents a
small fraction of the overall technological architecture (i.e., the entire application land-
scape), we expect the arrangements for governing SaaS applications to deviate notably
from overall governance arrangements. That is, overall IT governance arrangements are
likely to influence traditional on-premise application governance, however to a lesser
extent SaaS application governance. Accordingly we hypothesize
H1: The association of overall IT governance with application governance (H1a/b)
is weaker for SaaS than for on-premises applications, i.e. the type of delivery model
(SaaS/On-premise) moderates the relationship between IT governance and application
governance.

Origin of Initiative (application level)

Explorative case findings on application governance suggest that especially in SaaS

contexts the origin of the initiative to introduce a new applicationi.e. the question
whether it is started out of the business unit or the IT departmentis a strong predictor for
the resulting application governance mode (Winkler et al.2011a). In a procedural view,
the locus of initiative may be regarded as the exertion of a specific decision right that is
applicable during the first step in decision making (Fama and Jensen1983). However,
we delineate initiative from decision rights by two criteria that refer to time and
voluntariness. During the pre-decision phase of an implementation project, parties giving
input to investment decisions may be very distinct from those approving a decision (Xue et
al.2008). That is, project ideas can be brought in by business or IT stakeholders, while the
roles for deciding on and executing them may be more firmly defined. Still, depending on
the degrees of freedom in defining these later roles, we hypothesize a path dependency
between the initiation and later application governance outcomes, since the initiator may
inherently have a higher interest in taking over later decision rights (Fama and
Jensen1983). Accordingly we pose
H2: The locus of initiative is positively associated with application decision authority
(H2a) and task responsibility (H2b).
Initiating the implementation of an application may as well be thought of as an act
of applying particular knowledge towards a desirable outcome (Jensen1998). As argued
earlier, SaaS applications may require less technical knowledge than traditional, on-

145
5 Innovation Adoption and IT Governance in Enterprise Information Systems

premises applications for implementing them. For example, potential initiators may
assess the possibilites of application usage more easily by using online easier trial versions,
which are commonly available for SaaS (Gallaugher and Wang2002). This way, potential
initiators can better educate themselves about functional as well as technical features and
thus inform their initiation decision, as opposed to on-premises software that typically
requires a local installation before testing. Accordingly we pose that
H2: The association of application initiative with application governance (H2a/b)
is stronger for SaaS than for on-premises applications, i.e. the type of delivery model
(SaaS/On-premise) moderates the relationship between initiative and application gover-
nance.

Specificity (application level)

In line with TCT, application specificity is likely to affect application governance out-
comes. In the outsourcing literature, application specificity has been conceptualized
as the degree to which an application is customized prior to its implementation in an
organization (Benlian et al.2009). Consider the example an application for business
intelligence, which is inherently specific to a business, since the information provided
(e.g., via dashboards and data cubes) needs to be individually adapted to the companys
business context (Essaidi2010). Thus, operating this application requiresbesides IT
rolesmany business users of different management levels. In contrast, applications e.g.
for communicating and collaborating (e.g., email, conferencing and word processing),
reasonably require less business involvement (Cusumano2010). Accordingly we pose
H3: Application specificity is positively (negatively) associated with application decision
authority (H3a) and task responsibility (H3b) from business side (IT side).
There are grounds to assume that SaaS applications facilitate easier customization
and integration compared to the bulk of on-premises installed software. That is, SaaS
offerings try to avoid tenant-based customizations in terms of code changes at all cost by
providing a broad set of configuration options that are easily accessible to the user, e.g.
to add data fields, set up reports, define business rules, create workflows, etc. (Sun et al.
2008). Thus, even highly specific business applications are more likely to be governed
and maintained directly by business units. For comparison, we may think of the complex
parameter sets, configuration tables and custom written code in conventional on-premise
ERP systems (Dittrich et al.2009), which necessitate an involvement of an IT specialist
even to implement small changes to the application. Accordingly we pose that
H3: The association of application specificity with application governance (H3a/b)
is stronger for SaaS than for on-premises applications, i.e. the type of delivery model
(SaaS/On-premise) moderates the relationship between specificity and application gover-
nance.

146
 5.2 Comparing Authority for On-Premises Applications and SaaS

Scope of Use (application level)

Scope of application use relates to the frequency construct in TCT. Plausibly, the fre-
quency of transactions with an application is directly determined by the number of users
and business departments making use of this specific application. Different types of
business applications may vary strongly in the scope of use. A business intelligence
solution will typically have few expert users residing in management positions or staff
functions (Essaidi2010), while mentioned collaboration and communication tools may
be used across the entire organization.2. Based on the rationale of scale effects in the cost
for controlling and managing firm-wide applications (Nagpal2004), these are more likely
to be governed in a centralized, IT dominated fashion, compared to smaller applications
which may be more dispersed over the organization. Accordingly we pose that
H4: The scope of use of an application is positively (negatively) associated with applica-
tion decision authority (H4a) and task responsibility (H4b) from IT side (business side).
Based on the technological differences outlined earlier, the dispersion of application
governance across an organization may be facilitated by component based delivery models
such as SaaS (Tiwana and Konsynski2010). Arguing in a transaction cost theoretic
way, the paradigm of centralizing application authority in order to leverage internal
economies of scale may become much weaker for SaaS, since cost advantages refer only
to coordination costs, not to production costs (since the latter rather accrue externally
on provider side, (Armbrust et al.2009). This implies that with SaaS, there is a lower
incentive to allocate governance to a central IT department if the scope of use is medium or
low. Thus, the scope of use becomes more a decisive criterion in the allocation of SaS
governance, compared to on-premises models. In other words
H4: The association of the scope of use with application governance (H4a/b) is
stronger for SaaS than for on-premises applications, i.e. the type of delivery model
(SaaS/On-premise) moderates the relationship between scope and application governance.

Business IT Knowledge (organization level)

Drawing on the knowledge based view, complementary knowledge on business side is

likely to affect application governance arrangements. That is, the more technical and IT-
related knowledge business managers and employees have collectively absorbed, the
more inclined they will be to take over authority in application related decision making
and operational IT tasks (Sambamurthy and Zmud1999). According to a range of IS
studies, IT knowlede can derive either from internalization of external knowledge e.g.
on IT management practices, applications and technologies, or from prior experience,
for example from participation in IT projects (Bassellier et al.2003). Managers and
employees then build a capacity to apply this knowledge to achieve individually desirable
outcomes (e.g., for automating personal work routines, or organizing data). This in turn
2
This does not imply that specificity and scope are necessarily correlated. For example, ERP systems
are frequently used throughout many departments of a firm and still highly customized (Dittrich et al.
2009)

147
5 Innovation Adoption and IT Governance in Enterprise Information Systems

may lead them to seek higher involvement in decisions related to other applications
(Sambamurthy and Zmud1999). Accordingly we pose that
H5: The business IT knowledge is positively (negatively) associated with greater ap-
plication decision authority (H5a) and task responsibility (H5b) by business units (IT
units).
For mass delivery models like SaaS, where most detailed technical and implementation
related decisions are in the hands of an external party (Sun et al.2008), the need for
distinctive IT knowledge for making IT governance decisions may diminish. That is,
business managers and IT employees do not require more than basic IT knowledge in
order to make informed decisions regarding SaaS (Yanosky2008). This proposition fits
to the opposite observation that in custom software development projects, i.e. for com-
parably complex on-premise software delivery contexts, the client departments technical
knowledge is an important prequisite to reap the benefits from decentralized decision
rights (Tiwana2009). Thus, for on-premise software delivery, higher IT knowledge of
business users appears as a precondition to decentralize governance as opposed to SaaS.
Conversely:
H5: The association of IT knowledge in business units with application governance
(H5a/b) is weaker for SaaS than for on-premises applications, i.e. the type of delivery
model (SaaS/On-premise) moderates the relationship between business IT knowledge and
application governance.

IT Business Knowledge (organization level)

Complementary knowledge on the IT side refers to the IT employees collective knowledge

of the business domain (Tiwana2009). The earlier IT governance literature has rarely
discussed the relevance of absorptive capacities and knowledge on IT side (Brown and
Magill1994;Sambamurthy and Zmud1999). However, according to the knowledge
based view, business-knowledgeable IT employees are more likely to take over application
authority, potentially also for required business decisions that arise during the lifecycle
of an application. Business knowledge of IT professionals can be broadly conceptualized
as organization-specific and interpersonal/management competence (Bassellier and Ben-
basat2004). We focus on the first component and conceptualize IT business knowledge
as the insight into the business processes and day to day activities, since business appli-
cations primarily support these procedural aspects of an organization (Davenport and
Short2003). IT emplyoyees who have high knowledge of the business core processes
for example, from working previously in a business roleare evidently more educated to
make decisions on the what to implement to an application, when introducing, operating and
enhancing it. Accordingly we pose that
H6: The ITs business knowledge is positively (negatively) associated with greater ap-
plication decision authority (H6a) and task responsibility (H6b) by IT units (business
units).
The emergence of delivery models such as SaaS may demand different skill requirements

148
 5.2 Comparing Authority for On-Premises Applications and SaaS

for IT employees (Bassellier and Benbasat2004;Roepke et al.2000). Since technical

tasks are largely performed externally, the profile of an IT team managing SaaS may
resemble more a consultant to the business and a service integrator (Yanosky2008).
For example, in one of the cases explored in (Winkler et al.2011a), the central IT
department hired a team of experienced business consultants to govern and manage the
externally provided SaaS CRM solution. This is congruent with other researchers who
have emphasized the need for client-specific knowledge on the IT provider side (Susarla
et al.2010). Conversely, if internal IT does not provide the skills and insights in the
business processes and day to day operations, there is greater motivation for business
units to take over authority for SaaS based applications. Accordingly we pose
H6: The association of business knowledge in IT units with application governance
(H6a/b) is stronger for SaaS than for on-premises applications, i.e. the type of delivery
model (SaaS/On-premise) moderates the relationship between IT business knowledge and
application governance.
Table5.5presents an overview of all constructs with their theoretical framework, our
definition, the hypothesized impact on application governance and supporting literature.

Control Variables

To account for the differences among organizations, we include four control variables
in our research model which are commonly discussed in IT adoption and governance
contexts. These variables relate to industry, overall firm size, relative IT organization size
and time since implementation of the application. As we aim to demonstrate general- izable
results, we expect these variables not to be associated with the dependent variables of
application governance, i.e. to find them outside of our theoretical framework.

Industry Earlier research on organization and governance phenomena was unable to

find a significant and assocication between a corporations industry and the level of
decentralization of IT in the organization (Brown and Grant2005). We control for
the industry dimension by differentiating between manufacturing firms (selling physical
products) and service firms (selling informational products, respectively services). Since
there are possibly mixed forms of service and manufacturing businesses (for example
manufacturing companies that generate significant revenues with post-sales services), we
operationalize this control variable as a bipolar scale.

Firm Size The influence of firm size on IT governance arrangements has been viewed
ambiguously in the literature. While a number of earlier studies could not show a
significant relationship of firm size and the adoption of a particular IT governance
pattern (Brown and Grant2005;Tavakolian1989), later studies have substantiated
such relationship due to an increased need for coordinating mechanisms in complex
organizational structures (Brown and Magill1994;Sambamurthy and Zmud2000;Weill
and Ross2004a). We assume that the firm size will be largely reflected in the mode

149
150

5 Innovation Adoption and IT Governance in Enterprise Information Systems

Table 5.5:Hypotheses overview
H Construct Level Lensa Definiton Dimension Association with Difference
application governance SaaS/On-premises
0a Decision App. AT Locus of application Bus./ IT Dependent Tiwana2009; SaaS Bus Susarla et al.
Authority related control variable Fama and Jensen OnPr IT 2010;Choudhary
decisions rights 1983;Schilling 2007a;Xin and
0b Task App. AT Locus of application Bus./ IT Dependent 2000;Susarla SaaS Bus Levina2008;Sun
Respon- related decision variable et al.2010 OnPr IT et al.2008
sibility management rights
1 IT Org. AT Locus of application Bus./ IT Bus. Brown and Magill SaaS: weak Peak and
Governance related decision rights Bus. IT 1994;Brown1997; OnPr: Azadmanesh1997;
on the function level IT Sambamurthy and strong Tiwana2009;
Zmud1999;Weill Winkler et al.
and Ross2004a 2011a
2 Origin of App. AT, Locus of origin of the Bus./ IT Bus. Fama and Jensen SaaS: Gallaugher and
Initiative KBV idea to introduce the Bus. 1983;Winkler strong Wang2002;
application IT IT et al.2011a;Xue OnPr: weak Jensen1998;
et al.2008 Winkler et al.
2011a
3 Specificity App. TCT Degree to which the Low/ high High Benlian et al. SaaS: Essaidi2010;Sun
application has been Bus. 2009;Dittrich strong et al.2008;Xin
customized to the Low IT et al.2009;Sun OnPr: weak and Levina2008
companies processes et al.2008;
Susarla et al.2009
4 Scope of App. TCT Share of employees and Low/ high High IT Nagpal2004 SaaS: Tiwana and
Use departments making Low Bus. strong Konsynski2010;
use of the application OnPr: weak Winkler et al.
2011a
5 Business IT Org. KBV Amount of Low/ high High Sambamurthy and SaaS: weak Sun et al.2008;
knowledge complementary Bus. Zmud1999; OnPr: Tiwana2009;
knowledge on business Low IT Bassellier et al. strong Yanosky2008
side 2003
6 IT business Org. KBV Amount of Low/ high High IT Tiwana2009; SaaS: Roepke et al.
knowledge complementary Low Bus. Bassellier and strong 2000;Willcocks
knowledge on IT side Benbasat2004 OnPr: weak et al.2006;
Winkler et al.
2011a

a
Theoretical lenses: AT: Agency theory, TCT: Transaction cost theory, KBV: Knowledge based view
 5.2 Comparing Authority for On-Premises Applications and SaaS

of overall IT governance, rather than having a distiguishable impact on the appication

governance variables and control for this influence.

IT Organization Size Investigating a phenomenon that affects internal business and

IT organization, it stands to reason that we also consider the size of the internal IT
organization as a separate control variable. Since absolute size of the IT organization is
directly dependent on the overall firm size, we operationalize this parameter as the rela-
tive size (i.e., IT employees by total employees of the firm). Connecting this variable to
our theoretical framework implies the question, whether governance (i.e., the distribution
of decision rights) and organization (i.e., the distribution of resources) can be regarded
as independent phenomena, or need to viewed endogenously correlated (Sambamurthy and
Zmud2000).

Time We include the time since implementation of the application as an additional

control variable. IT governance is comonly viewed as a rather static concept that needs
to be well designed and implemented in an organization via certain mechanisms (Weill
and Ross2004a). This contrasts with the dynamic development of organizations. It is
plausible that governance arrangements that have been made at the time of implementing
an application, do not fit to the more recent conditions (i.e., at the time a survey is taken),
since organizations may walk through a learning process by using an application (Liang
et al.2007). Thus, in some cases governance arrangements need to be subsequently
adjusted, in some cases for example by formally assigning more authority to key business
users (Winkler et al.2011a). The time variable controls for the distorting effect of time
in post-implementation developments (Liang et al.2007).

5.2.5 Methodology

We operationalized the proposed research model in a multi-step procedure conducted a

survey with IT decision makers in 207 large-sized firms to test our hypotheses.

Construct Operationalization and Scale Development

Our research model is based on the explorative case studies and variables identified
in (Winkler et al.2011a). Relevant construct items were drawn from literature where
possible, i.e. for IT governance (Weill and Ross2004a), business IT knowledge (Bassellier
et al.2003), and specificity (Benlian et al.2009), or derived from the case material where
necessary (i.e., for decision authority, task responsibility, initiative, IT business knowl-
edge, and scope of use). For appropriate semantic granularity, business/IT dimensions
were operationalized on five-point scales (busines; business with IT involvement; business
and IT equally, IT with business involvement; central IT) (Brown and Magill1994), and
high/low dimensions via seven-point scales. In line with the extant literature, the IT
governance construct was operationalized by formative items, i.e. three classes of major

151
5 Innovation Adoption and IT Governance in Enterprise Information Systems

decision domains that mutually define an overall mode of IT governance (Weill and Ross
2004a). All remaining constructs are operationalized by reflective items, i.e. the items are
assumed to reflect the true value of the latent variable.
Several steps were taken to refine these measures as well as to assure reliability and
validity. During this process, each construct was reduced to 2-3 items (2 for seven-point
and 3 for five-point scaled items). The survey addressed high ranking managers and ought
to include more questions from a global survey related to this project, thus limiting total
questionnaire lenth and reducing redundancy was a major goal in this procedure. In a
first step, a category sorting was performed with 8 fellow researchers, i.e. the items were
shuffled and had to be sorted according to the definitions of the constructs. Second, an
online pretest was conducted with 29 IT professionals from industry and consulting as
well as experienced researchers, who were requested to fill the core questionnaire for a
company and two applications (SaaS and On-premise) of their choice. The pretest yielded
in acceptable reliability values (alpha>0.8) for all items and constructs considered in this
research except one (specificity), whose items were subsequently rephrased. The entire
final questionnaire was pretested in paper-based form by thinkaloud meetings with 3
different CIOs and only passed minor revisions regarding understandability and layout.

Data Acquisition and Sample Characteristics

We retrieved the addresses of Germany based firms with more that 50 million Euro of
revenues and more than 500 employees (excluding public sector) from a commercial
publisher of company information. The focus on large companies is motivated by the
assumption that we find a stronger variance in business/IT governance arrangemens
as well as a higher sensitivity for this topicin large firms with higher coordination
needs (Weill and Ross2004a). In April 2011, formal invitations were sent out to 2,886
companies including a paper version of the questionnaire, a return envelope, and cover
letter including references to the project website as well as the online questionnaire. As
an incentive, we offered a small gadget for the first 100 respondents as well as the chance
to win a tablet computer for all participants. After a six-week response period and
combined phone/email reminders, we received a feedback from total 534 companies, out
of whom 220 provided usable responses (60% online, 40% paper-based). We attribute
this moderate response rate (7.6%) to the total questionnaire length as well as to the
increasing saturation of IT decision makers with research inquiries, a fact which was also
stated in some of the reminder calls. Further 13 records were removed due to missing
or inconsistent values in the questions regarding our research model. Cover letter and
the complete questionnaire in German language, including the items for this study, are
presented in Appendices6.1and6.2.
Respondents stated to be top-level information officers (20%), senior IT managers
(47%), IT managers (20%), IT staff (3%), as well as business managers (3%), senior
business managers (3%) and CEOs (4%) with a median of 11.5 years of work expe-
rience. Textual job descriptions revealed that respondents with business roles were
chief operating officers (COOs), business unit IT managers, or in similar positions with

152
 5.2 Comparing Authority for On-Premises Applications and SaaS

Table 5.6:Application types and sample distribution

Application All SaaS On- Application All SaaS On-
types Prem types Prem
Business 7% 8% 6% Office & 5% 7% 5%
Intelligence & Productivity
Analytics
Communications 7% 16% 2% Production 4% 5% 3%
& Collaboration Execution
Customer 8% 13% 6% Service 2% 4% 1%
Relationship Management
Mgmt. (CRM)
Engineering & 0% 0% 1% Supply Chain 3% 8% 1%
Design Mgmt. (SCM)
Enterprise 48% 17% 66% Other 11% 16% 8%
Resource
Planning (ERP)
Human Resource 3% 7% 2% Total 207 76 131
Management 100% 100% 100%
(HRM)

IT responsibility, therefore we decided not to remove these records from the sample.3
The industry distribution (based on the German industry classification code) does not
reveal a significant bias compared to the distribution of invited companies. The size
distribution (measured by number of employees) indicates that larger firms responded
overproportionally to our inquiry. The mean number of employees is 4,780 (quartiles:
700; 1,300; 2,500; skewness: 6.7) and mean number of IT employees is 130 (quartiles:
7; 15; 35; skewness: 11.2). The data selection procedures and sample description is
reported more extensively in Appendix apx:bitdata.
After completing the questionnaire section on overall IT organization, participants
were asked to provide information about an SaaS or an on-premise application of their
choice that is relevant to their business and that they are knowledgeable about. For
the purpose of distinction, SaaS had been defined as any enterprise software that is
deployed over the internet by an external provider and is accessed by users at different
companies through the web browser, while an on-premises application was described
as any enterprise software that is installed in a traditional way and run on computers
belonging to the premises of your company, i.e. this installation is only used by your
company. Out of 207 respondents, 76 provided information about an SaaS and 131
regarding an on-premises solution. The underlying distribution of between SaaS and on-
premises according to a classification into common application types is presented in
Table5.6.

3
T-tests demonstrated no significant subgroup differences in the model variables for business and IT
respondents, except for the indicators Spec1 and DecAuth2. As a robustness check, we also recalcu-
lated the PLS model excluding respondents with business roles, which did not affect the final results
of the hypotheses tests.

153
5 Innovation Adoption and IT Governance in Enterprise Information Systems

Measurement Model and Common Method Bias Assessment

We use partial least squares structural equations modelling (PLS) to assess the psychome-
tric adequacy of the measurement model and combine this with multigroup comparison
methods to test the proposed hypotheses. The choice for PLS is motivated by the
relatively small sample size for each subgroup (i.e., SaaS versus on-premises appliations)
(Chin1998b), the ability to handle formative constructs and multiple dependent vari-
ables, as well as the rather explorative character of this study that includes several newly
developed constructs (Hair et al.2011).
To assess the adequacy of the measurement model, formative and reflective constructs
are considered separately. For the formative construct (IT governance), all tolerance
values (0.883; 0.749; 0.835) are clearly above the recommended threshold of 0.5 (respec-
tively VIF<5) (Hair et al.2011). Thus multicollinearity can be excluded as an issue for
the measurement of the IT governance construct.
For reflectively measured constructs, all item loadings are clearly above the threshold
of 0.7, indicating that more than half of their variance is explained by their substantive
latent variable. Convergent validity is supported by acceptable values for Crobachs alpha
(Alpha>0.7), composite reliability (CR>0.7) and average variance extracted (AVE>0.5) for
each of the constructs (Hair et al.2011), see Table5.7. We assess discriminant validity by
evaluating cross-loadings as well as the Fornell-Larcker criterion (Chin1998b, p. 321). The
mean of absolute item-to-construct cross-loadings is 0.136 with maximum values at 0.633
between the two dependent variables, thus below the critical value of 0.7. Since both
constructs (decision authority and task reponsibility) represent core dimensions of
application governance, this correlation is in line with our theoretical model and thus
does not refute model validity. Second, the Fornell-Larcker criterion, stating that construct-
to-construct correlations should be below the square root of AVE (represented as diagonal
elements in the matrix), is as well fulfilled for all constructs, see Table5.7. All item
loadings and cross-loadings are reported in Appendix8.
For all self-reported data, there is a threat for a common method bias (CMB) due to
the subjects motif to give socially desirable and cognitively consistent answers (Podsakoff
and Organ1986). We assess potential CMB by a Harmans one-factor test (Podsakoff
and Organ1986), as well as by including a latent method factor in the PLS model. The
first factor from an exploratory factor analysis on all variables of the theoretical model
accounts for 0.27 of total variance (loading distinguishably on the items of application
governance). This provides a preliminary objection against the existence of a single
dominant factor that would explain the majority of variance in the sample (Podsakoff
and Organ1986). For confirmation, we follow the procedure described by (Liang et al.
2007) and include a common method factor in the PLS model that comprises all model
indicators and potentially influences the constructs principal indicators.4 CMB can
be assessed by comparing the variance of model indicators explained by substantive
constructs with the variance resulting from the method factor. Regarding our model,
4
For the purpose of CMB assessment, formative constructs (IT governance) may be modeled reflectively
(Liang et al.2007).

154
 Table 5.7:Validity criteria

5.2 Comparing Authority for On-Premises Applications and SaaS

Operationalization Descriptive Convergent Discriminant val.(Latent variable correlations with AV E)
validity
Construct Items Scale Mean SD Alpha CR AVE 1 2 3 4 5 6 7 8 9 10 11 12
DecAuth 3 1-5 3.75 0.90 .79 .88 .70 .84
TaskResp 3 1-5 4.06 1.05 .83 .90 .74 .67 .86
ITGova 3 1-5 4.07 0.50 .56 .75 .51 .27 .17 .71
Init 2 1-5 3.44 1.24 .82 .92 .85 .60 .52 .22 .92
Spec 2 1-7 4.46 1.64 .85 .93 .87 .09 .11 -.04 .12 .93
Scop 2 1-7 4.59 1.38 .82 .92 .85 .43 .36 .01 .42 .34 .92
BusKnow 2 1-7 3.94 1.40 .87 .92 .86 .04 -.04 .08 -.03 .03 .02 .93
ITKnow 2 1-7 5.27 1.05 .82 .92 .85 .23 .20 .30 .08 -.03 -.05 .16 .92
Industry 2 1-5 2.56 1.50 .85 .93 .87 -.15 .00 -.10 -.12 .07 -.06 .09 -.10 .93
FirmSizeb 1 Lg(E) 3.17 0.52 1 1 1 -.05 .02 -.02 -.01 -.01 -.14 -.03 .02 -.01 1.0
ITOrgb 1 E/I .023 0.04 1 1 1 -.11 .04 .05 -.22 -.04 -.20 .17 .05 .35 -.11 1.0
Time 1 Yrs 8.16 6.38 1 1 1 .13 .14 -.01 -.03 .24 .29 .03 .01 -.13 .05 .00 1.0

a
IT Governance operationalized as formative variable, thus convergent validity criteria (in italics) are not applicable and only for demonstrative purposes
b
E = number of employees total, I = number of employees in IT
155
5 Innovation Adoption and IT Governance in Enterprise Information Systems

the average substantively explained variance is 0.652 while the average variance from
the method factor is 0.015 (ratio 1:44). Additionally, after bootstrapping all substantive
loadings remain significant (p<0.01), while most path coefficients from the method factor
are not significant. Altogether these results indicate that CMB is not a serious concern
in this study.

5.2.6 Analysis
The PLS model and its factor scores are used to perform descriptive analyses on the
two dimensions of application governance, as well as to assess the structural model and
perform subgroup comparisons. We may refer to the research questions posed at the
outset (page18) to structure our analysis.

Patterns of Application Governance Descriptive Results

First, regarding the question how organizations allocate horizontal authority for applica-
tions in general, we find mixed results. For descriptive purposes, we use the standardized
factor scores of decision authority (DA) and task responsibility (TR) to perform a hier-
achical clustering using Wards linkage, a widely applied and well performing clustering
technique (Punj and Stewart1983). We observe a relative drop in the distance measure
when reducing from four to three clusters (second derivation of the last four distance
measures: 13.9; 9.3; 21.2; 165.3) and thus obtain three clusters. According to the
unstandardized mean values (m) of DA and TR, we may say that these clusters rep-
resent IT dominated (nIT =150; mDA=4.17; mT R=4.54), business dominated (nBus=18;
mDA=2.21; mT R=1.46), and mixed (nmix=39; mDA=3.00; mT R=3.45) patterns of appli-
cation governance, see Figure5.4(left side). Worth noting, the scatter plot also reflects
the strong correlation of decision authority and task responsibility (Pearsons r=.663)
with a slight disaggregation towards the lower right, i.e. greater business authority
(DA<3.0) combined with greater IT task responsibility (TR>3.0) especially for the mixed
governance patterns.
A chi sqare test on the frequency distribution of SaaS and on-premises applications
across these three governance patterns (2=14.17; df=2; p=.001) as well as unpaired t-
tests on decision authority (T=-1.88; df=205; p=.061) and task responsibility (T=-3.0; df=205;
p=.003) reveal, that according to our sample, SaaS is significantly more likely to be
governed in a business dominated fashion compared to on-premises applications (i.e.,
H0 is supported). However, the distribution of SaaS applications by governance patterns
clearly depicts that governance of SaaS appliations varieswhich is in line with the
exploratory case findings in (Winkler et al.2011a)and that IT dominated governance is
still the prevailing pattern in the majority (59%) of the considered cases of SaaS
delivery, see Figure5.4(right side).
We may attribute the governance differences between the delivery models to the prop-
erties of the SaaS and on-premises application types represented in our sample, that
determine the model outcomes. Examining the factor scores of antecedent constructs

156
 5.2 Comparing Authority for On-Premises Applications and SaaS

Figure 5.4:Application governance patterns

we find that SaaS and on-premises applications differ significantly in scope of use (T=-
6.62; df=205; p=.000), i.e. SaaS applications are less widely used, and in specificity (T=-
6.63; df=205; p=.000), indicating that SaaS applications are less customized to the
companies processes. Regarding control variables, there is also a significant difference
in the time since implementation (mean 4.05 years for SaaS and 10.28 years for on-
premises). Lower specificity, however, would contradict a more frequent allocation to
business units according to our hypotheses, which motivates us to look closer at the
hypothesized factor relationships.

Determinants of Application Governance Model and Subgroup Tests

To test our hypotheses, we assess the structural model for all 207 organization-application
pairs, as well as for the 76 organizations using SaaS, and the 131 firms providing infor-
mation about on-premises usage separately. Statistical significance of the parameter
estimates is assessed based on the t-values from separate bootstrapping procedures each
using 1,000 resamples.5 We compare the path coefficients of the SaaS and on-premises
models (cSaaS , cOnP r ) by two kinds of tests: a t-test based on the standard errors as
described by (Keil et al.2000) and an adapted Mann-Whitney-Wilcoxon (MWW) test
based on the bootstrap samples (Henseler et al.2009). Both tests produce congruent
results, see Table5.8(page160). We regard paths with test error probabilities p <0.1
as significantly different and thus decide on the support for the differential hypotheses
(Hx). In addition we assess the effect sizes (f) for those constructs which are signif-
5
Note: For the subgoups samples, we estimated the model parameters and t-values separately for each
single control variable, in order to adhere to the recommended sample size requirements for PLS (i.e.,
sample size greater than ten times the number of incoming links of the dependent variable,Chin
1998b). A similar approach is taken by (Liang et al.2007). For the total sample (n=207), all model
constructs and control variables were tested together.

157
5 Innovation Adoption and IT Governance in Enterprise Information Systems

icantly associated (Chin1998b, p. 316). The results of the model test and subgroup
analysis for SaaS and on-premises are illustrated in Figure5.5.

AT TCT KBV

IT Governance Initiative Scope of Business IT IT Business

Specificity
(B/IT) (B/IT) Use Knowledge Knowledge

0.03 -0.01 0.35** 0.42** -0.01 -0.07 0.35** 0.23* 0.11 -0.04 0.31** 0.20
0.16* 0.10 0.46** 0.39** -0.01 0.03 0.17** 0.03 -0.08 -0.12 0.11 0.19**
(=) (=) (=) (=) (=) (=) A (=) A (=) A (=)

Application
Decision Authority Task Responsibility
Governance
R2=0.56; R2=0.40 R2=0.38; R2=0.34
(Business/IT)

-0.05 0.12 0.02 0.10 -0.06 -0.07 0.07 0.07

-0.07 0.11 -0.06 -0.03 -0.06 0.34** 0.13* 0.10
(=) (=) (=) (=) (=) A (=) (=) Subgroups:
SaaS (n1=76)
OnPrem (n2=131)

* = p < 0.10
Industry IT Org. ** = p < 0.05
Firm Size Time
Control (M/S) Size
Tested differences
Variables
(=) not significant
A significant

Figure 5.5:Model tests and subgroup analysis

H1: Regarding the first hypothesis, we find that overall IT governance is not sig-
nificantly related with application governance for the total sample, especially not with
task responsibility (H1b). However, we observe a weakly singnificant relationship with
decision authority for on-premises applications (c=0.16; effect size f=0.04) and conclude
that the insignificant relationship in the combined sample is caused by heterogeneous
influences in the underlying populations. The multipgroup analysis does not siginficantly
support the difference in path coefficients (p<0.16; p<0.18), nevertheless, the association
of IT governance with decision authority (H1a) is much weaker for SaaS, i.e. practically
not existent, compared to on-premises applications. Hence, we may state that H1a is
at least partly supported.
H2: We find strong support for the contingent influence of the origin of initiative
on both, application decision authority (H2a) and task responsibility (H2b) represented
by with the strongest path coefficients in the model (c=0.47, effect sizes f=0.28 on
DA, and c=0.47, effect size f=0.24 on TR). Multigroup tests do not confirm that this
influence is significantly differs with the delivery model (H2 not supported), i.e., the
origin of initiative appears to be equally important for the allocation of SaaS and on-
premise application governance (cSaaS =0.35 and cOnP r =0.46 on DA, cSaaS =0.42
and cOnP r =0.39 on TR).
H3: According to our model test, application specificity is not significantly associated

158
 5.2 Comparing Authority for On-Premises Applications and SaaS

with application governance, neither for decision authority (c=-0.05, H3a), nor for task
responsibility (c=-0.03, H3b). In this case we are unable to attribute this finding to an
unobserved heterogeneity in the sample, since the low and insignificant path coefficients
are congruent for both, SaaS and on-premises subgroups, which is confirmed by high p-
values in the multigroup comparison (i.e., H3 and H3 not supported).
H4: The scope of use is significantly associated with application governance (c=0.24 on
DA, c=0.20 on TA), i.e. the more departments and employees make use of an appli-
cation, the more likely it is to be governed in a IT dominated fashion, yet, the effect
sizes of this influence regarding the overall sample are comparably weak (f=0.05 on DA,
and f=0.04 on TR). The multigroup analysis shows that the association of scope with
decision authority, and to a much lesser extent also with task responsibility, is stronger for
SaaS than for on-premises applications (DA: cSaaS =0.35 and cOnP r =0.17 ; TR: cSaaS
=0.23 and cOnP r=0.03), thus H4 is supported for decision authority.
H5: Higher business IT knowledge was hypothesized to be asscociated with stronger
application governance by business units. This relationship is not significant for task
reponsibility (TR: c=-0.10), and not existent for decision authority (c=0.01). Regarding
subgroup differences, we find a positive association of business IT knowledge with decision
authority for SaaS using organizations (DA: cSaaS =0.11). Although this relationship is
statistically not significant, it leads to significant subgroup differences on the p<0.1
confidence level. We may interpret this significance as a weak support for our hypothesis
H5a stating that with SaaS, the businesses IT expertise is becoming comparaby less
important in defining application authority, as opposed to on-premises delivery models.
H6: Finally, we find significant support for an association of IT business knowledge and
application governance (c=0.16 on DA, c=0.18 on TR) with low effect sizes (f=0.03
on DA; f=0.02 on TR). Regarding the subgroup analysis, again we observe hetero-
geneity in the association with decision authority (DA: cSaaS =0.31 and cOnP r=0.11;
p<.04). Multigroup tests show significant path differences between SaaS and on-premises
subgroups (p<0.04). Aparently the strong link for SaaS applications also causes the
significant link in the total sample. This supports our hypothesis H6a implying that for
SaaS applications, complementary knowledge in the IT department is comparatively more
important to establish, respectively retain, decision authority in the IT department.
Regarding the control variables, we can rule out influences respresented by the type of
industry (i.e., manufacturing versus service) and the size of the organization. The relative
size of the IT organization significantly influences the allocation of task responsibility
(c=0.18), i.e. the more human resources work in the IT organization, the more likely is
IT to take over task responsibility, and vice versa. Interestingly, the influence represented
by organizational characteristics exclusively results from on-premises applications in the
sample (cOnP r =0.35 ; effect size f=0.12), not from SaaS (cSaaS =-0.07), which is
supported by the multigroup tests (p<0.01). Ultimately, the time since implementation
partly influences decision authority for on-premises applications (cSaaS =0.35). That is,
more recent governance arrangements for on-premises applications are made in a more
decentralized fashion, an effect we control for in our model.

159
5 Innovation Adoption and IT Governance in Enterprise Information Systems

Altogether, the model variables collectively explain 46.7% of the variance in decision
authority and 37.9% in task responsibility. Explanatory power is substantively bet-
ter for SaaS (R2DA=56.0%; R2 T R=38%) than for on-premises applications (R2 DA =40.1%;
2
RT R =33.6%). The influence originating from control variables is negligible except for
the effect on on-premises task responsibility (effect size fControls=0.16). The results of
the hypotheses tests and multigroup analysis are summarized in Table5.8.

Table 5.8:Results of the hypotheses tests

H Path Model tests Multigroup analysis
All SaaS OnPrem Association T-test MWW Group
n=207 n1=76 n2=131 p p difference
H1 ITGovDecAuth .10 .03 .16 Partly sup. .16 .18 Not sup.
ITGovTaskResp .02 -.01 .10 .26 .44 Not sup.
H2 InitDecAuth .47 .35 .46 Supported .21 .20 Not sup.
InitTaskResp .47 .42 .39 .42 .40 Not sup.
Spec DecAuth
H3 -.05 -.01 -.01 Not sup. .49 .50 Not sup.
Spec TaskResp
-.03 -.07 .03 .25 .26 Not sup.
Scop DecAuth .24 .35 .17
H4 Supported .08 .09 Supported
Scop TaskResp .20 .23
.03 .15 .28 Not sup.
H5 BusKnowDecAuth .01 .11 -.08 Not sup. .10 .06 Supported
BusKnowTaskResp -.10 -.04 -.12 .29 .19 Not sup.
H6 ITKnowDecAuth .16 .31 .11 Supported .04 .03 Supported
ITKnowTaskResp .18 .20 .19 .48 .32 Not sup.
C1 IndustryDecAuth -.05 -.05 -.07 - .44 .35 -
IndustryTaskResp .06 .12 .11 .48 .16 -
C2 FirmSize DecAuth -.02 .02 -.06 - .27 .32 -
FirmSizeTaskResp .06 .10 -.03 .19 .30 -
C3 ITOrgSiz DecAuth .04 -.06 .06 Partly infl. .20 .20 -
ITOrgSizTaskResp .18 -.07 .34 .01 .01 Influential
Time DecAuth .13
C4 .08 .07 Partly infl. .32 .30 Partly infl.
Time TaskResp .10
.07 .10 .41 .36 -

p <.10; p <.05

5.2.7 Discussion

This study draws on the extant IT governance literature to develop a research model
that applies multiple theoretical lenses in order to explain two important dimensions of
application governance: decision authority and task responsibility. Based on the steady
evolution in the technological architectures, our particular focus was on investigating
the dynamism that the use of emerging delivery models such as SaaS may imply for
business organizations in comparison with traditional, on-premises models. Descriptive
analyses of data from 207 organization/application pairs indicate that three patterns of
application governance can be distinguished: IT dominated, business dominated, and
mixed forms. Empirical tests of our model provide support for IT governance (partly),
the scope of application use, the IT employees knowlege of the business domain, and
the origin of the initiative being significantly associated with the application governance

160
 5.2 Comparing Authority for On-Premises Applications and SaaS

pattern.
Regarding the role of IT governance, we found that the allocation of overall IT decision
rights is associated with governance arrangements for on-premises, but not for SaaS ap-
plications. This finding may result from the fact that SaaS applications still account for a
minor fraction of total application architectures. For such applications that deviate from
the overall mode of IT governance, business and IT units may have special arrangements.
Conversely, we could argue that a more modular view of IT governance is appropriate:
that is, overall IT governance is an aggregation of a multitude of arrangements on
application and IT artifact levels. Adopting the latter view would imply that with the
increasing prevalence of SaaS based delivery models, overall IT governance is likely to
be shaped towards more business dominance. The significant correlation between time
and task responsibility in our model at least provides a retrospective argument for the
existence of such dynamism in application governance arrangements.
Combining this view with the findings regarding the scope of use and specificity, we
obtain a reasonable explanation for why SaaS governance patterns differ from on-premises
applications. In line with the transaction cost theoretic view, todays SaaS applications
are evidently those that have a smaller scope of use within organizations and are therefore
more frequently governed in a business dominated fashion. This fits the argumentation of
the long tail for Internet based commodities (Chong and Carraro2006;Brynjolfsson et al.
2006). Since cost advantages for SaaS largely accrue on the external IT provider side,
this emerging delivery model represents a cost-effective alternative for specific application
demands that otherwise would not be fulfilled. SaaS applications may often extend
existing application architectures, rather than cannibalizing large and highly complex
local instances, such as on-premises ERPs.6 This also implies, that the imperative to
leverage economies of scale becomes a much weaker argument for centralizing governance
to a corporate IT department when SaaS models are prominenta finding that is as well
supported by the case findings in (Winkler et al.2011a).
From a knowledge-theoretic lens, the business knowledge of IT employees emerges as
an important variable associated with application governance, especially for SaaS-based
sourcing. This finding underpins previous studies that emphasize the change in the
competencies of IT professionals and retained employees in IT outsourcing relationships
(Bassellier and Benbasat2004;Roepke et al.2000). Conversely, we were able to provide
weak evidence that the IT knowledge of business staff influences application governance
for SaaS less than for on-premises applications. We attribute this weak relationship to the
fact that business IT knowledge was conceptualized for the entire business organization,
rather than for a close group of application users. However, assuming that validity, this
finding may reflect a lower need for specialized IT competences of SaaS users for being
be involved in application related decision making, compared to traditional IT delivery.
Finally, the origin of the initiative for implementing an application emerges as a crucial

6
Side note: We included an additional binary item in the survey asking whether the application was
introduced as a new solution or replaced a previous solution. The group differences are significant,
40% of SaaS, versus 19% of on-premises applications had been introduced as new solutions.

161
5 Innovation Adoption and IT Governance in Enterprise Information Systems

predictor for both, SaaS and on-premises application governance. This finding does
not principally contradict our initial supposition that authority for SaaS applications is
allocated largely dependent on the voluntary initiative of either of the parties business
or IT. However, it does lead us to assume that this has been an influential factor also
for on-premises delivery models. This would represent an unexpected result inasmuch as
literature suggests that in formalized IT investment processes, the parties initiating a
project can differ strongly from those taking over later application operation and decision
making (Xue et al.2008). Altogether, our finding rather strengthens the agency theoretic
view, which suggests that initiative may as well be regarded as one of the facets in the
broader class of decision management rights (Fama and Jensen1983).

Theoretical Contribution

The central contribution of this work is represented by the theoretical framework it

provides for explaining governance phenomena. This framework comprises two distinctive
novel ideas for which we provide empirical evidence. The first idea is to conceptualize
governance on the application level as a subset of overall IT governance arrangements,
thus disaggregating the facets of governance along the technological architecture. As we
contend, this approach is meaningful in order to connect the analysis of contingencies on
organizational level with the influence represented by different properties of enterprise
applicationsor in general, the IT artifact (Orlikowski and Iacono2001). The results
provide evidence that such a disaggregated view is useful to discover arrangements that
deviate from (prior) overall schemesin this case application governance arrangements
(specifically decision authority) for SaaS delivery that deviates from overall IT gover-
nance. Furthermore, reflecting the agency theoretic imperative to disaggregate decision
management and decision control rights (Jensen1998), we conceptualized application
governance by the two empirically correlated but theoretically distinct dimensions deci- sion
authority and task responsibility and demonstrated their value to describe different
governance patterns.
Second, this work reframes the theoretical basis for explaining governance phenomena
by enhancing the view to three theoretical lenses. Although the extant IT governance lit-
erature has as well implicitly drawn on constructs from knowledge based and transaction
cost theoretic views, this study explicitly integrates these views in a single research frame-
work and conceptualizes corresponding constructs. Other studies, for instance on SaaS
adoption, have taken similar approaches in order to facilitate a broader understanding
of the phenomenon (Benlian et al.2009). Regarding our study, transaction cost theory
in particular emerges as a valuable lens to identify and operationalize artifact-specific
properties, such as application specificity and scope of use. The knowledge based view
provides a meaningful lens to capture the specific properties pertaining to the organiza-
tion and its human resources, which are critical in a governance context. The empirical
evidence supports the value of this multi-theoretical framework inasmuch as most of its
embodied constructs exhibit clear relationships with the regarded phenomenon, in this
case governance on application level.

162
 5.2 Comparing Authority for On-Premises Applications and SaaS

Managerial Implications

This study offers a number of findings that may guide managerial decision making in the
presence of various application adoption contexts (e.g., SaaS and on-premises). First of all,
the empirical analysis based on multiple antecedent factors facilitates a contingent
understanding of application governance arrangements. For example, it informs that
smaller scoped applications may as well be governed in a more decentralized fashion
(especially for SaaS) without necesariliy conflicting with overall governance arrange-
ments. Consequently, the notion of overall IT governance may be brought forward
from a monolithic understanding of the entire IT function, towards a modular view
that captures appropriate arrangements for each single cluster of IT artifacts. Such
disaggregated understanding may be helpful given the increasing amout of IT activities
that take place outside of the IT department.
According to our analysis, two factors play a crucial role to define governance ar-
rangements on the application level. First, apparently initiators (e.g., single actors or
departments) play an important role in IT adoption contexts. Firms may rethink this role
and question whether initiators should also be responsible for later application operation and
management. This can potentially be achieved by implementing appropriate proce- dural
mechanisms such as idea management processes and project funnels. (Reversely, if initiators
know ex ante that large parts of tasks and responsibilities remains with them, they might
have less incentive to take a lead in future initiatives) Second, according to our results the
knowledge of IT professionals and their insight into business processes poses an
increasingly important skill for future IT units to sustain effective decision making and
prevent a drift of IT authority to the business units. Companies may reconsider the
benefits of human resource practices, such as job rotation between business and IT as
well as trainings of IT staff, as an effective means to achieve this goal. Altogether, this
study may help practitioners to assess the organizational impact that the use of
emerging delivery models may have for their organization and the arrangements between
business and IT units. Our results dissociate from the radical view of some proponents
(Carr2005), however, we provide relevant insights under which conditions governance
shifts may be appropriate.

Limitations and Future Research

Two important limitations of this study merit consideration. First, due to questionaire
length restrictions, all model constructs have been operationalized by a relatively low
number of items. Arguably, this may have affected the predictive validity of some of
these constructs. Although criteria for convergent and disctiminant validity are fulfilled,
a semantically broader operationalization of the model constructs may have resulted in
significant path coefficients for those hypotheses that are currently only vaguely evident
or not significant (Diamantopoulos et al.2012). Second, a potential unobserved bias of
the results may result from the research design. Respondents were allowed to choose the
application they wanted to provide information on, which resulted in a broad distribution

163
5 Innovation Adoption and IT Governance in Enterprise Information Systems

of different application types across the SaaS and on-premises categories. We assume that
this individual choice results in a random sample of applications that is comparable to
the overall population of Saas and on-premises applications across organizations (which
is unknown). This assumption is especially important when interpreting the descriptive
results.
Future research may extend the modular view on governance phenomena in three
promising directions. First, the strong influence by the origin of initiative suggests that
companies allocate application governance largely in a path dependent way based on the
seemingly arbitrary initiative of either party in the organization. Researchers may bring
more light on the question of voluntariness in IT governance and adoption contexts.
Relevant questions might be, for example, what motivates employees to initiate new
technology adoption and how do governing roles evolve? Longitudinal and process-
theoretic approaches appear appropriate means to address these issues. Second, while
the influence of specificity is widely emphasized by in the outsourcing and SaaS literature
(Benlian et al.2009;Nagpal2004), its role in the allocation of application governance
remaines ambiguous. (We found significant subgroup differences, however, a nonsignif-
icant association with governance). Conceivably, there is an argument for two effects:
lower application specificityoperationalized as low customization needsmay allow
for more business authority, however, higher business authority may as well cause higher
specificity. Future research may address this issue in order to unveil the meadiating
factors that justify (or disprove) an influence of the specificity construct on IT artifact-
level governance phenomena. Finally, our theoretical model was free from a criterion
variable. Thus, we were not able to assess whether a specific application governance
pattern is beneficial or not. Reasonably, future research may develop more normative
theories for application governance phenomena, based on the antecedent factors provided
by this research.

5.2.8 Summary
This chapter addressed the neglected question of how organizations allocate information
technology (IT) authority between business and IT units for enterprise applications, with
a particular focus on the potential changes that emerging delivery models such as Soft-
ware as a service (SaaS) may bring about. We theorized that allocations of application
authority result from agency, transaction cost, and knowledge-based contingencies and
developed a research model that conceptualizes application governance arrangements
by the two dimensions decision authority and task responsibility. Empirical tests using
data from a survey with 207 large firms provide partial support for our model. The
data reveals that authority for SaaS-based applications is allocated more frequently to
the business units andmost notablyis established largely independent from overall
IT governance arrangements. This study contributes a more modular, application level
view to IT governance theory and demonstrates its usefulness to uncover the dynamics
represented by emerging patterns of IT delivery. Relevant implications for theory and
practice were discussed.

164
 5.3 A Process Model for Explaining Governance of Software as a Service

5.3 A Process Model for Explaining Governance of Software as

a Service

5.3.1 Preamble

This chapter has been initially published and presented at the German Multikonferenz
Wirtschaftsinformatik 2012 in Braunschweig (seeWinkler and Gnther2012). Note that
one case example has been renamed (Case A replaced by C) to ensure the coherence
with Chapter5.1of this dissertation.

5.3.2 Introduction

Firms are socio-technical systems. Any change to the technical infrastructure may
also imply a change to the internal organization (Robey and Boudreau1999). When
implementing new enterprise applications, business and IT decision-makers face the
challenge how to allocate decision rights for the use, management and enhancement
of such application. This phenomenon has been commonly identified as an important
aspect in IT governance.
In the past, the focus of IT governance has been directed on balancing between cen-
tralized (i.e., IT departmental) and decentralized (i.e., business units) decision rights.
This appears reasonable, as the internal IT department has been regarded as the focal
point of IT delivery. However, emerging delivery models such as Software as a Service
(SaaS) are likely to defy this view. With SaaS, a third party comes into play providing
large parts of IT delivery, so that business departments may be more inclined to take
over large parts of decision authority and application-related activities (Yanosky2008;
Golden2010;Winkler et al.2011a).
Previous work on SaaS governance has proposed a contingency model including or-
ganizational and technical categories to explain in which cases authority for the SaaS
application is rather allocated to the business or to the IT side (Winkler et al.2011a).
However, such models follow a variance-theoretic paradigm. Thus, they are hardly able
to embrace complex temporal and causal interrelationships between the factors and fail
to explain exactly how they are related (Newman and Robey1992).
In this work, we build on previous models and take a process-theoretic approach to
examine the governance of SaaS-based applications. For this purpose, we first define a
process model that considers the three actors business, IT and external provider. Then,
we illustrate the applicability of this model in two cases of SaaS adoption to explain
different governance outcomes. The comparison of the cases reveals some of the complex
relationships and path dependencies between the variance-theoretic factors. The model
can be used to study further cases of application adoption and better understand the
allocation of application governance in each respective case.
The remainder is structured as follows: In the next section we review related work
on IT governance, Software as a Service and process theory. Then, in Section5.3.4we

165
5 Innovation Adoption and IT Governance in Enterprise Information Systems

present our process approach for investigating SaaS governance. Section5.3.5empirically

demonstrates the approach in two cases of SaaS adoption. Section5.3.6summarizes the
results and outlines limitations as well as future work.

5.3.3 Related Work

IT Governance and Subdomains

IT governance is commonly understood as a subset of corporate governance aiming

to ensure that the IT organization sustains the organizations strategy and objectives
(De Haes and Van Grembergen2004). Governance mechanisms are installed on struc-
tural, procedural and relational level to connect the stakeholders (i.e., the business) and
the entities in authority of information technology (i.e., the IT department or external
providers) (Peterson et al.2000). While practitioner literature has much focused much
on procedural mechanisms and developed several governance frameworks, such as ITIL
and CoBIT (Peterson et al.2000), earlier IS research has related IT governance primarily
to the locus of authority for IT functions (Brown1999), thus to the structural level
(Peterson et al.2000).
Commonly, allocations of IT authority can be classified into centralized, decentralized
and federal archetypes (Brown1999;Sambamurthy and Zmud1999). Weill and Ross
(Weill and Ross2004a) propose a more sophisticated framework comprising six gover-
nance archetypes (business monarchy, IT monarchy, IT duopoly, etc.), which essentially
combine the horizontal (i.e., business vs. IT) with the vertical (i.e., executive vs. em-
ployee level) distribution of authority. Building on that, a few works demonstrate that
firms need to allocate decision authority depending on their strategic goals, context and
environment, for example in order to balance between the need for local flexibility versus
global standardization (Weill and Ross2004a). Some authors have also broken down
the concept of IT governance to different subdomains, such as infrastructure governance
(Gu et al.2008) and data governance (Khatri and Brown2010). Therefore it appears
conceivable to draw on governance theory also to explain the mode of governance for
Software as a Service, i.e. to explain SaaS governance.

Software as a Service and Application Governance

Software as a Service (SaaS) refers to an increasingly deployed delivery model, where

standard enterprise applications are provided as a service over the Internet (Cusumano
2010). Conceptually, SaaS is attributed to the highest layer of the cloud computing stack
(Armbrust et al.2009). SaaS applications differ from traditional IT delivery inasmuch
as they are designed for multiple tenants (i.e., user organizations) that share the same
underlying infrastructure (Cusumano2010). Economically, this often correlates with
a subscription-based pricing model as opposed to a perpetual-use pricing model for
traditional applications (Choudhary2007b).
In order to reach a broad market, many SaaS offerings are designed for web-based mass

166
 5.3 A Process Model for Explaining Governance of Software as a Service

customization, making it easier for user organizations to adopt and adjust the application
to their own company-specific needs (Sun et al.2008). This in turn is likely to have an
effect on internal governance structures, as some anecdotal evidence suggests (Yanosky
2008). Once business departments can source new software virtually on a mouse click and
practically without upfront capital investment, it becomes harder for IT organizations
to justify a man in the middle staffing for SaaS applications (Golden2010). Thus, the
SaaS-based delivery model is about to defy the conventional logic behind centralized and
decentralized governance.

Contingency Factors of SaaS Governance

Empirical work suggests that firms allocate responsibility for the same SaaS application
in different ways (Winkler et al.2011a). These authors operationalize application gov-
ernance by two variables capturing the decision as well as the execution level: decision
authority and task responsibility. Both variables can be either allocated to business,
the IT department or an external services provider. Furthermore, their work draws on
previous contingency theories (Sambamurthy and Zmud1999) and a grounded theory
analysis of four cases to propose a number of factors that influence in the allocation of
SaaS governance. The following five factors will also be used in course of this research
to develop our process approach:
Corporate governance comprises the degree of managerial autonomy and the strategic
IT goals, which can be either efficiency- or growth-oriented. Firms with higher autonomy
in the business units are expected to be more inclined to allocate SaaS authority to
business. The influence of strategic IS goals has been ambiguous. While IT governance
literature suggests that efficiency-oriented IT goals generally correlate with more central-
ized autonomy (Brown1999;Peterson et al.2000;Weill and Ross2004a), some evidence
suggest that this is not necessarily the case for SaaS applications (Winkler et al.2011a).
Absorptive capacities in this context refer to business and IT knowledge. The more
IT knowledge the business organization has absorbed, the more likely it is to take over
application governance. Reversely, the more business knowledge IT employees possess,
the more likely they are to govern the application (Winkler et al.2011a).
Initiative characterizes the part of the organization (either business or IT) that brings
up the idea for, and is driving the implementation of the application. It is proposed that
the initiating party is also more likely to take over application governance (Winkler et al.
2011a).
Specificity refers to the degree of adapting the application to company-specific require-
ments. For SaaS, this typically takes place through customization (Sun et al.2008). High
specificity is reflected in the degree of integration with the existing application landscape
as well as with the amount of training required for the users of that application. Therefore
it is proposed that a higher specificity also demands more IT involvement in application
governance (Winkler et al.2011a).
Finally, the scope of use measures whether an application is used by the whole company

167
5 Innovation Adoption and IT Governance in Enterprise Information Systems

or only a small fraction of employees (Winkler et al.2011a). Drawing on the rationale

of economies of scale, a wider scope of use is expected to correlate with more centralized
application governance.

Variance and Process Theories

The contingency model presented in (Winkler et al.2011a) follows a variance-theoretic

paradigm. This class of research seeks to provide empirical associations based on the
levels of an outcome (here: allocation of application governance) and its potential pre-
dictor variables (Newman and Robey1992). While possessing the strength to aim for
more generalizable results, variance-based approaches do not explain how the outcomes
exactly occur (Lyytinen and Newman2008). Process theories are a complementary
alternative which focuses on sequences of events over time in order to explain how and
why particular outcomes are reached. Thus, the outcomes become at least partially
predictable from the knowledge of the process, not from the level of predictor variables
(Mohr1982).
Process theories provide a vocabulary which is apt to study the phenomenon of interest
(Lyytinen and Newman2008). When integrating factors from variance theories in such
vocabulary, however, one should be cautious. Factors should not be understood as
predictors of certain events (e.g., the degree of specificity of the system causes more work
on system integration), but rather as a social action that helps to produce the outcome
of interest (i.e., the activity of specifying the system is followed by system integration)
(Newman and Robey1992). In this study, we bear in mind these fundamental differences
when connecting process models with factor models.

5.3.4 A Process Model for SaaS Adoption and Governance

In the following we propose a model to analyze the adoption process of SaaS applications
with a special focus on explaining the arrangements regarding the governance outcomes
of that application. The model comprises elements that define the phases, states, rela-
tionships, actors, and domains of governance factors in SaaS adoption.

Phases

Several approaches have been taken to describe the phases in the adoption of enterprise
systems (Markus and Tanis2000;Ross and Vitale2000). To structure the temporal
sequence of action regarding our phenomenon, we define five phases.
The first phase of the model refers to antecedent conditions and pre-decision activities.
Antecedent conditions are important for any process theory. They refer to the context
and historical relationships, which are essentially the outcome of a history of prior ac-
tivities likely to affect subsequent events (Newman and Robey1992). We also aggregate
relevant activities here that occur prior to the decision for implementing a certain SaaS
application.

168
 5.3 A Process Model for Explaining Governance of Software as a Service

Second, the decision phase refers to activities and events that are related to the decision for
the SaaS application, such as evaluating vendors and preparing the implementation
project. This largely correlates with the project chartering phase in (Markus and Tanis
2000). The third phase is the implementation itself (Ross and Vitale2000). It typically
comprises a number of activities related to specifying and customizing the SaaS solution
as well as rolling it out to the organization. In (Markus and Tanis2000) this is simply
referred to as the project.
Any implementation project is followed by application operation and system use,
denominated as the assimilation phase. Assimilation in this sense refers to the pro-
cess in which the application is becoming a routinized element of the firms activities
(Armstrong and Sambamurthy1999). Finally, we aggregate a future phase capturing
such developments prospected to occur induced by the current use of the system.

States

As indicated, our process model follows the goal to describe the sequence of action that
takes place in these phases. The choice on how to discretize this sequence is ultimately a
question of the conceptualization of change (Boudreau and Robey1999). While the
radical view describes change as revolutionary punctuations followed by episodes of
stability, the incremental view suggest that change rather occurs as a sequence of small
evolutionary adjustments. The punctuated equilibrium view combines elements of both
views, stating that change can alternate between both forms (Boudreau and Robey1999).
We adopt the latter view and define four types of states: events, decisions, episodes and
actions. Events and decisions represent punctuations which can either follow episodes
of stability or concrete actions of small incremental change. For example, the decision
to use SaaS is a punctuation within the SaaS adoption process. It can be followed by
a series of actions to implement that application, thus causing incremental change. The
use of that application can be seen as an episode of stability that, however, may lead to
further socio-technical changes.

Direct and Indirect Relationships

The relationships between these states are directed and characterized by temporal and
causal dependencies. We differentiate between direct and indirect relationships. Direct
relationships exhibit a clear temporal sequence and causal dependency, and thus can
also be regarded as transitions that form the process. For example, the decision for SaaS
(state A) leads to the action of making a contract with the SaaS provider (state B).
This refers to a counter-factual understanding of event causality, if A had not occurred
B would not have happened (Kim1999).
An indirect relationship can be regarded as a weaker causal dependency. For example,
the decision for SaaS (state A) is one of the reasons for an IT representative to leave the
firm (state C). Here, causality is used in a probabilistic way, A increases the likelihood

169
5 Innovation Adoption and IT Governance in Enterprise Information Systems

of C to happen, however, C could also have occurred without the event A and vice versa
(Kim1999). Regarding the sequence of action, the time between two indirectly related
states may be longer.

Actors

Most process theories relate the states to different categories regarding the outcome.
For example, a social process model on system development maps each event to any
of the three outcomes of acceptance, equivocation, or rejection (Newman and Robey
1992). However, as our change process is less concerned with success outcomes, but with
the question of governance between business, IT and the external provider, our mapping
relates to the actors. For each state it defines the actor who is mostly concerned with the
respective decision, event, episode or action. This does not exclude hybrid mappings, e.g.
to business and IT parallely. Graphically this can be illustrated by the use of swimlanes
and overlapping boxes.

Domains of Governance Factors

AsBoudreau and Robey(1999) note researchers must specify the actual content of
theory, i.e. the elements that are connected with each other within the theorys logic.
We relate the process states to the factors that are hypothesized to influence in the
governance of SaaS applications (see5.3.3). These factors are per se scaled to different
dimensions. Therefore, we widen their notion to factor domains, or second-order factors
asLyytinen and Newman(2008) suggest, which abstract from these narrow dimensions.

5.3.5 Empirical Illustration of the Process Approach: Two Cases

The purpose of this section is to demonstrate the applicability of the proposed model to
study governance phenomena by analyzing two cases of SaaS adoption.

Case Selection

The case material presented here has been drawn from a previous study. SaaS user
organizations were drawn from a customer references sites and contacted formally (see
Winkler et al.2011afor the detailed sampling strategy). Several interviews have been
conducted, transcribed, and complemented by secondary material such as company re-
ports and press clippings (Winkler et al.2011a).
Out of this collection, we chose to compare two cases which exhibit similarities in
variables external to the model (e.g., size, industry, and application type), and a strong
variance in the outcome variable (i.e., SaaS governance). The two companies chosen are
both large and internationally operating, German manufacturing firms that have adopted
the wide-spread SaaS solution [Link] (SF) for customer relationship managment
(CRM). Company B has allocated decision authority and task responsibility for SF to

170
 5.3 A Process Model for Explaining Governance of Software as a Service

Table 5.9:Case key figures

Case B Case C
Machine tools manufacturing High-tech manufacturing
70 m EUR revenue 150 m EUR revenue
600 employees 1,700 employees
7 employees in IT 40 employees in IT
1 week SF implementation time 3.5 months SF implementation time (pilot)
60 SF users 150 pilot SF users, 400 global
Interviews: Sales Organizer and SF Key User Interview: Head of Competence Center CRM
(B1), IT-Application Manager SAP (B2) (C1)

the business organization whereas in company C, SF is governed largely by an special

unit within the IT organization (a Competence Center CRM). The key figures of both
companies are given in Table5.9.

Case Descriptions

In the following we compare both cases, describe the major developments throughout
the phases of SF CRM adoption and complement these with relevant quotations from
our interviewees. For the ease of comparison we keep the two-column structure.

Antecedent conditions and pre-decision phase

Company B has a long tradition in producing ma- Since its foundation in the late 90s, company
chine tools and serving customers worldwide. C had rapidly grown in an emerging high-tech
IT had formerly been a department with more segment and strongly diversified through mergers
than 20 employees, separated into applications and acquisitions.
and infrastructure management. However, during IT-wise, the conglomerate was hardly integrated.
the times of economic crisis (2003 to 2005), IT For example, sales people did not have real-time
has been gradually reduced to a small department information about stocks. Interviewee C1 tells:
of 7 employees. Regarding IT governance, this IT had a bad image before I started here, tick-
department is run as a business monarchy (Weill eting took too long, etc. Also, CRM was a burnt
and Ross2004a). Our interviewee from IT (B2) issue. Several initiatives for CRM had been at-
explains that when the board is in the drivers tempted earlier by the business and failed.
seat, the head of IT, who is positioned much below In late 2007, the company was forced to restruc-
this, only has to serve. ture and focus on core business. The new strategy
Regarding CRM, the business representative (B1) called for more global harmonization. In 2008, a
tells: By the time it turned out that we urgently new CIO was nominated to lead the new operating
needed a CRM system. We only used self-made model implementation and a corresponding ERP
solutions. For example, we exported data from initiative. The new CIO reports on board level,
the ERP to Excel files and our sales people wrote i.e. the IT governance model can be regarded as
their reports on an in-house developed software. a duopoly between C-level business and IT (Weill
Reports were then transferred via email to the and Ross2004a).

171
5 Innovation Adoption and IT Governance in Enterprise Information Systems

headquarters and read into the ERP. Every sales A CRM expert with a strong background in CRM
representation had its own database, also the sub- and business consulting (C1) was staffed to ad-
sidiaries. This caused us to set up something more dress the open issue of CRM, and to find a sup-
integrated. plementary solution to the new ERP system.

Decision phase
The business started the CRM vendor eval- The new CRM Manager started the vendor
uation. Finally three vendors were at choice, evaluation. I evaluated the classics, SAP, Siebel,
two server-based systems and Salesforce. The Microsoft and Salesforce, until it was decided that
IT raised concerns regarding data security for the we want to go for on-demand [i.e. SaaS], not on-
SaaS solution, but finally needed to make an ex- premise. Then, we went further in the area of
ception. B2 says it went back and forth who de- SaaS and rated different criteria until we said, ok
cided, and finally business has won. B1 opposes SF is what we liked best. The decision to go for
that it was only the decision of our CEO, who on-demand came directly from the IT strategy.
was at the vendor presentation. I had agreed on We had this outsourcing project and the guide-
SF beforehand with the Head of Sales, so it was line was to operate internally as few servers as
just a matter of giving the final Go. possible.
The main motivation for SF was to disburden the The reasons that spoke in favor of SF were usabil-
IT department. Besides, other criteria such as ity, support for mobile devices and foremost our multi-
language support mattered. According to CIO wanted transparent costs. Security issues B1
functionality was not decisive and cost was were not a concern, particularly not in compar- no
major criterion either: Over a period of five ison to traditional outsourcing: If you look at years
there was no major difference in total cost. the security concept of SF, I would even say that In
course of the decision for SaaS, the SF responsi- this is better than the security concepts of our ble
on IT side left the company and handed over outsourcing partners. Costs were not major cri- the
topic to our interviewee B2, who states: If terion either: Of course, at some point you are the
thing [CRM system] had been with us, my break-even, for example after four years, but we colleague
would probably not have left the com- did not calculate this scientifically.
pany that fast. The contract with SF was closed The company decided to conduct a pilot rollout of
in 2006. SF in one region (Spain) first, in order not to in-
terfere with the ongoing ERP rollout in Germany.

Implementation phase
The Sales Organizer (our interviewee) wrote In order to guide the pilot rollout, a Compe-
the technical specification together with a consul- tence Center CRM was established and staffed
tant from Salesforce. with a second CRM expert for the Spain rollout.
The next task was data migration. The business The Spain pilot was rolled out in two legal entities,
respresentative (B1) had to prepare the existing replacing a number of local databases and Excel
spreadsheets, documents and data from the ERP tools. C1 emphasizes: We worked with an exter-
to import them into SF. nal partner there who conducted workshops, doc-
At this point, the IT was not involved into the umentation, and took over customization, testing
rollout activities at all. The actual rollout activ- and user trainings. We gained a lot of experience
ities were carried out by the Sales Organizer, the by this how SF works and how its customized.

172
 5.3 A Process Model for Explaining Governance of Software as a Service

SF consultant and an external integration part- The project took about 31/2 months.
ner. B1 notes that alone we would not have been Integration with backend systems was done later,
capable of doing that. after the go-live of the ERP system in 2009
Together with the SF consultant, the system was and the global rolling of SF. The integration
customized according to the specification. B2 with the ERP caused some IT efforts since they
states: We only needed five workdays for the [the ERP team] implemented the interfaces them-
specification, not more. selves. [. . . ] Now we exchange data such as of-
The responsibility of the integration partner was fers, orders, bills, delivery receipts, products and
to program an interface to the ERP, which created prices. So there is a lot happening.
the largest effort. IT was involved here to provide One of the things that was underestimated during
adequate connectors. B2 explains: we created the rollout was the effort for change management.
a one-way interface to SAP that polls the data C1 states you need change management people
from SAP and sends it over an i-doc connector to on the project who explain things to the other
SF. Our man only provided the things that were users. This is always a very critical point, espe-
required and later took over maintenance of the cially for CRM projects. There are many things
interface that change, sales people need to disclose their
Trainings were not a major issue. B2 states that figures a thing which no sales person likes to
the business [i.e. (B1) himself] trained all the do. The issue was addressed by trainings, com-
users. That went without complications. It was munication and later developing strong key users
not more effort than for other applications, maybe in the regions and business units.
even less.

Assimilation phase
The system is currently used by 55 employees in SF was first provided to the 150 users in Spain
the sales department, a hand full of users in the and then incrementally taken into use by Ger-
production areas and by the foreign subsidiaries many, the US and other legal entities, currently
in China, the US and Italy. counting about 400 users.
First level support for these users is provided via First level support is provided by global help desk,
the classic incident management by IT, second second level requests regarding SF are forwarded
level requests for SF are then passed to the SF to the Competence Center CRM, by now a team
Key User. of three experts.
Requests for changes from the users are collected This team also decides on requests for changes
and evaluated by B1, who is also in charge of im- and implements them in SF. More than that, it
plementing them. B1 states: we have to consider understands itself as a consultant to the business.
the tight personnel situation overall and in IT. We are positioned very consulting-like here and
Existing positions have not been staffed [. . . ], so do the specification, implementation, training and
that inevitably I have to take over things which testing. Most of us also come from consulting, i.e.
are usually not part of my job description they have the business process expertise as well
Regarding involvement of internal IT he continues as the technical expertise. Therefore we are also
for SF we only have one touch point with our able to customize the system ourselves.
IT, which is the interface to SAP [i.e. the ERP In terms of the technical interfaces to SF, there are
system]. some discernable efforts also for the ERP team.
In case of special customizations, the business I guess the effort is about 1 one person-day per

173
5 Innovation Adoption and IT Governance in Enterprise Information Systems

would directly contract external partners, for ex- week. Thats just because we built this buffer-
ample for a module to print reports of onsite vis- acknowledge-database. That was programmed by
its: That was an external partner working for the ERP team, so they have quite some effort with
us, and it also went without IT. The require- maintaining this.
ments came from us, and ultimately IT was not B1 reports that tickets for SF are even increasing
involved, B1 adds. due to a certain loop for further enhancements:
Regarding future enhancements, B1 gives into The people know that you can do a lot with SF,
consideration that I would love to do more things so they to push further processes into SF. Some
in SF, but unfortunately I dont have the time for business experts are really demanding a lot.
it.

Future
Since introducing SF, further SaaS-based enter- Overall, the Head of the Competence Center
prise solutions have been used, such as document (C1) would agree that his IT people can perform
management and enterprise content management. higher value work through working with the ex-
However, those are not as widely used and as in- ternal solution. The CIO has received positive
tegrated as SF. feedback from business, which was not normal.
B1 also states: I would appreciate if I could It was just because we can react fast.
perform minor customizations also in the ERP, Based on the company-wide use and assimilation,
without going via IT. That wouldnt be of a dis- a new strategy evolved to exploit SF for further
advantage for the company. It is just the decision global harmonization. It was mid last year that
that all customizations of the ERP stay with IT we said, there is no use if the CRM templates are
for other applications this is different. different in each country. So now, every country
will get the same core template.

Processes of SaaS Adoption

We modeled the two cases of SaaS adoption according to the proposed model. The
resulting processes are depicted in Figure5.6. For space constraints we only display a
rough overview. Detailed illustrations can be found in Appendix7(pages266-267).
States are represented by rectangular shapes (diamonds for decisions). Direct rela-
tionships are depicted as solid lines and indirect by dotted lines. Furthermore, the states
have been mapped to the respective factor domains, which are also expressed by different
color shades, see Figure5.6.

Factor Interrelationships and Case Comparison

In order to obtain information about the relationships between the factors of SaaS
adoption fromWinkler et al.(2011a), we aggregate both processes according to the
factor domains. This aggregation omits relationships between states of the same domain
and therefore focuses on the direct, as well as indirect inter-domain relationships. As a
result we obtain a partially directed graph which describes the relationships for each of the
cases qualitatively, see the Figure5.7.

174
 5.3 A Process Model for Explaining Governance of Software as a Service

B External Business IT C Business IT External

Context and Context and Pre-Decision Phase

Pre-Decision Phase
IT organization IT organization
Firm History and Corporate Firm History and Spreadsheet
and governance and governance
context Governance context solutions for CRM
context context

Need for CRM Corporate Governance

Strategy Change / New IT Strategy
(Use of Consolidation of ERP
Business incl. ERP and
Spreadsheed and IT department implementation
Reengineering Outsourcing
word lists)

(1a) (1a)

Decision Initiative Several intents

Phase Vendor selection (3a) for im plementing Staffing of a CRM
(long and short (1b) (1b)
CRM Manager
list)
Absorptive Capacities
(2a)
(2)
+ Disburden IT Decision
+ Internationalization Auth. SF responsible Decision Phase
o Cost equal SaaS? (3b) leaves the
- Security (from IT)
(3a)
com pany
(3b) Vendor(long
evaluation
listing)
Decision Authority
(2b) +Usability
+Mobility
+Cost
Transparency SaaS? Initiative
Contract with SF Absorptive Capacities
o Security

Short listing and

(3a)
(5) cost evaluation
Implementation
Phase
SF Specification (4)
Contract with SF

Further staffing
Migration of old (4) CRM Com petence
Specificity
data Implementation Phase Center

SF Pilot
implementation
Spain
Workshops.
Custom izing
SF introduction (5) specification,
(SF consultant) Absorptive (6)
testing, training
(7) Capacities (Consultancy)
(6a,b) (6d)
Global SF Rollout Integration to ERP
Absorptive
Interface to ERP Capacities Developing Key Specificity
Integration with
(External Testing (6c) Training Users & Change (7)
ERP / involvement
Consultancy) Management
Assimilation Phase
Task Responsibility (8)
(8)
Specificity Dec. Auth . Scope
Assimilation Scope of (9)
First Level Second level Third level
Further Use (10) support support for SF support
Phase adaptation (Helpdesk) (Comp. Center) (SF)
(11a,b) (10a,b) System use (9)
(decision and Increasing Use
specification) (Users)
Dec. Spec.
Enhancement
Auth. Implementation of
(consulting,
(11a,b) (12a,b) enhancements
decision,
Implementation of Task Second level First level (13) (testing, training)
enhancements specification)
Authority support for SF support
(partly external)
(14a)

(12)
Future Corp.
Further process
Gov.
Future Initiative
(14b)
standardization
(core tem plate
Introduction of rollout)
further SaaS
solutions

Figure 5.6:SaaS adoption processes (overview)

For coherence, the relevant relationships have been numbered in each of the figures.
While the two interrelationship graphs are not identical, we can still identify some com-
mon paths through the graph. For the purpose of interpretation we will compare both
cases along this dominant common path.
Corporate Governance Absorptive Capacities Initiative: Corporate governance
is the starting point for both processes. For company C, the IT strategy led to staffing a
new CRM Manager who brought about special business- and IT-specific knowledge (i.e.,
absorptive capacities) and took over the initiative for the CRM project. In company B,
the business-centric overall governance mode, as well as the efficiency focus in corporate
and IT governance, led to anchoring the initiative in the business. This tendency was
reinforced by losing key IT personnel, such as the SF responsible on IT side, over the
discussion on security issues. This, we argue, led to a further shift of the initiative
towards the business.
Initiative Decision Authority: We observe that the actual decision authority over
the later SaaS operation is already manifested in the party that largely decides on the
question for or against SaaS. In case C, this is the IT department, where the final decision
for SaaS came from the new CIOs IT strategy. For case B, this is the business, so that
we assume that these two domains are closely related.

175
5 Innovation Adoption and IT Governance in Enterprise Information Systems

Case B
(1a) Firm growth increases Initiative
the need for CRM; (1b) business
Corporate monarchy drives the initiative (2) Business key user took the initiative
so that the decision for SaaS was ultimately
Governance taken by business
(5) The business defines the
new systems specification Decision
(3b)
(3a) Consolidation of the IT department Authority
and (3b) decision for SaaS (despite security (6) Customizing, training and integration with ERP
concerns) caused the SF responsible are triggered by the business initiative
from IT side to leave the company

(4) Decrease of IT staff caused the initiative

to drift further to business
(8a) Specificity (11a;b) Decisions on further adaptation
requires tasks to be taken over
even by external contractors

(7) Training improves absorptive

capacities in business
System usage leads to
Task
decisions (10a) about further
adaptation (10b) of the system
Responsibility
Absorptive The initiative (8a)
and training efforts (8b)
(9) Use leads to increased
support through business
Capacities lead to system use

Scope of Use

Case C (14a) Central IT decision authority and

(14b) increased use enabled further
global harmonization

Initiative
Corporate
Governance (3a) The decision to go for SaaS was taken
at the beginning of the initiative
(3b) and cam directly from the IT strategy
(5) The IT initiative largely Decision
(14b)
(4) New capabilities
decided on the specification,
training and integration
Authority
were staffedto IT
for the pilot rollout (6) integration with the ERP (8) After rollout, IT created
were done after the Pilot a new application support

(2) IT relaunched
the CRM initiative with
the new CRM Manager (12a;b) Decisions on enhancements
(1a) To implement the new IT Strategy,
new capabilities were required
Specificity cause further IT tasks

(1b) Several unsuccessful past initiatives

led to staffing of the CRM Manager
(13) Enhancements
(7) After customizing the solution to the system lead to
business users needed to be trained an increased use
Task
System usage leads to
decisions (11a) about further Responsibility
adaptation (11b) of the system
Absorptive (10) With increasing use,
employees issue more tickets
Capacities (9) Trained business
started to use SF
Scope of Use

Figure 5.7:Interrelationship of application governance factors

176
 5.3 A Process Model for Explaining Governance of Software as a Service

Initiative ( Absorptive Capacities) Specificity: Next, the initiative driving the

SaaS implementation strongly influences the specificity of the system. In case C the IT,
more precisely the Competence Center CRM did large parts of the customization work,
yet relying strongly on external partners for the pilot rollout. Thereby, and by staffing
new capabilities to the Competence Center, IT has also gained crucial knowledge for
future application governance. For case B, the system has been largely specified and
adapted to company requirements by the business representative and external support.
Comparing the project durations and the amount of work for training and integration,
we may assume that the degree of specificity is considerably larger in case C. Reversely,
we may also conclude, that in case B less change management and training efforts for
business users was required, due to the fact that the initiative was already directed from
the business organization.
Specificity Absorptive Capacities Scope of Use: Consequently we deduce that
through change management and training, more absorptive capacities are built on the
user side. This in turn leads to an increased use, and thus scope of use of the system.
Especially in case C we observed that system use did not occur instantly, but as a
development. This may also be related to the larger training efforts in case C.
Scope of Use Decision Authority Task Responsibility: Increasing use of the system
consequently leads to more decisions on changes and their respective implementation.
Case B shows that, in absence of internal capabilities, the task responsibility for such
further enhancements is contracted out to external partners. In contrast in case C, the IT
department is handling SF-related activities (i.e., change implementation and support)
largely on its own.
Scope of Use Specificity ( Scope of Use): At least for case C, the ongoing enhance-
ment of the system and adaptation to specific business processes can also be interpreted
as a reinforcing cycle. A higher specificity is leading to an increased use, which in turn
creates more demand to enhance specificity, as long as the demand can be satisfied.
Further indirect effects: In case C we learned that this system enhancement is also
impacting again the overall governance mode, inasmuch as a further business harmoniza- tion
is enabled. For case B we might argue that the SaaS initiative itself has triggered further
initiatives to implement SaaS for other enterprise applications.

5.3.6 Conclusion

In this work we took a process-theoretic approach to better understand the complex in-
teraction of factors that influence the allocation of authority for SaaS-based applications.
Therefore we first proposed a process model that is suitable to examine governance in
application adoption processes. Then we illustrated the applicability of the model in
two cases of companies using SaaS for CRM, and explained the different governance
outcomes.
A few conclusions can be drawn regarding the causal relationships and path depen-
dencies between the factors. First, regarding corporate governance we outlined how

177
5 Innovation Adoption and IT Governance in Enterprise Information Systems

strategic IT goals as well as the overall mode of IT governance have a bearing on the
initiative and where it is coined. The locus of initiative, as well as the decision for SaaS
as a culminating point, seem to determine at a very early stage which party is likely
to take over decision responsibility for later application operation. Thus, the initiative
emerges as a central variable that connects the overall mode of IT govance with the later
SaaS governance outcome. However, the initiative as such also interacts with absorptive
capacities. On the one hand, existing capacities influence the degree of involvement of
the parties, both business and IT, into the initiative. On the other hand, capacities are
also increased through the initiative, for example through staffing or training new staff.
We also find that application specificity and scope of use cannot necessarily be regarded as
exogenous variables. They are determined at a rather late stage of the process and
interact with variables such as absorptive capacities and the governance outcomes.
The chosen approach possesses some inherent limitations, foremost regarding general-
izability. Since we focused on the SaaS segment for CRM, these results cannot instantly
be transferred to all types SaaS applications. Also, the process modeling and assignment
to factors may not always be straightforward due to the interpretive approach taken
in this research. Finally, the sampling of two cases cannot be regarded as sufficient to
produce stable results regarding the relationships between the factors.
However, the results generated here represent valuable insights as they add a new
complementary dynamic view to the contingency model presented in (Winkler et al.
2011a). Such temporal and causal interrelationships can be particularly of interest
when advancing from a contingency model to more complex path modeling and analysis
techniques, such as structural equation modeling. An analysis using a much broader
basis of quantitative empirical data is currently underway as further research.
Furthermore, the proposed model can be regarded as a first step to conduct more
process-theoretic research in the domain of IT governance. This appears reasonable, as
governance can be regarded as a highly dynamic construct that changes throughout var-
ious IT implementation contexts. As more research and practical experiences regarding
SaaS governance accumulate, our hope is that more precise elements can flesh out the
content of the proposed model and improve its predictive power.

5.3.7 Summary

Defining the allocation of decision rights for enterprise applications is a crucial issue
in IT governance and organization design. Today, emerging delivery models such as
Software as a Service (SaaS) defy the notion of the internal IT department as the focal
point of centralized governance. Recognizing the importance of this issue, we find that
the phenomenon of SaaS governance itself is not yet well understood. Based on two
cases of SaaS adoption, we take a process-theoretic approach to investigate the complex
interaction between factors that influence in the allocation of SaaS authority. The results
suggest that some factors, such as the locus of initiative and the decision for SaaS, interact
with absorptive capacities and determine the later mode of application governance at a

178
 5.3 A Process Model for Explaining Governance of Software as a Service

very early stage. Thus, the initiative for introducing SaaS emerges as an important
intermediate variable between the overall IT governance mode and the resulting SaaS
governance outcome.

179
5 Innovation Adoption and IT Governance in Enterprise Information Systems

5.4 The Dual Role of Information Systems Specificity for

Governing Software as a Service

5.4.1 Preamble

This chapter presents an extended version of a paper that is to be presented at the

International Conference on Information Systems 2012 (seeWinkler and Benlian2012).
Enhancements in this version largely relate to the analysis presented in Section5.4.6,
which did not enter into the conference submission for space constraints.

5.4.2 Introduction

Organizations today need to manage expanding information technology (IT) landscapes

and a multitude of internally and externally operated business applications (OBrien
2005). The way that each of these subsystems is governed may certainly depend on the
type of application itself.7 For example in large firms, ERP systems are often operated
and managed by centralized competence centers within the IT organization (Krmmer-
gaard and Rose2002;Miller et al.2004), as opposed to emerging social software tools that
tend to be managed with very strong business ownership (Deans2011). Defining such
governance arrangements between business and IT stakeholders is a common challenge
for organizations, especially given the increasing amount of services that are provided
through third-parties. Although the IS literature has widely recognized the importance
of governance arrangements for the overall IT function (e.g.,Brown and Magill1994;
Sambamurthy and Zmud1999;Weill and Ross2004a), few works have addressed gover-
nance arrangements specifically on the application level. This is remarkable, considering
that IT artifacts and their structural and social embeddedness in organizational contexts are
often seen at the core of the IS discipline (e.g.,King and Lyytinen2006;Orlikowski and
Iacono2001).
Notwithstanding, the mere knowledge of the softwares category and a vague sense of
its functionality are certainly not sufficient to advance towards a general understand- ing
of governance patterns across different application types. This is mainly because
applications of the same type may still cover different functionalities (e.g., an ERP
system may include some functionalities of a CRM), and in case two applications would
exhibit a comparable functionality set, they can still be utilized by organizations in very
different ways (Jasperson et al.2005;Strong and Volkoff2010). The misfits arising
between the (often rather generic) functionality set of an enterprise application and the
(concrete) organizational context are commonly addressed either through changes of the
software (i.e., customizations) or changes to the organization itself (Sia and Soh2007;Soh
7
With the type of application we refer to the bundle of enterprise functions that are supported by
a software application. These software types are often clustered and denominated by acronyms
such as ERP (Enterprise Resource Planning), CRM (Customer Relationship Management), SCM
(Supply Chain Management), HRM (Human Resource Management), MES (Manufacturing Execution
Systems), CCC (Content, Commnunications and Collaboration Suites), etc.

180
 5.4 The Dual Role of Information Systems Specificity for Governing SaaS

et al.2000;Strong and Volkoff2010) during the process of organizational embedding

(Orlikowski and Iacono2001, p. 126). Thus, the relevant question from a research
perspective is, what are the pertinent and salient properties of the entire information
system that lead to different application management and governance outcomes during this
process?
Transaction cost theory has evolved as a valuable lens to capture the specifics of enter-
prise applicationsforemost in IS outsourcing contextsby postulating an important,
if not essential, characteristic of an IT artifact: asset specificity (Williamson1979;1981;
1985;1991;1996). According to this view, for example, enterprise applications that are
highly specific to a companys processes are less likely to be outsourced (e.g.,Aubert et al.
2004;Benlian et al.2009;Dibbern et al.2005;Loebbecke and Huyskens2006). However,
the IS literature has conceptualized and measured this construct in different ways and
attained mixed explanatory value from it. Especially the influence on the decision to
outsource remains inconclusive in many (i.e., more than half) of the past IS studies
(Karimi-Alaghehband et al.2011;Lacity et al.2011). Therefore recently some authors
have demanded better appropriations of transaction cost theory (Karimi-Alaghehband
et al.2011) or even called for theories that are endogenous to IS to explain outsourcing
governance phenomena (Lacity et al.2011).
In this work, we build on the extant literature and extend the notion of asset specificity
to embrace specificity characteristics of ensemble information systems (IS), i.e. of
the application including its organizational and technological context (Orlikowski and
Iacono2001, p. 135). We conceptualize IS specificity by three components (functional
specificity, human asset specificity and technical specificity) and study the impact of these
components on application-level governance. More precisely, we focus on the horizontal
distribution of decision authority and task responsibility between business and IT units as
a fundamental dimension of application governance arrangements.
As the question of application governance arrangements becomes more prevalent with
increasing application outsourcing to/ and sourcing from external providers (Brown
2003, p. 202), we draw on survey data from 76 organizations that use different types
of software as a service.8 The results of our analysis unveil the dual effects that different
components of IS specificity have on application governance arrangements for SaaS. We
show that greater technical embedding (technical specificity) is associated with stronger
IT authority, while greater organizational embedding (human asset specificity) correlates
with stronger ownership by business units. A mediation model as well as an extended
model (including the influence exerted by usage characteristics) support the significance
of this and other transaction cost theoretic constructs in an application governance
context. Furthermore, we provide a visualization of different application types and their
specificity similarities to provide concise practical guidance for managing heterogeneous
application landscapes.
8
SaaS refers to the use of business applications over the Internet from an external provider and is
commonly seen as the highest layer in Cloud computing (e.g.,Cusumano2010). Several authors
argue that SaaS and Cloud computing also pose new management challenges for business and IT
units (e.g.,Bento and Bento2011;Khajeh-Hosseini et al.2010;Winkler et al.2011a)

181
5 Innovation Adoption and IT Governance in Enterprise Information Systems

In the remainder of this chapter we first explain our theoretical lens (5.4.3), followed
by the research model (5.4.4), the description of the methodology (5.4.5), the analysis
(5.4.6) and a discussion of theoretical and practical implications (5.4.7).

5.4.3 Theoretical Lens

Transaction Cost Theory (TCT) puts transactions, defined as the ultimate unit of eco-
nomic activity (Commons1931), and the cost for planning, adapting, and monitoring
this activity into the focus to explain the existence of institutions, whether they are
bureaucracies, markets or other organizational forms (Williamson1985). Williamson
uses the term transaction cost synonymously with governance cost and coordination cost
and delineates these from the costs of production. According to this view, the boundary
of an organization (or an organizational subunit) extends to the point where total costs
for hierarchically coordinating transactions exceed the costs for coordinating via the
market. Variables that influence transaction costs have been defined as asset specificity,
uncertainty and the frequency of transactions, whereas asset specificity is deemed the
most important (Williamson1985, p. 52). For example, when the required assets are
idiosyncratic (i.e., highly specific) and frequently required, the costs for coordinating
these transactions internally, as well as uncertainty surrounding the transaction, are pre-
sumably lower than when a firm needs to find, contract out and perform the transaction
with a partner on the external market. Based on TCT, organization theorists have also
viewed firm boundaries as a continuous choice that may lead to diverse hybrid forms,
i.e. arrangements between the polar market and hierarchy extremes (e.g., profit centers,
Hennart1993;Williamson1991).
Assuming that boundary choices are a parallel notion to allocations of authority, TCT
may also inform governance arrangements between business and IT units. Or, in the
words ofWhetten et al.(2009), we may borrow TCT as a theoretical lens and transfer
it to the context of IT governance phenomena between business and IT units (cross-
context) on the subsystem level of single applications (cross-level). According to this
view, business units decide whether to enter into a contract with IT units (i.e., to use
the firm-internal market) or govern application operations in a hierarchical form (i.e.,
business unit internally) depending on application-related transaction costs. Analogously
toLacity et al.(2009, p. 146), we argue that the notion of a transaction, however, may
be less adequate for the IS domain, since it implies that these internal contracts have
defined entry and exit points. Information systems, in contrast, are characterized by
changing requirements and a continuous evolution (e.g., through updates and enhance-
ments) rather than discrete transactions with a certain frequency and uncertainty. This
leads us to make several appropriations when casting TCT into application governance
problems.
First, costs of application-related transactions can be best understood as coordination
cost, referring to the continuous efforts for planning, adapting, and monitoring applica-
tion operationas opposed to production cost, e.g. for personnel, software licenses
and operating technological infrastructure. In line with TCT, organizations seek for the

182
 5.4 The Dual Role of Information Systems Specificity for Governing SaaS

best governance mode of applications to economize on total cost. Arguably, coordination

costs are lower under direct coordination within business units (i.e., hierarchical coordi-
nation), while production is more efficiently coordinated by a (centralized) IT unit, due to
the possibility to leverage internal economies of scale effects (e.g., through a better
utilization of IT resources across the enterprise). However, when operating applications
at an external partyas it is the case for SaaS, the rationale of leveraging internal
production cost advantages becomes much weaker, since the economies of scale accrue on
the provider side. That is, production costs (expressed by the fees of the provider) are
largely independent from the internal coordination form. Thus the question on
application governance becomes merely dependent on the costs of coordination.
Second, the frequency of transactions, often defined as the repetitiveness of a cer-
tain type of transaction (Karimi-Alaghehband et al.2011) appears less applicable to
the context of long-lasting information systems. When introducing the frequency con-
struct, Williamson originally pointed at the amount of buyer activity in the market
(Williamson1979, p. 247), i.e. in our context the amount of business activity in infor-
mation systems use. This property, we argue, may be captured better by the scope of
use of an application within the wider organization. According to TCT, the scope of
use then moderates the relationship of asset specificity on application coordination cost
(Karimi-Alaghehband et al.2011;Williamson1981). That is, given a certain degree of
specificity, more business activity (i.e., a greater scope of use) will also increase the cost
for coordinating the management of the application. Thus, scope of use is also expected
to play a reinforcing role for the effects of specificity on application governance.
Finally, as a third appropriation, we assume that in an intra-organizational context,
behavioral uncertainty defined as strategic non-disclosure, disguise or distortion of infor-
mation (Williamson1985, p. 57) plays a minor role compared to contractual situations
with external, i.e., potentially unknown, agents on the market. Moreover, meta studies
in an IS outsourcing context (Karimi-Alaghehband et al.2011) as well as in the broader
strategic management literature (David and Han2004) document a general explanatory
weakness of this variable, which corroborates the argument to omit this variable in our
theoretical lens. The appropriation of the core construct of TCT, asset specificity, will
be treated in the following section, after we conceptualize appropriate dimensions of
application governance.

5.4.4 Research Model

This work aims to investigate the relationship of IS specificity on application governance

arrangements. We derive appropriate constructs from the extant literature, conceptu-
alize their meaning in the context of the governance of SaaS and hypothesize relevant
relationships. The overall research model is depicted in Figure5.8.

183
5 Innovation Adoption and IT Governance in Enterprise Information Systems

Information system specificity Application governance

(and use) (-Business/+IT)
Human asset
specificity
H1 (+) H3 (-) Decision
H5 (+)
authority

Functional Application
Scope of Use governance
specificity

H6 (+) Task
H2 (+) H4 (+) responsibility
Technical
specificity

Notes: Extended model variant with dotted lines; Second-order construct with doubled lines

Figure 5.8:Research model

Application Governance

The notion of IT governance has been shaped by researchers and practitioners and
evolved over the last two decades. Departing from the discussions on the centralization
versus decentralization of the IT function, the early IT governance literature has pri-
marily focused on the distribution of IT decision rights between divisional and corporate
IT units (e.g.,Brown1997;Brown and Magill1994;Sambamurthy and Zmud1999;
Weill and Ross2004a). Lateralso influenced by the increasing pervasiveness of best-
practice governance frameworks such as ITIL and COBITauthors emphasized the
importance of complementary mechanisms to implement such governance arrangements,
such as certain roles, defined processes, and service level arrangements (e.g.,Peterson
et al.2000;Van Grembergen2004;Weill and Ross2004a). While the past literature has
primarily viewed IT as a bulk function, recently also more fine-grained conceptualizations
of governance arrangements have been proposed, such as for data governance (Khatri and
Brown2010), development project governance (Tiwana2009), infrastructure sourcing
governance (Xue et al.2011), and application governance (Winkler et al.2011a).
Analogously to IT governance, application governance is concerned with the design of
decision rights and complementary mechanisms to deliver expected value from the use of
business applications. One of the most fundamental dimensions of application governance
is certainly the degree to which application-related decisions are shared horizontally
within the organization, i.e. between business and IT units (Winkler et al.2011a). 9 This
dimension is fundamental in the sense that organizations would probably first assign the
major accountabilities for managing a business application, before deciding on detailed
governance processes, responsibilities, performance measurement and charging schemes.

9
Note, that this dimension delineates from the vertical distribution or decision rights, e.g. across C-level,
senior management, middle management and staff roles (Weill and Ross2004a).

184
 5.4 The Dual Role of Information Systems Specificity for Governing SaaS

Conceivably, such internal agreements may be especially valuable also in an outsourcing

context, where the view of only two stakeholdersthe client firm and the supplier
firmis too simplistic (Lacity et al.2009, p. 149).
The agency theoretic imperative to separate decision control from decision implemen-
tation (Fama and Jensen1983) suggests the existence of two classes of application-
related decision rights, which have been introduced elsewhere as decision authority and
task responsibility (Winkler et al.2011a). This is, the principal makes control decisions
regarding the management of an application which are to be implemented by an agent
(Tiwana2009). For example, as common governance frameworks suggest, application
change decisions might be agreed by a change advisory board (comprising business and
IT stakeholders) and implemented by a change manager on IT side (TSO 2011). In our
theoretical model, however, we consider a bundle of decision rights as well as a bundle
of actors (i.e., business and IT units). Thus, in the given example, we might find mixed
decision authority on control level and greater IT authority on implementation level.
Therefore, both of these dimensions can be conceptualized as a continuum with greater
ownership by one or the other side, rather than a discrete set of allocations.
Generally, there may be a certain delta between decision authority and task responsi-
bility (e.g., business units are likely to have more decision control rights, while IT units
possess more implementation decision rights). However, we argue, that these dimensions
are still highly correlated (i.e., the delta is not likely to change substantially in magnitude
or sign across cases), so that both dimensions together ultimately reflect an overall
application governance [Link] and Ross(2004b) make a similar argument.
They define seven rather sophisticated patterns for distributing IT decision rights hori-
zontally and vertically within an organization (e.g., business monarchies, IT monarchies,
duopolies, federal, feudal and anarchy patterns). However, finally they conclude that all
of these patterns can be ultimately arranged on a single centralization/decentralization
scale. Thus, we may conceptualize the two subdimensions as first-order factors of an
overall second-order application governance dimension (Edwards2001, p. 146).

Decision Authority Decision authority captures the distribution of application-related

decision control rights between business and IT stakeholders. Over the lifecylce of a busi-
ness applicationi.e., from the requirements analysis, design and/or vendor selection,
building and/or deployment, operation, optimization and phase out (van der Pols2004)
a multiplicity of single decisions (e.g., on the initial investment) and recurrent decisions
(e.g., on regular change approvals) accrue. For example,Weill and Ross(2004a;b) use
six classes of overall IT decision rights that refer to IT principles, IT infrastructure,
IT architecture, business application needs, as well as IT investment and prioritization.
Arguably, from these classes the latter three appear most applicable to the application
level (moreover in a SaaS sourcing context). Similar to the overall application governance
dimension, we argue that the allocation of such decision rights is strongly correlated, so
that they can be ultimately reduced to a single dimension. Analogously, the allocation
of any application-related decision control right thus ultimately becomes a (good or bad)

185
5 Innovation Adoption and IT Governance in Enterprise Information Systems

indicator for the decision authority subdimension.

Task Responsibility Task responsibility refers to implementing application-related

decision rights, or simply, to the execution level. Thus, this dimension is concerned
with who is doing the requirements analysis, the design and/or vendor selection, the
building and/or deployment, operation, optimization and phase out (van der Pols2004).
Arguably, for conventional IT applications (meaning those that are operated in-house), the
scope of task responsibility is fairly wide and may include application development and
integration, infrastructure operation as well as user support tasks. However, for
externally operated applicationssuch as SaaSfew activities remain with the client
[Link] et al.(2010, p. 93), for example, propose a task structure for SaaS
CRM services that consists of seven tasks, out of which only two (mapping SaaS to
business requirements, and streamlining customer-facing activities) are performed by
the client organization. In our lifecycle framework these categories would largely relate to
requirements analysis for initial deployment and ongoing application optimization (i.e.,
implementing changes). Furthermore,Winkler et al.(2011a) point out that for many
SaaS offerings end-user support is an activity that remains with the client organization,
and which is not allocated to IT units by default. Analogously to decision authority,
we argue that such task responsibilities are likely to be strongly correlated. Therefore,
any allocation of any single task responsibility would as well represent an indicator for
this subdimension.

Information System Specificity

In line with Williamsons (1996, p. 105) definition of asset specificity, we define IS

specificity as the degree to which an IS can be redeployed to alternative user organi-
zations without sacrifice of its productive value. When developing the TCT, Williamson
emphasized that specificity can derive from various categories of assets, foremost the
geographical location of an investment (site specificity), the employees knowledge, ex-
pertise and learning (human asset specificity) and the specialization of equipment and
tools (physical asset specificity) (Williamson1979;1981).
IS researchers have mostly aggregated different facets to a single dimension, when
applying TCT to study IS phenomena (seeKarimi-Alaghehband et al.2011for a sys-
tematic review). For example, in a study on SaaS adoption,Benlian et al.(2009)
operationalize application specificity as the degree that specific applications can be
customized, integrated, and modularized prior to and in the outsourcing relationship.
Dibbern et al. (2005) focus on human resource specificity to study a similar phenomenon and
operationalize this construct by four subdimensions (unique business knowledge, unique
software knowledge, social collaboration IS/user, social collaboration within IS). In a
study on netsourcing,Loebbecke and Huyskens(2006) consider items related to
technical specificity, site specificity, and human capital [Link] et al. (2004)
study outsourcing of diverse IT functions and possibly use one of the broadest
operationalization of asset specificity with 24 indicators relating to categories such as

186
 5.4 The Dual Role of Information Systems Specificity for Governing SaaS

client investment, human resource (HR) specificity, HR hiring delay, HR trainings delay,
supplier investment, [and] structural liaison devices.
The explanatory value that IS researchers have obtained from this variable, however,
remains mixed. More precisely, meta reviews demonstrate that in more than half of
the studies the construct does not produce the hypothesized relationships (Karimi-
Alaghehband et al.2011;Lacity et al.2011). In the given examples,Benlian et al.(2009)
as well asLoebbecke and Huyskens(2006) find no significant relationships,Dibbern
et al.(2005) obtain partly significant relationships, andAubert et al.(2004) find a
significant opposite hypothesized relationship regarding their multi-faceted construct.
The authors partly explain the low impact of asset specificity by the definition and
operationalization of the variable (Loebbecke and Huyskens2006, p. 421), and mea-
surement problems (Aubert et al.2004, p. 929). We put forward that IS specificity can
be decomposed into three major factors referring to functional, human, and technical
categories. This conceptualization is based on the idea that an IS is gradually made
specific while it is embedded in its organizational and technical context, so that the
facets of IS specificity mutually influence each other.

Functional Specificity The core role of an enterprise IS is to provide certain functional-

ities that typically map to one or several business processes of an organization (OBrien
2005). Since organizations are inherently different, packaged applications, such as ERP,
typically do not perfectly match the organizations process, data and user requirements
(Sia and Soh2007;Soh et al.2000;Strong and Volkoff2010).Strong and Volkoff(2010)
emphasize that two different kinds of misfits can arise: deficiencies which result from
required functionality that is actually missing, and impositions that arise from addi-
tional functionality that organizations implement to enable new work practices (best
practices). The former (deficiencies) are typically overcome by adapting the application
to the organization, which can vary strongly in the degree and effort put into [Link]
(1998) differentiates customization (or configuration), extension, and modification. We
pose that the degree of adaption to company specific requirements directly expresses the
degree of specificity of the overall IS inasmuch as a highly customized system according
to our definition can be less easily redeployed to an alternative user organization without
sacrifice of its productive value.
Important to note, customizations are not limited to conventionally deployed IS. Al-
though SaaS applications inherently comprise a very standardized set of functionalities,
vendors increasingly provide ways to enable self-service adaptation and extension, e.g.
through web-based configuration (instead of editing local configuration files), use of
components from platform markets (instead of locally installing add-ons) and integration
through standard application programming interfaces (APIs) (Bezemer and Zaidman
2010;Sun et al.2008;Xin and Levina2008). For examplejust like in any other
software implementation projectSaaS consultants typically write up a requirements
specification document, which can take few days or up to several months depending on
the specific case (Winkler and Gnther2012). We argue that this degree of functional

187
5 Innovation Adoption and IT Governance in Enterprise Information Systems

specificity has both organizational and technical implications. As our first hypotheses,
we pose:
H1: Higher functional specificity of an IS leads to greater human asset specificity, and
H2: Higher functional specificity of an IS leads to greater technical specificity.

Human Asset Specificity According toWilliamson(1981, p. 555), human asset speci-

ficity is associated with an organizations work routines and primarily arises from learn-
ing by doing. For example, although programming skills may be improved by practice,
they are still regarded as less specific inasmuch as they are equally valued by the current
and potential other employers. In contrast, the knowledge of a firms business process is
regarded as highly specific as it cannot be instantly transferred to other organizations.
Functional specificity may lead to greater human asset specificity (H1), since the work
routines that are created in overcoming information system misfits, i.e. through both
customizing the application to own routines and enabling new routines, will also require
greater firm-specific knowledge. In other words, in the process of specifying and customiz-
ing new software to an organization, those additional functionalities that are added to the
software also require greater knowledge and skills from users. This knowledge comprises
knowledge of unique business processes and application software that is specifically
customized to a company (Dibbern et al.2005). Such skill requirements could be
reflected, for example, in the amount of user trainings given and the need to specialize
and educate staff for managing this application.
The degree of human asset specificity of an information system (i.e., the application
including its users and personnel responsible for its management), in turn, is likely to
affect the mode of application governance. The more company-specific knowledge is
incorporated in an information system, the more involvement of those organizational
units is expected that make use of the information systemwhich are commonly the
business units. Or, from a TCT view, highly specific knowledge needs to be embedded
in a protective governance structure to economize on transaction costs for acquiring and
maintaining this knowledge (Williamson1981, p. 563). Altogether, this argument im-
plies that more customized and functionally specific applications cause higher IS human
asset specificity and therefore more business involvement in application-related decision
authority and task responsibility. Accordingly we pose that
H3: Higher human asset specificity of an IS leads to greater business unit application
governance, and
H3: Human asset specificity mediates the effect of higher functional specificity on greater
business unit application governance.

Technical Specificity Technical specificity has been considered explicitly by few IS

researchers (e.g.,Loebbecke and Huyskens2006) and has thus so far only been weakly
conceptualized. However, we may relate this dimension to physical specificity, generally
referred to by (Williamson1979) as the use of special equipment or tools in an organi-
zation to produce certain goods or services. Increased functional specificity of a business

188
 5.4 The Dual Role of Information Systems Specificity for Governing SaaS

application may also lead to higher technical specificity (H1) since some functionality may
require a special technical implementation. For example, in the case of a SaaS CRM
implementation, a company built a special buffer-acknowledge database to exchange
data, such as offers, orders, bills, delivery receipts, products and prices, in both ways
between the CRM and the ERP systems (Winkler and Gnther2012). In the view
of TCT, such technical integration approaches reduce the possibility of the IS to be
redeployed to alternative user organizations.
For conventionally deployed information systems, technical specificity may refer to a
number of different categories, e.g. the need of specific custom programming, databases,
and server infrastructure. In a SaaS context, where the gross of technology is outsourced, the
technical dimension largely refers to the integration with other backend systems
(Benlian et al.2009;Bezemer and Zaidman2010;Sun et al.2007). Thus, for SaaS,
we might also term this construct simply as integration specificity. Arguably, higher
technical (integration) specificity might be related with stronger application governance
through IT units to minimize internal coordination costs. For example, in the given
case the technical interfaces created some discernible efforts also for the ERP team
(Winkler and Gnther2012). Altogether this adds an argument to our conceptual model
that greater functional specificity also increases technical specificity, which in turn leads
to more IT involvement in application-related decision authority and task responsibility.
Accordingly we pose
H4: Higher technical specificity of an IS leads to greater IT unit application governance,
and
H4: Technical specificity mediates the effect of higher functional specificity on greater
IT unit application governance.

Scope of Use

Scope of use relates to the frequency construct in TCT and can be understood as
the breadth to which a business application is used within an organization. This scope
may be expressed by the share of users of an application or simply by the amount of
organizational units in which it is utilized (e.g., on a scale from single-department to
company-wide use). This construct does not relate to any of the six categories of system
usage provided byBurton-Jones and Straub(2006), since it captures lean system usage
characteristics on the organization, rather than on the individual level.
As (Karimi-Alaghehband et al.2011, p. 135) note, some IS studies that use TCT as a
theoretical lens do not account for the frequency construct and/or its interaction effects. In
line with TCT, we assume that scope of use acts as a moderator (or more precisely
multiplicator) of the effects of functional specificity on coordination cost imposed by
both human asset and technical specificity. This is, while functional specificity leads
to according human asset and technical specificity, the strengths of these relationships
are influenced by scope of use. Consider the examples from above: a highly customized
system will require more user trainings, whereas the need for such trainings logically still

189
5 Innovation Adoption and IT Governance in Enterprise Information Systems

increases with the number of users that actually need to be trained. Same applies to
technical specificity: the highly customized CRM system causes certain initial as well as
continuous integration efforts, i.e. technical specificity. These efforts would be arguably
even higher, the more employees use the system, e.g. since more data needs to be handled and
more change requests turn up. Accordingly we derive that
H5: Scope of use moderates the relationship of functional specificity and human asset
specificity.
H6: Scope of use moderates the relationship of functional specificity and technical speci-
ficity.

5.4.5 Methodology

To test the proposed research model, we used data from a survey with 76 organizations
that provided information about a SaaS application in use. In the following we describe
our approach for the development of the measurement instrument and acquisition of the
sample.

Instrument Development

Our approach for developing the measures has been oriented in the procedure propounded
by (Churchill1979). The domains and subdomains of application governance and IS
specificity had been initially described and specified in course of a number of explorative
case studies presented elsewhereWinkler and Gnther(2012). We compared these case
findings with the extant literature and derived a number of items for each of the six
presented constructs (plus further constructs not included in this research). In this
manner, items for functional specificity (Benlian et al.2009), human asset specificity
(Dibbern et al.2005) and decision authority (Weill and Ross2004a) have been derived
from literature, while items for technical specificity, task responsibility and scope of use
can be regarded as entirely new measures. In order to validate these measures, we first
conducted a category sorting with 8 fellow researchers, who had to assign the items
according to given construct definitions, and eliminated ambiguous measures.
Second, we conducted an online pretest with 29 IT professionals from industry, con-
sulting and research, which resulted in acceptable construct reliabilities (alpha>0.8).
Due to size constraints for the final questionnaire, we needed to further reduce the
number of items to 3 for five-point-scaled and 2 for seven-point-scaled items. Five-point
scales were used for business/IT application governance items (question: who decides /
who is responsible; scale anchors: busines, business with IT involvement, business and
IT equally, IT with business involvement, central IT) due to their adequate semantic
granularity discussed in related works (Brown and Magill1994;Winkler et al.2011a).
Seven-point scales (very low, low, low-medium, medium, medium-high, high, very high)
were used for the other measures. The final questionnaire has been pretested in think-
aloud meetings with 3 different CIOs and passed only minor revisions. The measurement

190
 5.4 The Dual Role of Information Systems Specificity for Governing SaaS

items are provided in Appendix8(page269.

Sample Description

The survey was sent out to 2,886 large-sized companies in Germany (i.e., companies
with >50 m Euros of revenues, >500 employees, from non-public sector) accompanied
by a formal invitation, a return envelope, as well as references to the online version of
t7he questionnaire and the project website. Relevant addresses and CIO contacts had
been retrieved from a commercial publisher of company information. Besides providing
the survey results, we were able to offer a small gadget for the first 100 respondents
and the drawing of a tablet computer as further incentives. During a six-week response
period in April/May 2011, we received feedback from total 534 companies, out of whom
220 provided usable responses (60% online, 40% paper-based, total response rate 7.6%).
In the main part of the questionnaire, the respondents were asked to provide details
about one business application they are knowledgeable about. 131 of these respondents
provided information about a conventional application, and 76 about a SaaS in use (13
were discarded from further analyses due to unclear application types or inconsistent
answers). Regarding the SaaS subsample, respondents stated to be CIOs (27.6%), senior
IT managers (39.5%), IT managers (23.7%), IT staff (2.0%), and senior business man-
agers (6.6%) with an average 12 years of work experience in their current firms (median
10 years). T-tests revealed no response bias in the model variables neither regarding
business versus IT roles, nor regarding early versus late responses. The respondents
provided details about a wide range of common SaaS application types, in line with the
categories introduced earlier7, these can be characterized as ERP (13), CCC (12), CRM
(10), SCM (6), HRM (5), MES (4), business intelligence and analytics (6), office and
productivity (5), service management (3), and other (12) application types.

5.4.6 Analysis

For the purpose of model tests we employ partial least squares (PLS), a widely used
structural equations modeling technique. The choice for PLS was primarily motivated
by the explorative character of this research including various new constructs as well
as the good ability of PLS to handle moderating effects (Chin1998b;Hair et al.2011).
In order to isolate the effects constituted by adding scope of use as an additional TCT
construct, we test three variants of the proposed research model separately: M1 contains
only the core constructs, i.e. functional specificity (FuncSp), human asset specificity
(HumSp), technical specificity (TechSp), as well as application governance (AppGov),
decision authority (DecAuth) and task responsibility (TaskResp) and suffices to analyze
the core model including the hypothesized mediation effects. M2 integrates the scope of use
(Scope) and its direct effect on HumSp and TechSp. M3 uses a product indicator
approach (Henseler and Chin2010) and includes the interaction terms (i.e., the Carte-
sian product indicators of ScopeFuncSp and ScopeTechSp) to test the hypothesized
moderating effects. Significances were assessed based on the pseudo t-values from a

191
5 Innovation Adoption and IT Governance in Enterprise Information Systems

bootstrapping procedure with 1,000 re-samples. We follow the recommended approach

to assess the measurement model and structural model separately (Chin1998b).

Measurement Model

We use M2 to assess the psychometric adequacy of our measurement instrument since this
model variant contains all relevant indicators (i.e., includes Scope). Convergent validity
of the first order constructs is supported by acceptable values for Cronbachs alpha
(alpha>0.8), composite reliability (CR>0.7) and average variance extracted (AVE>0.5),
see Table5.10(Hair et al.2011). Furthermore, all indicator loadings are significant and
clearly over the threshold of 0.7, see Appendix8(269). This suggests that the indicators
are sufficiently correlated and reflect the properties of their substantive constructs. For
the reflective second-order construct (AppGov), convergent validity can be assessed by
calculating the AVE of its subdimensions (MacKenzie et al.2011, p. 315). With two
highly significant path coefficients to DecAuth (.920, R2=.847) and TaskResp (.936,
R2=.877), see Table5.11(page193), this second-order AVE amounts to highly an
acceptable AVE of .862.

Table 5.10:Measurement model validity

Operationalization Convergent
Discriminant validity
validity ( AV E on diagonal)
Construct Items Dim Scale Alpha CR AVE 1 2 3 4 5 6
DecAuth 3 B/IT 1-5 .836 .902 .754 .868
TaskResp 3 B/IT 1-5 .915 .946 .854 .724 .924
FuncSp 2 /+ 1-7 .878 .943 .891 .229 .095 .944
HumSp 2 /+ 1-7 .885 .946 .897 .042 .011 .533 .947
TechSp 2 /+ 1-7 .830 .922 .854 -.008 -.010 .486 .731 .924
Scope 2 /+ 1-7 .779 .889 .802 .506 .426 .543 .310 .342 .895

1: Decision authority; 2: Task responsibility, 3: Functional speciftcity; 4: Human asset speciftcity, 5: Technical speciftcit y; 6: Scope
of use

We analyze discriminant validity by the Fornell-Larcker criterion, indicator cross-

loadings, and an exploratory factor analysis (EFA).Fornell and Larcker(1981) suggest
that the squared latent variable correlations should be less than AVE for each construct
to ensure support validity, which is the case for our model, see Table5.11. Indicator cross-
loadings are as well lower than loadings by their substantive constructs, see Appendix8
(page269). However, we note that some loadings within the application governance do-
main, as well as within the IS specificity domain are above the threshold of 0.7, indicating
thatstatisticallythese cannot not necessarily be viewed as separate dimensions. After
varimax rotation, the EFA of our model variables (eigenvalue>1; KMO=.816) yields
three clearly distinguishable factors relating to the indicators of application governance
(30.8% of explained variance), IS specificity (24.15%), and scope (17.49%). This supports
our nomological framework inasmuch as we stated that both DecAuth and TaskResp are
subdimensions of application governance. Likewise, FuncSp, HumSp and TechSp can be
viewed as components of overall IS specificity (which, based on our particular research

192
 5.4 The Dual Role of Information Systems Specificity for Governing SaaS

interest, has not been explicitly modeled here).

Finally, we assess common method bias (CMB) by using a latent method factor
approach for PLS. Especially in self-reported data where independent and dependent
variables stem from the same source, CMB may arise due to the subjects motif to give
consistent and/or socially desirable answers (Podsakoff and Organ1986). Following the
procedure described byLiang et al.(2007, pp. 85-87), we add a method factor to the
model to account for this bias and compare the average variance presented by this factor
with the variance presented by the substantive constructs. To account for each indicator
only once, we omit the reflective first-order constructs and test the method influence
directly on the six indicators of AppGov. The method factor accounts for .009 of the
average variance while the substantive constructs explain .799 (ratio 1:88), suggesting
that CMB is unlikely to affect our results.

Table 5.11:Model tests

H Path M1 M2 M3 Sup-
(mediation model) (including Scope) (moderated med.) port
coeffa f2 coeffa f2 coeffa f2
FuncSp AppGovb -.010ns -.015ns -.015ns
-.017 -.002 -.002
H1 FuncSpHumSp .736 .691 .415 .559 .474 Sup.
H2 FuncSpTechSp .541 .404 .141 .319 .167 Sup.
H3 HumSpAppGov -.105ns .008 -.121 .006 -.120 .011 Part.
H3 FuncSpHumSp
c
-.077ns .878 -.084 .852 -.067 .819 Part.
AppGov
TechSp AppGov .235 .229 .228
H4 .042 .038 .040 Sup.
H4 FuncSpTechSp
c
.128 1.086 .094 1.186 .073 1.258 Sup.
AppGov
Scope HumSp .128 -.031ns
.007 .066
Scope TechSp .419 .313
.150 .175
H5 ScopeFuncSp .338 .058 Sup.
HumSp
H6 ScopeFuncSp .222 .025 Sup.
TechSp
AppGovDecAuth .922 .920 .920
AppGovTaskResp .937 .936 .936
a
p<.05; p<.01; nsnot significant
b
direct path for comparison and VAF calculation, not included for further model tests
c
total mediated effect, significance according to Sobel test (1982), VAF reported instead of f 2

Structural Model

The results of the structural model tests are provided in Table5.11. Path coefficients
and significances can be interpreted similar to coefficients in a simple regression; effect
sizes f2 express the relative change in the explained variance of an endogenous construct
when eliminating one of its antecedent variables from the model (Chin1998b, p. 317).

193
5 Innovation Adoption and IT Governance in Enterprise Information Systems

Analogously to regression analyses, effect sizes of 0.02, 0.15, and 0.35 can be regarded
as small, medium, and large, respectively (Cohen1988, p. 421). For the mediation
effects, we report the specific indirect effect (i.e., the product of the path coefficients
of the mediation) (Hayes2009, p. 3) and its significance assessed by aSobel(1982)
test statistic; in the effect size column, we alternatively report the variance accounted for
(VAF) of the specific indirect effect relative to the specific total effect (i.e., the total effect
excluding the opposite mediation effect).10 In simple mediation models, values for VAF
range from of 0.0 (no mediation) to 1.0 (perfect mediation), while values greater than
1.0 demonstrate a suppressor effect (Shrout and Bolger2002, p. 430). In the following,
we will evaluate our hypotheses successively.
Worth to begin with, we find that the direct effect from FuncSp to AppGov is close
to zero and not significant (for none of the model variants). The effect size of FuncSp is
even negative, which shows the rare case that excluding this construct from the model
even increases explained variance in AppGov. (This will be explained more in detail in
the following.)

HumSp and FuncSp TechSp are both strongly significant, which
confirms H1 and H2 and underlines the nomological argument made earlier that these
constructs are ultimately dimensions of an overall IS specificity property. When including
Scope and interaction variables in M2 and M3, we note that the strength of both paths
decreases only moderately so that support for these hypotheses is sustained. Accord-
ingly, the effect sizes of this construct compared to the influence exerted by Scope are
considerably strong.
We find partial support for H3 and the path HumSp AppGov. This is, in the simple

mediation model (M1) the path is negative as hypothesized, yet, not significant at the 0.05
level (t=1.573). However, when adding Scope and the interaction variable to the models
M2 and M3, the path coefficient increases slightly and becomes statistically significant
(t=1.777). We take this as evidence that including the Scope variable increases the
overall explanatory power of the model, although the effect sizes presented by HumSp
remain extremely weak. Similarly, we find that HumSp significantly mediates the effect
from FuncSp through AppGov (H3) when Scope is included into the model (M2 and
M3). The high VAF ratio (>0.8) expresses that this mediation effect is comparably
strong to the non-existent direct effect (FuncSp AppGov).

Analyzing the role of TechSp, we find also support for our hypotheses H4 and H4.
TechSp has a clear and significant impact on AppGov across all three model variants with
small, but considerable effect sizes. Furthermore, as hypothesized, TechSp significantly
mediates the effect from FuncSp to AppGov in opposite direction to the mediation exerted
by HumSp. Thus, we find an explanation for why there is no direct effect from FuncSp
to AppGov: While HumSp negatively exerts the effect from FuncSp on AppGov, TechSp

10
Hayes(2009, p. 14) adds for consideration that the quantification of effect size in intervening variable
models remains a cutting-edge area of thinking and research where none of the methods proposed
thus far are particularly satisfying. Since total effects can approach zero and even have opposite
signs than the denominator, any VAF ratio of a mediation effect should be interpreted with caution.

194
 5.4 The Dual Role of Information Systems Specificity for Governing SaaS

positively counteracts it. For this reason the total effects (i.e., the sum of the direct
and indirect effects) are close to zero across all three model variants. Or in the words of
Hayes(2009, p. 10), we provide a case where two or more indirect effects with opposite
signs can cancel each other out, producing a total effect and perhaps even a total indirect
effect that is not detectably different from zero.
Regarding the hypothesized moderating effects and the product indicator approach,
we first add Scope as a second antecedent of HumSp and TechSp to the model (M2). This
yields in significant paths and a considerable increase in explained variance for TechSp
(f2=0.15). However, when introducing the two interaction terms (ScopeFuncSp), the
coefficient Scope TechSp is reduced and Scope HumSp
is practically absorbed. Cou-
pled with the that the interaction are significant for both relationships
finding terms
(with considerable effect sizes), we can say the moderating hypotheses H5 and H6 are
both supported, which underlines the explanatory power provided by the scope of use
construct.

Multidimensional Scaling of Application Types

To provide a more visual interpretation of our results, we perform a multidimensional

scaling (MDS) of application types and a property fitting according to the model di-
mensions. Multidimensional scaling refers to set of analytical techniques that aim to
explore multidimensional data by positioning objects in a low-dimensional space based
on their similarity (or dissimilarity) characteristics (Kruskal and Wish1978;Kappelhoff
2001). Objects in our case refer to the different application types by which respondents
classified their SaaS application (see5.4.5). Based on the standardized factor scores from
the PLS analysis, we first calculate the factor means for each of the six model variables
(i.e., FuncSp, HumSp, TechSp, Scope, DecAuth, and TaskResp) per application type, see
Table5.12. Based on these factor means, we generate six dissimilarity matrices (one per
factor) by calculating the pairwise absolute differences between the factor means. These
dissimilarity matrices are used as an input for a MDS with two dimensions, an euclidean
metric and interval scales11.
As a result, we obtain the (x, y)-values of the final MDS solution (see Table5.12) and
a scatter plot of the 10 application types (see Figure5.9). The STRESS of this final
solution is 0.177, i.e. (1STRESS)=82.3 percent of the scaled data can be accounted
for by the MDS procedure, which can be regarded as a good level of fit (Kappelhoff
2001). Trying to interpret this chart (still without the dimensional lines), we find a
cloud of almost equidistant application types in a two-dimensional space, with HRM
and other solutions at the left end and SCM, ERP, and M ES types at the other, as well
as productivity applications at the far outer right margin.
Inevitably the question arises, how these positions can be explained against the back-
ground of the pertinent properties of these applications types (i.e., FuncSp, HumSp,
TechSp, and Scope) and the governance outcomes we proposed earlier (i.e., DecAuth
11
Stop criterion is gradient of STRESS<.0001

195
5 Innovation Adoption and IT Governance in Enterprise Information Systems

Table 5.12:Application types (factor means and scaling results)

Group Factor means Position
Typea n FuncSp HumSp TechSp Scope DecAuth TaskResp x y
BI&A 6 -.356 -.130 -.081 .072 -.180 -.002 -.201 .035
CCC 12 -.378 -.437 -.081 .236 .504 .499 -.047 -.449
CRM 10 .310 .315 .042 -.070 -.596 -.333 -.092 .411
ERP 13 .253 .503 .202 .360 .292 .101 .319 .127
HRM 5 -.358 -.193 -.899 -.595 -.703 -.583 -.788 .003
Prod 5 .705 .552 .793 1.096 .804 .854 .987 -.195
MES 4 .677 1.067 .806 .792 -.062 .075 .723 .377
SM 3 -.312 -.010 -.445 -.284 .597 -.286 -.358 -.223
SCM 6 .198 -.349 .534 -.387 .267 .393 .204 -.253
Other 12 -.325 -.633 -.513 -.813 -.546 -.525 -.746 .165
a
BI&A: Business Intelligence and Analytics, CCC: Content, Communication and Collaboration, CRM: Customer
Relationship Management, ERP: Enterprise Resource Planning, HRM: Human Resource Management, Prod:
Productivity and office applications, MES: Manufacturing Execution Systems, SM: Service Management, SCM:
Supply Chain Management

and TaskResp). Property fitting provides a way to visualize and interpret previously
assessed dimensions in MDS. Following the procedure described byKappelhoff(2001,
pp. 64-69), we perform linear regressions of the (x, y)-values on each of the specificity
and governance dimensions. The resulting coefficients can be translated into gradients of
the plotted dimensional lines (y = ax) by a division (a = cx ),cyintercepts can be omitted.
Figure5.9illustrates how these lines are situated in our two-dimensional space.
For the purpose of interpretation, the (average) properties of each application type
can be assessed by drawing an orthogonal line (i.e., dropping a perpendicular) on the
dimension of interest (therefore intercepts are not relevant). We may exemplify this by
SaaS ERP systems, see dashed lines in Figure5.9. According to our data, SaaS ERP
systems exhibit an above-average human asset specificity (HumSp) as well as a above-
average technical specificity (TechSp). Since the correlation (or influence) of TechSp on
the governance dimensions is reasonably stronger for TechSp, in sum these specificity
characteristics lead to a slightly more IT-oriented governance outcome (here the perpen-
dicular on TaskResp is depicted). Thus, this graphical interpretation is largely consistent
with the factor mean values provided in Table5.12(i.e., HumSp=.503; TechSp=.202;
TaskResp=.101).12 In contrast, for example, SaaS Human Resource Management (HRM)
solutions exhibit far below-average functional, human asset, and technical specificities
and are therefore largely governed by business stakeholders. Apparently the same applies
to the the bunch of solutions classified as other applications by our respondents.
This visualization also expresses the relationships of the specificity and governance di-
mensions as a whole. First, we note that the FuncSp dimension itself is almost orthogonal
to the TaskResp and DecAuth lines. That is, increasing (or decreasing) the functional
12
Note that the dimensions do not perfectly scale, i.e. due to the aggregation in a two-dimensional space
there is an average STRESS=17.7% error in the the prediction of this graphical interpretation.

196
 5.4 The Dual Role of Information Systems Specificity for Governing SaaS

Figure 5.9:Multidimensional visualization of application types (including specificity and

governance dimensions

specificity of an application is unlikely to affect these application governance outcomes

(since the intersection with the perpendicular remains the same). However, in line with
our nomological framework, increasing FuncSp regularly also increases the HumSp and
TechSp, which are not perfectly othogonal to the governance dimensions. Rather, HumSp
correlates with more business involvement (acute angle towards Business) while TechSp
stongly correlates with IT involvement (acute angle towards IT). This visualizes how
and why the impact on the application governance dimension may vary. A greater
Scope of use, in this sense, can be viewed as a moderating influence which increases the
correlations between all the mentioned variables, practically by pushing an object (i.e., the
application) to a more outer (or inner) orbit within this spacial model.

5.4.7 Discussion

This work was motivated by the principal argument that the fundamental characteristics
of ensemble information systems (IS) are not yet sufficiently understood, so that we
still fall short in explaining individual application governance outcomes. We employed a
transaction cost theoretic lens and hypothesized that different components of IS speci-
ficity (i.e., functional specificity, human asset specificity, and technical specificity) lead to
different and partially contrary application governance patterns. Our hypotheses are
grounded in the rationale of organizational and technical embedding. That is, enter-
prise IS (including packaged applications as well as SaaS) are typically first functionally
specified before they are technically implemented and taken in use in an organizational

197
5 Innovation Adoption and IT Governance in Enterprise Information Systems

context.
We find empirical evidence for our argument that greater technical embeddedness
(represented by technical specificity) leads to greater IT involvement in application-
related decisions and tasks (H2, H4, H4), while greater organizational embeddedness
(represented by human asset specificity) is likely to induce more involvement of business
units (H1, H3, H3). This unveils a theoretically interesting, and practically relevant
new dualism regarding the contradictory role of IS specificity for explaining governance
phenomena. Furthermore, we find strong support for the relevance of considering the
scope of use of an IS as a moderator in these relationships. Consequently, the more
employees use an IS, the greater is the impact of the systems functional specification
on both the extent of organizational and technical embedding of the ISand thus the
need to define appropriate governance arrangements. We see three major directions of
theoretical impetus provided by this research.

Contribution to IT Governance Theory

Our findings contribute to past research on IT governance since they reconfirm the need
to balance decision rights between different stakeholders involved in IT related decision
making (e.g.,Brown1997;Brown and Magill1994;Sambamurthy and Zmud1999;Weill
and Ross2004a). In addition to stressing the need for balancing decision rights and
proposing appropriate new governance dimensionsi.e. decision authority and task
responsibility between business and IT unitsthis research provides reasons for why
application governance arrangements vary. We find these reasons (at least partly) in the
specificity characteristics of different information systems. This represents a new contri-
bution inasmuch as (a) this is (to the knowledge of the authors) the first work to employ
a purely transaction cost theoretic (TCT) lens for investigating internal IT governance
phenomena (i.e., excluding outsourcing). TCT thus emerges as a particulary valuable
lens to capture the different facets of IS specificity that play a role in IT governance
contexts.
(b) Past IT governance research has focused on business related contingencies (such
as firm strategy, size and diversification) and thus largely neglected the technological an-
tecedents (Brown and Grant2005, p. 704). Just more recently, authors have emphasized
the complementarities between the technical and organizational architecture (Tiwana
and Konsynski2010). Thus, our findings add a new dualism to this complementary
view. (c) As a result, we particularly advance in the understanding of IT governance
phenomena on the application level. Such modular view, i.e. disaggregating the bulk IT
function along its technological architecture, appears long overdue given the increasing
dispersion of information systems within, and outside of, organizational boundaries.

Contribution to Transaction Cost Theory in IS

We put forward, that our appropriation of TCT to an IS context, and particularly the
distinctive view of IS specificity as a construct composed of different facets, may also

198
 5.4 The Dual Role of Information Systems Specificity for Governing SaaS

provide a valuable lens to address other IT governance phenomena, particularly related to

IS outsourcing. In line withWilliamson(1979;1981;1985;1991;1996), we argued
that application-related transaction costs (or better: coordination costs) can arise from
different types of assets (here: functionality, human assets, and technological integration).
However, in contrast to Williamson we do not argue that one single form of governance
(e.g., hierarchical coordination) is optimal to economize on all types of coordination
cost. Instead, we argued that, for example, costs arising from human asset specificity are
economized through business coordination while costs arising from technical specificity
are economized through stronger IT coordination.
This proposition is in line with prior organization researchers who have emphasized
that different kinds of transaction cost can arise from the possible forms of governance
(Hennart1993). While this seems to be a very simple logic, such multi-faceted un-
derstanding of IS specificity might provide a new key to resolve some of the inconsis-
tencies regarding past TCT research in studying IS outsourcing phenomena (Karimi-
Alaghehband et al.2011;Lacity et al.2011). Transferring our logic (back) to the domain
of IS outsourcing, this could imply that human asset specificity is handled more efficiently
under hierarchical coordination (i.e., in-house) while (at least some aspects of) technical
specificity can be handled better through market coordination (i.e., by providers who are
potentially more specialized to economize on production costs and technical coordination
costs). The presented examples from the TCT/IS outsourcing literature (Benlian et al.
2009;Dibbern et al.2005) provide at least a first indication for this argument. In
contrast, aggregating a multitude of facets into a single dimension, as in (Aubert et al.
2004), potentially just aggravates the prevailing ambiguities around asset specificity in
an IS context.
As a way forward, future TCT/IS studies might consider the issue of multidimensional-
ity of the IS specificity construct. Researchers might also revise past data in this regard,
and thus exploit the body of empirical IS studies carried out on IT outsourcing over the
last nearly 20 years (Lacity et al.2011, p. 148). This perspective, however, contradicts
the immediate claim to discard constructs from reference disciplines that have not been
empirically robust in the ITO context, like asset specificity from [TCT]. (Lacity et al.
2011, p. 149). The aim of this study was rather to provide a way for a better appropriation
of this reference theory to the IS field (Karimi-Alaghehband et al.2011). In this sense,
we also demonstrated how we can interpret the frequency construct in an IS context
as a scope of use. Ultimately, by this we expect a better understanding of IT artifact
specificity, its interaction with other TCT variables, and its impact on (internal and
external) coordination mechanisms. Such understanding may be particularly helpful in
the beginning era of SaaS and Cloud computing (Bento and Bento2011;Khajeh-Hosseini
et al.2010).

Contribution to Theories of the IT Artifact

We believe that our distinct approach also allows us to position our findings in a wider
context regarding the conceptualization of the IS/IT artifact. Since the call fromOr-

199
5 Innovation Adoption and IT Governance in Enterprise Information Systems

likowski and Iacono(2001), IS researchers have widely debated about conceptualizations

of the IT artifact (e.g., inKing and Lyytinen2006).Strong and Volkoff(2010) theo-
rize on a conceptual model originally propounded byWand and Weber(1990), which
characterizes the IT artifact to consist of deep, surface, and physical structures. Surface
structures represent the facilities that allow users to interact with the system, i.e. the user
interface and its usability criteria. Deep structures represent things, properties, states and
transformations of real-world systems, i.e. functionality and data, and physical
structures refer to the underlying technology. Some readers may also recognize these
elements as presentation, logic, data and hardware layers in software engineering (e.g.,
Fowler2003, pp. 17-24).
Strong and Volkoff(2010) extend this conceptualization by a fourth element that they
refer to as latent structures containing roles and controlwhich we may consider as
core elements of IT governanceas well as organizational culture. As the authors argue,
the latent structures emerge from and depend on the other three structures, or in
other words, the arrangements for governing and managing IT artifacts are inherently
related with the usability, functionality, data, physical configuration, and potentially
other properties of an IS. However, even studies taking an ensemble view focus on what
the technology is connected to, while black-boxing the structures that constitute the
technology itself, as the authors continue (Strong and Volkoff2010, p. 752).
The study byStrong and Volkoff(2010)which can be regarded as a type II theory,
according to the taxonomy byGregor(2006)makes a valuable point by postulating
an integrated ensemble view of latent structures and the IT artifact. However, it falls
short in developing testable propositions to learn to design [such] latent structures as
the authors acknowledge (p. 752). We suggest that the dimensions of IS specificity set
forth in our study (foremost functional and technical specificity) can be well regarded as
measurable properties of the deep structures of an IS, while the dimensions of application
governance clearly relate to the latent structures. In this sense, our study provides
first causal explanations and tested propositions on the relationship of deep and latent
structures. This, we argue, also advances our understanding of IT artifacts from a
type II towards a type IV theory (Gregor2006).

Practical Implications

This research also provides relevant insights for practitioners, both in business and IT
positions. First, our model facilitates a contingent understanding on how management
and governance patterns for single applications depend on the type of application itself.
Taken the examples from the outset, this would mean that the management of con-
ventional ERP systems is best organized in central IT competence centers due to their
high functional and technical specificity (Krmmergaard and Rose2002). In contrast,
emerging social software tools can, and potentially should, be given to the hands of
business stakeholders, since/if they are operated as largely isolated applications (i.e.,
technically unspecific) and deal with highly company-specific knowledge (i.e., human
asset specificity) (Deans2011).

200
 5.4 The Dual Role of Information Systems Specificity for Governing SaaS

As our data was based on the assessment of various SaaS applications, the results
particularly provide a rationale on how to design such governance arrangements in the
Cloud era. Based on a multidimensional scaling, we provide a novel visualization of
different SaaS application types that aggregates all relevant model dimensions within
a single diagram and can be used to infer from the specificity characteristics of an
application to its governance dimensions. Such simple, yet meaningful, visualization
provides a useful overview and thus a comprehensive understanding of the pertaining
characteristics of different application types. Our hope is that this model may also be
used to guide practical decision making across diverse SaaS adoption and governance
contexts.
Consequently, under some conditions (i.e., low technical specificity and scope, and
high human asset specificity) this may imply to grant more decision rights to business
stakeholders and thus manage an increasing number of decision rights dispersed across
the organization. For some CIOs, to loosen the imperative of strictly centralized IT
control known from many traditional enterprise IS, may well represent a paradigm shift.
On the other hand, business practitioners may find our insights helpful to realize that
even for SaaS, not only in case of high technical integration, some coordination with
IT units remains a key necessity. Thus, altogether our dualist view strengthens the
importance of aligning IT with business (and vice versa), all the more in the era of Cloud-
based computing.

Limitations and Future Work

Three limitations of this study merit consideration, foremost related to issues of opera-
tionalization, generalizability, causality and endogeneity. First, due to practical restric-
tions, we operationalized constructs with a comparably low number of items. Arguably,
a semantically broader operationalization may have improved predictive validity of our
research model (Diamantopoulos et al.2012). Second, our sample is based on data from
large-sized enterprises in Germany using SaaS applications. Thus, any generalization to
any other macro and micro context is purely argumentative. Third, the cross-sectional
analysis can only ascertain association, not the causal ordering that derives from our
theoretical argument. Conversely, governance arrangements may also influence specificity
characteristics of the ensemble information system that was subject of this research.
We see multiple directions for further research. First, as the conceptualization of the
IS specificity construct is still in an exploratory phase, future works may unveil further
subdimensions and develop semantically broader measures. According to the presented
conceptualization of the IT artifact, deep structures (e.g., data) and surface structures
(e.g., usability) represent promising characteristics that could be dimensionalized in a
higher-order theoretical model. In addition, as suggested earlier, researchers in the IT
outsourcing field might find it helpful to reconsider the demonstrated multidimensionality
of IS specificity to isolate further potentially opposing effects. Finally, future works
should also address the question in how far the proposed dualism, besides for SaaS, also
holds for the domain of traditional delivery models. As we motivated at the outset, a

201
5 Innovation Adoption and IT Governance in Enterprise Information Systems

comprehensive understanding of IS specificity might provide a way to unite the impor-

tant, yet separate research streams of internal IT governance and external outsourcing
governance, towards a broader understanding of IT governance phenomena in expanding IT
landscapes.

5.4.8 Summary
This chapter addressed the theoretically neglected role of information systems (IS) speci-
ficity for application governance, by referring to the allocation of application-related
decision authority and task responsibility between business and IT units. Based on the
premise of organizational and technical embedding, and employing a transaction cost
theoretic (TCT) view, we developed the idea that customization and greater functional
specificity of an IS lead to both more business authority through higher human asset
specificity and more IT authority through higher technical specificity. Survey data from
76 organizations using different types of software as a service provide support for these
ideas. Our results unveil a new dualism for explaining IT governance phenomena on
the application level. Furthermore, we demonstrated a relevant appropriation of the
frequency construct from TCT to this context and provided a visualization that may
guide managerial decision making in diverse SaaS adoption and governance contexts.
Besides practical implications, we outlined contributions to IT governance, transaction
cost, and IT artifact theories in the IS field.

202
6 Conclusion and Contributions
This thesis was motivated by the principal idea that IT governance plays an important
role for adopting IT-based innovations and that this relationship has been theoretically
neglected in contemporary IT governance research. I conducted eight studies with diverse
municipal governments and enterprises to investigate IT governance challenges regarding
two recent IT-based innovations: Mobile Government (M-Government) and Software as
a Service (SaaS). The particular focus of these studies was on the question of how the
locus of authority for IT decisions influences, or is influenced by, technology adoption.
Furthermore, the studies give concise theoretical insights on each of these specific adop-
tion contexts and aim to provide practical guidance for making governance and adoption
decisions. In the following, I will briefly summarize the results of each of these studies.
In the first study (4.1), we found evidence that there is a relationship between the
strategic framework (i.e., efficiency goals, innovation goals, and IT sophistication) of a
municipality and the planned use of M-Government services. We expressed a suspected
heterogeneity of public agencies by four clusters describing Innovator, IT experienced,
Efficiency-oriented, and Laggard municipalities. An in-depth case analysis of these
types of municipalities particularizes this relationship by providing detailed evidence
for the strategic (financial situation) and organizational (IT governance) contingencies
that determine the extent and focus of M-Government adoption (4.2). Given the result
that most municipalities still focus on internal efficiency and effectiveness (i.e., process
innovations) and refrain from offering citizen-centric and transactional M-Government
services (i.e., service innovations), we shed more light on the benefits of urban sensing,
a particular class of mobile applications for external M-Government. The results of a
citizen survey indicate that urban sensing can provide a means for increasing citizen
participation, whereas perceived privacy risks are not a significant barrier to adop-
tion (4.3). In addition, we demonstrate that such services can improve a a municipalitys
level of environmental information at comparable cost to internal information acquisition
procedures andin this sensesimultaneously allow for implementing a service and a
process innovation (4.4).
In an enterprise context, I studied the impacts of adopting Software as a Service (SaaS)
on internal IT governance. A first exploratory case analysis suggests that companies
allocate authority for the same SaaS application (a well-known CRM system) in distinct,
but clearly distinguishable ways, i.e. either to business or to IT units, based on certain
contingent factors (5.1). Chapter5.2puts these factors on a broader theoretical basis
and partly confirms the proposed contingency model in a cross-industry survey with
large German enterprises. However, the results remain partly inconclusive regarding the
influence of some of the constructs that were derived from an agency and transaction

203
6 Conclusion and Contributions

cost theoretic perspective (particularly the locus of initiative and application specificity).
Addressing some of these limitations, I demonstrate a process-theoretic approach where
the initiative and application specificity factors emerge as intermediate variables between
the overall mode of IT governance and application-level governance outcomes (5.3). Such
process view is taken as a premise for a deeper analysis of the role of information system
specificity in Chapter5.4. Finally, here we are able to unveil that the functional and
technological properties pertaining to (SaaS) applications have a dual influence on the
locus of IS authority.
I am confident that the results of this compound research allow for some overarching
theoretical, practical and methodological conclusions, which I will discuss in the remain-
ing sections.

6.1 Theoretical Contribution: On the Relationship of IT

Governance and Innovation
Two principal research questions guided this dissertation, RQ1: how does the mode
of IT governance influence the adoption of new technologies and RQ2: how does the
adoption of new technologies affect organizational IT governance. In this context, we
understood IT governance more narrowly as the locus of IT decision rights (Brown
and Magill1994) and adoption of IT innovations as the decision of an organization
to make use of a new technology (Rogers1962). At first sight, these two constructs
may seem two incompatible concepts from different domains without a logical link. IT
governance is often thought to prescribe a rather static set of rules and guidelines that are
derived from corporate governance and fit to certain business contingencies. In contrast, IT
innovation inherently deals with different technologies and a dynamic development over
time. However, our argument for linking these two domains was motivated by the
observation that IT organizations have historically oscillated between centralized and
decentralized forms (Peak and Azadmanesh1997;Brynjolfsson and Hitt1998;Evaristo
et al.2005) as well as by the practical necessity to provide more guidance for organizations
on how to govern IT innovation and technological change (Chapter2).
Regarding RQ1, I particularly drew the focus on the adoption of mobile technologies
in a public sector context. Both the results of a survey-based investigation as well as
an in-depth case investigation prove that there is a relationship between the strategic
framework of municipalities and their adoption of M-Government innovations. Important to
note, this strategic framework is composed of drivers that are less easy to change, such
as the efficiency-versus-innovation dimension, oras we put it laterthe financial
situation. On the other hand, it also comprises drivers that are more actionable, such as
IT sophistication (referring to the institutional arrangements and technological vision for
M-Government) or, as we concretized in the second study, the mode of IT governance.
Regarding the latter, we provide evidence that those public agencies that are able to
effectively connect responsibilities for information technology and organizational change
also succeed in the adoption of IT-based innovations. In contrast, municipalities that

204
 6.1 Theoretical Contribution: On the Relationship of IT Governance and Innovation

fail to do so run the risk to end up with an unfocused adoption of M-Government

innovations with less efficiency and greater frictions. Thus, this capability, which we
termed transformational IT governance, appears to be a key prerequisite to ensure
successful implementation especially of process innovations in public sector.
According to our case findings, transformational governance can be achieved in dif-
ferent ways. For example, responsibilities for IT and organization can reside within
the same organizational unit, such as in Case A in an office for organization and IT
steering that is subordinate to the general office for central services. Or, the office for
IT steering is subordinate to the central office for (personnel administration and)
organization, such as in case C. It may even be sufficient if these responsibilities are
effectively aligned via appropriate governance mechanisms, such as in case B, where the
staff council participates in a joint IT steering committee that meets every 4-6 weeks.
However, if joint IT and organization meetings are not sufficiently empowered to make
decisions, such as in case D, then alignment and governance outcomes are likely to be
suboptimal from an innovation adoption standpoint. This is mainly because potential
user resistances can not be effectively mitigated and overcome.
Given these examples, we find that transformational IT governance cannot be instantly
mapped to the presence of a single organizational configuration or governance mechanism.
However, we provide an argument that allocating IT decision rights closer to where
decision rights for personnel and organization reside (which is usually centralized ) will
generally improve process innovativeness. This represents a novel finding inasmuch as
it contradicts the classic rationale of decentralizing IT governance for greater innovation
(e.g.,Weill and Ross2004b). (How much this finding can be attributed to the par-
ticular public sector context, will be the subject of the next section.) Regarding the E-
Government context, the claim for transformational IT governance is largely consis- tent
with the upcoming literature on transformational government (Irani et al.2008;
Weerakkody et al.2007). Although several authors have emphasized the importance of
IT governance in E-Government, few works have made concrete propositions on how
to operationalize (i.e., implement) effective IT governance in this field (Janssen and
Shu2008). Especially in the context of M-Government, the role of IT governance has
remained without precise suggestions, although some experts highlighted barriers relating
to governance when initiating a mobile service project (El-Kiki and Lawrence 2007, p.
785). Besides providing exploratory case evidence, I described very practically how
different stakeholders can be aligned when launching a new M-Government service
(Chapter4.4). For this purpose, we drew on the case of the State Capital of Saarbrcken
and the implementation of an urban sensing application. For these reasons, I believe
that the theoretical and practical findings of this work make a particularly valuable
contribution to the field of IT governance in E-Government.
Regarding RQ2, I hypothesized the adoption of Software as a Service and emerging
Cloud-based solutions to have an impact on IT governance. This suspicion was rooted
in the nature of SaaS as an (increasingly successful) external delivery model as well
as in its technological differences in the way that software is provided. Conceptually,
this principal hypothesis first required to conceptualize IT governance arrangements on

205
6 Conclusion and Contributions

the application-level in order to isolate the differences between SaaS and traditional
delivery models. Therefore we theoretically developed the two subdimensions decision
authority and task responsibility. Both an exploratory case analysis as well as a large-
sample survey showed that companies (sometimes) allocate authority and responsibility
for SaaS in ways that differ from traditional delivery models, i.e. with significantly more
business involvement. Throughout the course of the studies, we proposed a transaction
cost theoretic explanation for this finding. According to this view, companies allocate
application authority between business and IT units in a way that minimizes their total
IT production and IT coordination cost. For traditional delivery models, firms have
a great incentive to centralize application operations to a corporate IT department to
leverage internal IT production cost advantages, so that the costs of coordination remain
secondary. For SaaS (as well as other outsourcing models) however, production cost
advantages largely accrue on the provider side (expressed by the providers fees). Thus,
the question of internal governance becomes merely dependent on the costs incurred for
coordinating application operations (also termed by Williamson as governance costs).
For this reason, the rationale of leveraging internal economies of scale becomes much
weaker for SaaS, which in turn explains the greater decentralization of authority for
SaaS.
Nevertheless, the proposed shift of application governance clearly does not yield the
extent that some radical proponents would presumably expect (Carr2005). This is, the
majority of SaaS applications (59.2%) is still governed in IT dominated patterns, while
there is also a considerable amount of traditional on-premise software, which is gov-
erned in non-IT-dominated patterns (19.8%). This led us to research more into the tech-
nological contingencies of application governance, which have been largely neglected by
the past IT governance literature (Brown and Grant2005, p. 704). To qualify this, some
research has considered the technological dimension inasmuch it conceptually separated IT
decision rights for applications and infrastructure, leading to the wide recognition of the
federal archetype (e.g.,Brown1997;Sambamurthy and Zmud1999). However, I may argue
that this separation still lacks in sufficiently capturing the technological dimension, since the
border between what is an application and what is infrastructure remains blurred. We
may consider two provocative examples: According to a federal governance logic,
productivity applications, such as common office and spreadsheet programs, would be
provided in a decentralized fashion by business units. Conversely, high-performance
computers used for example for laboratory automation systems (Wood2007)in their
nature as infrastructure artifactswould be governed centrally by a corporate IT unit.
Intuitively, this does not necessarily reflect the reality in a majority of firms. Thus, the
mere knowledge of whether an IT artifact is an application or infrastructure does not
appear sufficient to explain why individual governance outcomes occur.
In the last study of this thesis, we proposed that four essential constructs derived
from transaction cost theory are better suited to capture the pertinent and salient
characteristics of IT artifacts: functional specificity, human asset specificity, technical
specificity as well as scope of use. According to this view, more business involvement
is caused by human asset specificity, while more IT involvement follows from technical

206
6.2 Practical Contribution: Reflections on IT Governance in Public and Private Sector

specificity. This dualism is grounded in a procedural view of SaaS (or generally informa-
tion system) adoption. That is, information systems are gradually made (functionally)
specific when being adopted in an organization, which in turn leads to both higher human
asset specificity as well as technical specificity. Such procedural views are consistent with
the literature on enterprise-system fit (Soh et al.2000;Sia and Soh2007;Strong and
Volkoff2010), and have as well been applied in a governance context (Xue et al.2008). Our
data on SaaS application governance provides evidence for this view. This suggests that the
type of application (expressed by its specificity) clearly has a bearing on the mode of
application governance, or more generally, that the type of IT-based innovation adopted can
even influence (at least to some extent) the overall mode of IT governance.
The compound findings of this second part make three main contributions. First,
as outlined above, we conceptualize overall IT governance as the sum of more modular
governance arrangements on application level (here dimensionalized by decision authority
and task responsibility). Second, we provide a transaction cost theoretic explanation for
why the classical rationale of centralizing governance for greater efficiency and economies
of scale is not necessarily applicable for external delivery models (such as SaaS). More
than that, we provide some exploratory case evidence that firms seeking greater efficiency
potentially tend to decentralize application governance for SaaS more than other orga-
nizations would do (Chapters5.1and5.3, case B). This can be regarded at least as one
example falsifying the classic strategy-structure fit for IT governance on the application
level. Finally, we show how the type of IT-based innovation (characterized by different
specificity dimensions) has a bearing on the particular mode of artifact-level governance,
i.e. we demonstrate some of the technological contingencies.
Altogether, I argue that the findings made for both research questions enrich IT
governance theory by providing novel insights and theoretic rationales regarding the
mutual and bidirectional relationship between the mode of IT governance and diverse IT-
based innovations. The main contributions of this research and how they extend
existing IT governance theory are summarized in Table6.1.

6.2 Practical Contribution: Reflections on IT Governance in

Public and Private Sector

The two different innovation contexts considered in this dissertation allow us to de-
rive a few practical guidelines for designing IT governance in public and private sector
(compareWeill and Ross2004a, pp. 185-214;Ali and Green2007;Sethibe et al.2007).
Moreover, the described findings, i.e. the proposition of transformational governance
in E-Government as well as the observed shift of enterprise SaaS governance from IT
to business units, immediately raise the question of how far these observations also hold
for an opposite sector context. In order to reflect on these questions, I will first briefly
review the principal differences between public and private sector organizations before I
discuss both findings separately.

207
6 Conclusion and Contributions

Table 6.1:Contributions to IT governance theory

Subject Classic understanding (literature) Contribution (chapter)
1)Types of Business strategies can be Internal M-Government (i.e., process
innovation innovation-oriented (i.e., focus on innovations), external M-Government
considered product and service innovations) (e.g., (i.e., service innovations) (4.1,4.2,4.4),
Sambamurthy and Zmud1999;Weill as well as SaaS (delivery model
and Ross2004b) innovations) (5)
2)Technology Widely neglected (Brown and Grant Information system specificity
contingen- 2005, p. 704), except from separation in (especially functional and technical
cies application and infrastructure specificity) (5)
governance (e.g.,Brown and Magill
1994;Brown1997;Sambamurthy and
Zmud1999), recently also modularity of
technological architecture emphasized
(Tiwana and Konsynski2010)
3)Level of Widely bulk IT function view, more Organization level and application level
governance recently also investment approval (Xue IT governance (5), transformational
et al.2008), system development governance (4.2)
(Tiwana2009), data (Khatri and
Brown2010), and infrastructure
sourcing governance (Xue et al.2011)
4)Direction StrategyIT governance
Innovation Strategy Innovation (4.1),
of assumed (e.g.,Sambamurthy and Zmud1999; Strategy IT governance Innovation
(4.2), St AppGov (5.1),
effectab Weill and Ross2004b) rategy
InnovationAppGov(IT governance)
(5.2), St rategySp ecificityAppGov
(5.3), SpecificityAppGov (5.4)
5)Sector Focus on private sector, few research in Public sector (4), private sector (5)
focus public sector (Weill and Ross2004a;Ali
and Green2007;Sethibe et al.2007)
Overall Innovation strategies favor decentralized Process innovations can favor
IT governance; transformational governance and more
centralized decision making.
Efficiency strategies favor centralized For external delivery models (SaaS),
IT governance. efficiency strategies can favor higher
decentralization of IT governance.
a
Strategy = strategic goals of a (government) organization or the competitive strategy of a firm, respectively
b
AppGov = application governance

208
6.2 Practical Contribution: Reflections on IT Governance in Public and Private Sector

Differences between public and private sector organizations are widely discussed in
the academic literature. Apparently, the most fundamental difference lies in the primary
goal and justification of these two institutional forms for creating public value versus (in
a simple view) achieving profit (Moore1995;Moore and Khagram2004). Public sector
(i.e., non-profit) organizations can be also characterized by lower degree of competition,
greater legal and political constraints, fewer performance incentives (both individually
and organizationally), greater scrutiny and higher aversion to take risks compared to
the private sector (Rocheleau and Wu2002). Consequently, regarding the role of in-
formation systems, it is often argued that public sector organizations view technology
more as a necessity rather than a tool for competitive advantage. Risk aversion and
the great amount of stakeholders make it more difficult to plan and implement IT-based
innovations. Therefore, public organizations are often seen as the late adopters of new
technology (Rocheleau and Wu2002). However, the information intensity of public
agencies task environment and the IT architectural complexity should not be underesti-
mated (Vilvovsky2008). For example, some of the interviewees in our qualitative studies
(Chapters4.1,4.2) pointed out that their municipalities managed several hundreds of
administrative procedures and according IT applications. Operating such relatively
broad spectrum of IT services (e.g., measured by applications per IT staff) would be
hardly imaginable for most private companies (since those inherently tend to focus on
core their processes). Also, contrary to the stereotype of being bureaucracies, authors
more recently recognize a move toward more business-like performance measurement and
more cost efficiency across all levels of public agencies (Vilvovsky2008).
Our research suggests that establishing transformational IT governance, i.e. ensuring
that responsibilities for IT and organization are effectively aligned, is a valuable practice
for public organizations to foster focused IT-based innovation adoption. Based on the
outlined characteristics of the public sector, we may explain this by the lack of economic
pressure, multiplicity of stakeholders and the lower creativity about technological possi-
bilities. That is, administrative departments of public organizations have a much lower
incentive to innovate, especially to implement process innovations that may potentially
change their own work practices and jobs.1 In contrast, business units in private or-
ganizations typically have own profit responsibility, so that they are not only inclined
to exploit potential IT-based product innovations, but also to constantly improve their
own cost base. Accordingly, the call for a transformational governance may seem less
relevant for the private sector. However, even many private enterprises resemble large
bureaucracies that need to involve a great number of stakeholders in realizing IT-enabled
business transformation (Venkatraman2005). Thus, one may also agrue that in this case
the call for aligning IT and organizational responsibilities appears equally important to
the public sector. The fact that in recent years an increasing number of large companies
have united IT and human resources functions under the same roof of a corporate COO

1
This aspect became particularly apparent in an anecdote from case D (4.2). The IT department in
case D had been given the goal to liberate resources through standardization and automation. Our
interviewee told us that when he once tried to explain this to one of the department heads, the
department head was literally threatening him by saying Just try to take one of my staff away.

209
6 Conclusion and Contributions

(Chief Operating Officer) may at least provide weak evidence for this argument (CIO
2008).
The second major finding from this research refers to the if and why of a governance
shift for enterprise SaaS solutions from IT to business units. Based on the differences
of public and private sector organizations, we may ask whether the observed shift is also
expected for the public sector. First, we have to state that SaaS and Cloud-based service
adoption is still limited in public sector due to various challenges related to quality,
privacy and security (Janssen and Joha2011). Our main argument for the decentral-
ization of SaaS authority was based in the transaction cost theoretic consideration of
reduced economies of scale for SaaS. Furthermore, we partly attributed this shift to the
lower technical specificity of SaaS solutions (i.e., the technological contingencies) and the
increased IT skills of business users (i.e., the knowledge-based view). As outlined earlier,
IT application landscapes in public sector (at least in municipal governments) exhibit
a comparably high technological complexity and low economies of scale. This means
that synergies across municipal departments are comparably low since the departments
administrate distinct procedures. Therefore, I argue that this is one reason that militates
against an occurrence of the propounded shift. Also there is no reason to assume that IT
skills of the users in municipal departments are higher than in private sector businesses,
so that this is not likely to be a reason either. Rather we assume that the greater
legal constraints in the public sector are more likely to allow only for singular adoption
of SaaS based solutions under the regime of the IT department, if at all (Janssen and
Joha2011). On the other hand, the fact that administrative procedures do not vary
much across different municipal bodies greatly spurs the market opportunities for highly
standardized, low cost SaaS-based solutions also in the public sector. Thus, there is still
an argument that we will see greater SaaS adoption and an effect on IT governance in
the public sector in the future.

6.3 Methodological Contribution: On the Use of Multimethod

Research
In this section I review the conducted studies according to the methodologies used. Since
this dissertation made use of several research methods, this review particularly intends
to identify contributions for the use of multimethod research.
Combining multiple methods and different research paradigms to investigate a specific
subject has been generally propounded as a valuable approach, especially in the IS field
(Kaplan and Duchon1988;Orlikowski and Baroudi1991;Gable1994;Mingers2001;
2003). The most advocated approach certainly refers to the combination of qualitative
and quantitative research methods, specifically case study and survey research (Kaplan
and Duchon1988;Gable1994). Later authors proclaim a general pluralism that does
not prescribe a mandatory sequence for applying different methods, nor a limitation
to a single paradigm (e.g., a strictly positivist view) (e.g.,Mingers2001). Although
IS researchers have early accredited the value of multimethod research (Orlikowski and

210
 6.3 Methodological Contribution: On the Use of Multimethod Research

Baroudi1991), some meta-reviews still assert a paucity of such pluralist approaches.

According to a meta-review, only about 20 percent of the papers in major IS journals
apply multiple research methods (Mingers2003). Barriers to multimethod research have
been identified in philosophical (e.g., incompatibilities of different paradigms), cultural
(e.g., traditions of different schools of thought), psychological (e.g., personal inclination
towards a certain style of research), and practical categories (Mingers2001). This
review particularly intends to explore the practical possibilities on how to link different
research methods.
An overview of the methodologies used in this dissertation has been presented earlier in
Table3.1(page35). The combinations of different research methods can be summarized
by four broad categories: a) links from qualitative to quantitative methods, reversely
b) links from quantitative to qualitative methods, as well as c) combinations of different
qualitative methods, and d) combinations of different quantitative methods, see Table6.2.

Table 6.2:Multimethod research overview (including chapter numbers)

a) Linking qualitative with b) Linking quantitative with
quantitative methods qualitative methods
Explorative model generation and Empirical clustering and detailed
5.2) case analyses (4.1 4.2)
model test (5.1
Operationalization of a theoretical Validating a sim ulation model in a
model and test (4.1) case study (4.4)
Procedural regrouping of constructs Integration of factors in a process
and test (5.3 5.4) model (5.1/5.2 5.3)

c) Combining qualitative methods d) Combining quantitative

Grounded theory and content methodsa
analysis (4.2) Factor models and clustering (4.1,
Grounded theory and multiple case 5.2)
studies (4.2,5.1) Factor models and multidimensional
scaling (5.4)
a
Factor scores from partial least squares structural equation models were used for further analyses

Multimethod research of type a) is certainly the most common approach (Kaplan

and Duchon1988;Gable1994;Mingers2003). In Chapter5.1(Impact of SaaS on IT
Governance) we conducted a set of explorative case studies to derive a conceptual model,
which was then operationalized and tested in a large-sample survey (Chapter5.2). This
approach, in spite of being allotted to two different research papers, largely matches the
multimethod approach propounded byGable(1994).
Second, qualitative approaches (in this case a content analysis of case interviews) may
also be used more simply to operationalize an existing theoretical model, as we demon-
strated in Chapter4.1(Section4.1.6). Thus, the role of the qualitative approach in this

211
6 Conclusion and Contributions

case is to validate the model previously derived from literature (the E-Government and
IS strategy literature) and to transfer it to a survey instrument, rather than inductively
generating an entirely new model. This approach is more congruent withKaplan and
Duchon(1988, p. 582), who emphasize the need for context-specific measures when
investigating a phenomenon of interest. This means practically that adopting question-
naire items from the related literature without such phase of qualitative reflection and
applying these in the own research context is less likely to lead to success.
Finally, with Chapters5.3and5.4we demonstrate that the link between qualitative
and quantitative methods can also derive from causal-logical reordering. That is, in
Chapter5.3we developed the idea that allocation of SaaS governance is the result of a
process rather than of a set of contingent factors. This idea led to a regrouping of some of
the factors from the prior analysis (Chapter5.2) as well as an addition of further factors
(i.e., human asset and technical specificity) and the explanation of an effect which had
previously remained inconclusive (i.e., the effect of specificity on application governance
arrangements).
Regarding b) the reverse direction from quantitative to qualitative research, in Chap-
ter4.1we performed an empirical clustering based on a sample of municipalities and
the key model dimensions. These empirical clusters were then used in a subsequent
work to analyze one (selected and revelatory) case from each subgroup in detail by the
use of qualitative research methods (Chapter4.2). I argue, that in terms of a proper
sampling logic (Yin2003)which is often a challenge when seeking for particularly
revelatory casesthis type of multimethod approach is very promising and particularly
beneficial. This is because, given that these four clusters exhaustively represent the
entire population, it strengthens the generalizability of the evidence from the four cases
to the whole population (here municipalities in Germany). Practically, this kind of
linkage between quantitative and qualitative research can be achieved simply by asking
survey respondents for their willingness to serve as a contact person for subsequent case
intverviews.
A different approach was taken in Chapter4.4, where we proposed a quantitative
model to estimate the effects of urban sensing adoption. This model was then validated
in a case example (the state capital of Saarbrcken) in several steps, including the use
of process mapping as well as observing the effects of modeling in a decision making
context. Both can be clearly seen as validation methods, i.e. here a quantitative model,
in its nature as a human-constructed artifact, is evaluated and proven for correctness
by a real-world case study (Hevner et al.2004;Sein et al.2011).
A third b)-type of link can be noticed between Chapters5.1/5.2and5.3. In this case,
we moved from a contingency theoretic and variance-paradigmatic view to a process-
theoretic model of SaaS governance. Such switch of theoretical paradigms is not without
challenges, as we outlined in the chapter: Factors should not be understood as predictors
of certain events [. . . ], but rather as a social action that helps to produce the outcome
of interest (Section5.3.3, page168). However, the approach taken in that chapter can
be regarded as one attempt to link variance with process theories.

212
 6.3 Methodological Contribution: On the Use of Multimethod Research

As outlined before, the term multimethod also implies that methods can be combined
that use the same type of data (i.e., qualitative or quantitative data), but follow poten-
tially different research paradigms. Regarding category c), we have utilized different
qualitative methods particularly in two chapters. Chapter4.2combines the use of
grounded theory with content analysis in a five step approach to derive and valuate
different drivers and inhibitors of M-Government adoption. The strengths of using
2
grounded theory as an analytical method here are that it provides different overarching
categories (the coding paradigm,Glaser and Strauss1967) in function of the relationship
to the core phenomenon. Thus, the coding paradigm helps to create a multilevel and
relational categorical system (our proposed framework) without the use of prior theory.
The key strength of content analysis, opposed to this, is to provide a method for valuating
different codes and determining the overall coding reliability. Therefore the combination
of the two qualitative approaches appears particularly useful to achieve reliable results,
not only in our M-Government context.
The study in Chapter4.2, as well as a second qualitatively oriented study in Chap-
ter5.1, also combine grounded theory with a multiple case study approach. In both cases,
the categories produced by the grounded theory approach are subsequently instantiated
for each case and analyzed in a comparative manner. This, I argue, appears to be a
particularly useful approach, since the case study methodology itself hardly provides any
advice on how to analyze and potentially cateogrize the data gathered from a case site
([Link] et al.1987;Yin2003).Benbasat et al.(1987, p. 374) write succinctly that
the analysis of case data depends heavily on the integrative powers of the researcher and
that a clear chain of evidence should be established. Arguably, the prescriptive
approach demanded by the grounded theory method can aid a researcher in improving
his/her integrative powers and thus help to discover the relevant chains of evidence. This
way of combining the case study method with grounded theory has also been regarded
as a beneficial approach by other researchers (e.g.,Hughes and Jones2003;Fernndez
2005).
Finally, regarding the combination of different quantitative methods (category d),
most of the chapters of this dissertation that apply quantitative survey and path model-
ing techniques combine these with more descriptive and/or cluster-analytical methods.
The motivations for combining these two quantitative approaches (based on the same
data) depended on the case. In Chapter4.1we presented a descriptive overview of the
attractiveness of different mobile services (Table4.3) primarily to address a practitioner
audience (see alsoWinkler2011). In the same chapter we also conducted an exploratory
cluster analysis based on the factor scores of the main model dimensions (i.e., efficiency
goals, innovation goals, IT sophistication) in order to detect potential heterogeneity
in the sample (see Figure4.4). A post-hoc analysis of variance of the factor means
suggested the existence of such heterogeneity. In Chapter5.2we conducted a similar
2
Note that grounded theory according toGlaser and Strauss(1967) is a whole research paradigm, rather
than only a method to analyze qualitative data. For example, this approach demands no use of prior
theory, iterative data collection, theoretical saturation and the use of memoing. However, utilizing
grounded theory for analysis is the most common practice in the IS field (Matavire and Brown2008).

213
6 Conclusion and Contributions

clustering of the factor values of the dependent variables (i.e., decision authority and task
responsibility) to detect (and describe) three different patterns of application governance
(i.e., IT dominated, business dominated, and mixed patterns).
Finally, in Chapter5.4, we conducted multidimensional scaling and visualized the
similarity characteristics of different application types (e.g., ERP, CRM, SCM, etc.) and
the core model dimensions (i.e., functional, human asset and technical specificity, as well
as decision authority and task responsibility) in a single two-dimensional plot (Figure5.9,
page197). This visualization is largely consistent with the results from the model tests,
since it is based on the same data. However, it allows us to draw conclusions which are not
intuitively attainable from the mere look at the number sheets (see Tables5.11and5.12).
The results of factor-based path models have been criticized generally for providing only
partial guidance to the practitioner who must assume responsibility for attaining positive
outcomes (Newman and Robey1992, p. 250). In this sense, our combination of factor-
based path models with more descriptive and/or visualized cluster-analytical methods
was mainly motivated by the intention to address a practitioner audience and thus also
to increase relevance of this research.
Overall, this compound research makes a case that it is valuable and useful to alternate
between different qualitative and quantitative research methods to investigate the same
phenomenon. The IS literature provides few practical advice on how to link different
research methods. The presented examples highlight a few possibilities based on the
studies conducted in this dissertation. Important to note, I neither state that any of
these possibilities are free from limitations, nor that they should be prescribed for other
researchers. This ex-post review rather intended to introduce a general framework for
multimethod research and provide a few practical insights. In this respect it may
contribute a few interesting (and potentially novel) methodological perspectives to the
IS field.

6.4 Limitations and Future Research

In this last section I outline the major limitations that should be considered when
interpreting the findings of this work and provide directions for future research. The
limitations of each single study are reported in the according chapter, so that here I may
focus on the outcomes of this compound research.
First, regarding generalizability we have to note that all of the empirical studies were
conducted with specific subjects (i.e., municipalities, citizens, and companies, with few
exceptions) from Germany. Thus, a transfer of our results to any other national or a
global context should consider the potential differences resulting from varying cultural,
legal, and economic frameworks. Second, it should be stated again that the path models
used in this research (in Chapters4.1,4.3,5.2and5.4) only ascertain statistical associ-
ation, not causal ordering. Causal relationships should derive from the theoretical and
empirical arguments made in these and the other, qualitatively oriented chapters.
Furthermore, I add for consideration that the two overall contributions emerging from

214
 6.4 Limitations and Future Research

this compound research, which extend the classical rationale of strategy-structure fit,
still require further empirical confirmation. That is, the claim for transformational
governance to favor process innovativeness in public sector emerged from the analysis of
multiple cases of municipal M-Government adoption. The argument for greater gover-
nance decentralization of external delivery models (including SaaS) under an efficiency
strategy was derived from the transaction cost theoretic argument of reduced internal
economies of scale. I also discussed to what extent these propositions hold for opposed
sector contexts (see Chapter6.2). Nevertheless, both of these propositions still demand
an empirical confirmation by a larger sample of public and private sector organizations
for adding up to the body of IS knowledge.
Overall, having argued that governance is a dynamic phenomenon (e.g., that the vast
SaaS adoption may lead to a governance shift from IT to business units), more research
regarding the questions of how governance arrangements evolve in firms throughout the
IS adoption and assimilation process appears appropriate. Such longitudinal approaches
might also help answer the question of the voluntariness of IT governance arrangements
raised in Chapter5.2, i.e. to answer to what extent the allocation of responsibility in
organizations is the result of formal governance arrangements (as the literature implies),
rather than the outcome of (seemingly arbitrary) initiatives of individual actors.
Finally, regarding the cross-sectoral differences in IT governance, I may citeRocheleau
and Wu(2002) who find that few studies [. . . ] empirically examined differences between
public and private information management practices, and the findings have not been
very consistent. Thus, assuming that public agencies continue to learn from enterprise
practices (and vice versa)3, the comparative view of sector differences taken here could
(and potentially should) be enhanced by more studies that directly assess IT governance
practices across public and private sector organizations.

3
Note that the development of ITIL, a widely used reference framework for IT service management, has
been largely driven by a government organization.

215
Appendices

1 M-Government Survey with German Municipalities (Online)

217
Appendices

218
1 M-Government Survey with German Municipalities (Online)

219
Appendices

220
1 M-Government Survey with German Municipalities (Online)

221
Appendices

222
1 M-Government Survey with German Municipalities (Online)

223
Appendices

224
1 M-Government Survey with German Municipalities (Online)

225
Appendices

226
1 M-Government Survey with German Municipalities (Online)

227
Appendices

228
 2 Citizen Acceptance Survey (Online)

2 Citizen Acceptance Survey (Online)

229
Appendices

230
2 Citizen Acceptance Survey (Online)

231
Appendices

232
2 Citizen Acceptance Survey (Online)

233
Appendices

234
2 Citizen Acceptance Survey (Online)

235
 3 Urban Sensing Simulation Model (Scripts)

3 Urban Sensing Simulation Model (Scripts)

DIFFUSION MODEL
new potential adopters = increase of mobile users in city * potential
user share
Potential adpoters = INTEG (new potential adopters - new
adopters,city population * percentage mobile users * potential user
share)
new adopters = IF THEN ELSE (Time >= implementation time, innovators
+ imitators, 0)
Adopters of the service = INTEG (new adopters, 0)
innovators = Potential adpoters * innovator rate
imitators = imitator rate * Adopters of the service * probability
probability = Potential adpoters / (Potential adpoters + Adopters of
the service)
Non adopters = city population - Adopters of the service

DETERIORATION MODEL
total new issues = new issues per km * street network size
total potential issue spots = max spots per km * street network size
issue density = Undetected spots / total potential issue spots
new spots with issues = total new issues * (1 - issue density)
Undetected spots = INTEG (new spots with issues-new reported spots,0)

REPORTING MODEL
reports from citizen = reports from adopters + reports via
traditional way
issue spots found by citizen = Undetected spots * (1-(1-(1/Undetected
spots))reports from citizen)
total walking km per month = number of inspectors * walking km per
inspector
issue spots found by inspectors = Undetected spots * (total walking km
per month/street network size)
new reported spots = issue spots found by citizen + issue spots found

237
Appendices

by inspectors issue spots found by citizen * issue spots found by

inspectors / Undetected spots
Issue detection rate = new reported spots / Undetected spots
Reported spots = INTEG (new reported spots-repaired spots, 0)
repaired spots = Reported spots / repair duration

COST MODEL
costs for mobile service = IF THEN ELSE( Time> = implementation time,
(IF THEN ELSE(Time = implementation time, implementation costs,
ongoing costs)),0)
costs for reporting = (costs per mobile report*reports from
adopters)+(costs per traditional report*reports via traditional
way)
costs of inspectors = cost per inspector per month*number of
inspectors
monthly costs = costs for mobile service+costs for reporting+costs of
inspectors

238
 4 Workshop Documentation (Excerpt)

4 Workshop Documentation (Excerpt)

Workshop: Konzeption eines mobilen

Meldedienstes in der stdtischen Instandhaltung

Stadt Saarbrcken und Humboldt-Universitt zu Berlin

Martin Weinberg, Till Winkler, Philipp Eckert, Philipp Junghanns
Saarbrcken, 16. Mai 2011

Workshop-Dokumentation

Dez VII, Stabsstelle IT-Koordination

239
Appendices

AGENDA

Nr. Zeit Agendapunkt Moderation

1. 9:30 Begrung, Vorstellungsrunde und Erwartungen an den Workshop Weinberg

2. 9:45 Prsentation bisheriger Arbeiten im Bereich mobiler Meldedienste Winkler,

Wirtschaftlichkeitsanalyse Eckert und
Prototypische Implementierung (inkl. Live-Demo) Junghanns
3. 10:15 Aufgabenstellung fr die Arbeit in Teilgruppen Winkler
Vorstellung Ist-Prozess
4. 10:30 Teilgruppe A: Prozessmodellierung Teilgruppe B: Aspekte der Umsetzung A: Eckert
Besprechung von IST-Prozessen Informationsobjekte B: Junghanns
der stdtischen Instandhaltung Rechtliche Aspekte (Datenschutz)
Erarbeitung von SOLL-Prozess mit Technische Umsetzung
mobilem Meldedienst
12:30 Mittagspause

5. 13:30 Gegenseitige Vorstellung der Ergebnisse aus den Teilgruppen (jew. 5 min.)
Zuordnung der Umsetzungsaspekte auf den SOLL-Prozess Winkler
6. 14:00 Zusammenfassung
Nchste Schritte Winkler,
Verabschiedung Weinberg
14:30 Ende des Workshops

Seite 2 HU-IWI 2011 Eckert, Junghanns, Winkler

Es gibt bereits eine Reihe von Anbietern derartiger Dienste

Die Akzeptanz in Deutschland ist allerdings berschaubar
Marktberblick
Name (Land) Betreiber Untersttzte Plattformen Monatl. Kosten Kommetar

NY311 (USA) ffentl. Webseite, iPhone [Link]. Nur New York

Seeclickfix (USA) Privat Webseite, BlackBerry, Pro Verion: 29 Euro Nur USA
iPhone, Android Plus Version: 72 Euro/
100k Einwohner
Connect Version: n. bek.
Citysourced (USA) Privat Webseite, BlackBerry, [Link]. Nur USA
iPhone, Android, Windows
Phone 7
Fixmystreet (GB) ffentl./ Webseite, iPhone, Android kostenfrei Community-
Privat Plattform, kein
Redaktionssystem
Mngelmelder (D) Privat Webseite, iPhone [Link]. Brger mssen sich
registrieren
Mark-a-Spot (D) Privat Webseite, iPhone, Android, Community Vers. Bisher geringe
BlackBerry, PalmOS, kostenfrei, Enterprise Verbreitung
Windows Phone, Symbian, Vers. ab 3178 Euro,
MeeGo, bada Updates + Hosting: 119
Euro
Maerker ffentl. Webseite [Link]. Nur Brandenburg,
Brandenburg (D) keine mobile Lsung

Seite 3 HU-IWI 2011 Eckert, Junghanns, Winkler

240
 4 Workshop Documentation (Excerpt)

Eine Modellrechnung zeigt, dass die Stadt Saarbrcken durch einen

solchen Dienst pro Monat ca. 5-10% mehr Probleme auffinden knnte
Wirtschaftlichkeitsanalyse
Adopters of the service
Modellannahmen 200,000

175.000 Einwohner
150,000

6 Inspektoren (Straenlufer)
100,000
665 km Strassennetz
10% der Handybesitzer senden einmal im
50,000

Jahr eine Meldung 0

1 16 31 45 60
berlagerung von Meldungen bercksichtigt Time (Month)

Adopters of the service : With_Mobile_Reports_ai=10_inp=6 people

Issue detection rate

Modellergebnis 0.6

Mgliche Verbesserung der

0.5
Problemfindungsrate um 5-10%
Einschrnkung: Auswirkungen auf die 0.4

Schadensbehebung nicht bercksichtigt

0.3

0.2
1 16 31 45 60
Time (Month)

Issue detection rate : With_Mobile_Reports_ai=10_inp=6 Dmnl

Issue detection rate : Without_Mobile_Reports_ai=10_inp=6 Dmnl

Seite 4 HU-IWI 2011 Eckert, Junghanns, Winkler

Prozesse - bersicht
Workshop-Ergebnis

HU-IWI 2011 Eckert, J(uV

ne
ghr
annnd
seruinnkgleer n
,W gegenber IstProzess sind grn eingefrbt bzw. ausgegraut)

241
Appendices

Beim Thema Vorfall-Meldung durch Brger sind eine Reihe von

Stellen involviert
Backup: Organisatorische Sicht

Zentrale
Beschwerdestelle

Straenerhaltung

Eingreiftruppe
Beschilderung

Seite 6 HU-IWI 2011 Eckert, Junghanns, Winkler

242
 5 SaaS Interview Guideline (Shortened)

5 SaaS Interview Guideline (Shortened)

Einfhrung
Willkommen und vielen Dank dass Sie sich die Zeit nehmen
Wer ich bin
Hintergrund des Interviews
Struktur des Interviews
Fragen und wie antworten
Vertraulichkeitsvereinbarung
Kontext (Context)
Wie wrden Sie Ihre Position in Ihrer Organisation beschreiben
Wie kann ich mir die IT in Firma XY in etwa vorstellen
Was ist Ihre Erfahrung mit SaaS
Entscheidung (Adoption)
Seit wann wird Salesforce CRM in Ihrem Unternehmen eingesetzt?
Wie war die Situation vorher?
Gibt es weitere SaaS Produkte?
Was war die Motivation zur Einfhrung?
Was waren die Vorbehalte gegen eine Einfhrung?
Knnen Sie (heute noch) den Auswahl- und Entscheidungsprozess fr die SaaS-
Software beschreiben?
SaaS-Einfhrung (Adoption)
Knnen Sie kurz schildern wie die Einfhrung von Salesforce von statten ging?
Inwieweit unterschied sich die Einfhrung von der einer On-premise Software?
Gab es bestimmte Dinge die bei der Einfhrung nicht unmittelbar bedacht worden
sind und dann mittelfristig verndert werden mussten?
SaaS-Auswirkungen Kurzfristige Vernderungen (Impact)
Wie wrden Sie generell die wichtigsten Auswirkungen der Nutzung von Salesforce
als SaaS-Produkt beschreiben?
Waren mehr nderungen auf IT Seite oder auf Fachseite notwendig?
Welche organisationalen Manahmen waren auf IT- und Fachseite notwendig, um
die SaaS-Software einzufhren insbesondere im Unterschied zu herkmmlicher
Software?
Wenn sie diese nderungen mit herkmmlicher on-premise Software vergleichen,
inwiefern wrden Sie dort einen Unterschied sehen?
SaaS-Auswirkungen Langfristige Vernderungen (Impact)
Welche langfristigen Vernderungen gibt es durch Nutzung der SaaS-Software?

243
Appendices

Haben sich darber hinaus mit Einfhrung von SaaS langfristige organisationale
Vernderungen innerhalb der IT bzw. zwischen IT und Business ergeben?
Wenn ja, welche? (In den oben gegebenen Kategorien)
Hat sich das Budget fr Applikationen durch SaaS gendert? In welche Richtung?
Hat sich die soziale Integration zwischen Business und IT verndert?
Hat sich die Rolle der gesamten IT durch die SaaS-Nutzung verndert? Inwiefern?
Hat die Einfhrung von der ersten SaaS-Lsung (Salesforce) zur Folge, dass das
Business (auch) in Zukunft strker auf on-Demand Lsungen setzen mchte?
Ist die Gesamt-Entscheidungshoheit durch SaaS strker ins Business gerutscht?
nderungen im Verhltnis mit dem Supplier (Salesforce) im Vergleich zu anderer
Software
Ergebnis (Success)
Inwieweit wurden die Ziele, die durch die Nutzung von SaaS angestrebt wurden,
erreicht? (Punkte von oben aufgreifen!)
Inwieweit konnten Kosten eingespart werden?
In welchem Bereich? Wie hoch?
Wie erfolgreich wrden Sie die Einfhrung der SaaS-Software insgesamt bewerten?
Schluss
Gibt es sonst noch Anmerkungen von Ihrer Seite / Aspekte die wir bisher nicht
berhrt haben, die aber von Ihrer Seite aus wichtig erscheinen?
Gibt es noch jemanden in Ihrer Organisation, mit dem wir zu diesem Thema
sprechen sollten?
Vielen Dank fr Ihre Zeit und Bereitschaft
Die Ergebnisse der Studie werden wir Ihnen gerne zukommen lassen.

244
 6 Business and Information Technology Survey (Supplements)

6 Business and Information Technology Survey (Supplements)

6.1 Cover Letter

%XVLQHVV DQG ,QIRUPDWLRQ 7HFKQRORJLHV

6WXGLH 2011

)LUPD
$GUHVVH7LWHO 9RUQDPH 1DFKQDPH
3RVLWLRQ
$GU=HLOH1
European School of
$GU=HLOH2 Management and Technology
3URI. 'U. )UDQFLV %LGDXOW
12. $SULO 2011 )XOO 3URIHVVRU

6HKU $QUHGH*HHKUW $QUHGH+HUU)UDX $QUHGH7LWHO 1DFKQDPH,

ZLU P|FKWHQ 6LH KHU]OLFK HLQODGHQ, DQ GHU GLHVMlKULJHQ %XVLQHVV DQG 7HFKQRORJLHV 6WXGLH
(%,7) WHLO]XQHKPHQ, GLH JHPHLQVDP YRQ (XURSHDQ 6FKRRO RI 0DQDJHPHQW DQG 7HFKQRORJ\
((607) XQG +XPEROGW-8QLYHUVLWlW ]X %HUOLQ GXUFKJHIKUW ZLUG.

%,7 ZXUGH 2003 LQV /HEHQ JHUXIHQ XQG KDW VLFK VHLWGHP ]X HLQHU GHU IKUHQGHQ LQWHUQDWLRQDOHQ
Humboldt-Universitt zu Berlin
XQG LQGXVWULHEHUJUHLIHQGHQ ,7-0DQDJHPHQW-6WXGLHQ HQWZLFNHOW. /DQJIULVWLJHV $QOLHJHQ GHV
%,7-3URMHNWHV LVW HV, GLH $XVZLUNXQJHQ QHXHU ,QIRUPDWLRQV- XQG .RPPXQLNDWLRQVWHFKQRORJLHQ 3URI. 2OLYHU *QWKHU, 3K.'.
DXI 8QWHUQHKPHQVSUDNWLNHQ ZHOWZHLW ]X XQWHUVXFKHQ. 'HNDQ GHU :LUWVFKDIWVZLVVHQ-
VFKDIWOLFKHQ )DNXOWlW
'LH GLHVMlKULJH 6WXGLH VHW]W HLQHQ 6FKZHUSXQNW DXI GDV 7KHPD ,7-,QGXVWULDOLVLHUXQJ DP %HLVSLHO
YRQ 6RIWZDUH-DV-D-6HUYLFH (6DD6). =LHO LVW KHUDXV]XILQGHQ, LQZLHZHLW VLFK GXUFK GLH 1XW]XQJ
YRQ 6DD6, LP 9HUJOHLFK ]X WUDGLWLRQHOOHQ VRJHQDQQWHQ "2Q-SUHPLVHV"-$QZHQGXQJHQ,
,7-9HUDQWZRUWOLFKNHLWHQ XQG GLH 5ROOH GHV )DFKEHUHLFKV YHUlQGHUQ.
.RQWDNW:
'DV $XVIOOHQ GHV )UDJHERJHQV QLPPW FD. 45 0LQXWHQ LQ $QVSUXFK. (U JOLHGHUW VLFK LQ IROJHQGH 7LOO :LQNOHU
QHXQ $EVFKQLWWH:
+XPEROGW-8QLYHUVLWlW ]X %HUOLQ
,. 8QWHUQHKPHQV- XQG 7HLOQHKPHUSURILO
:LUWVFKDIWVZLVVHQVFKIWO. )DNXOWlW
,,. *HVFKlIWV]LHOH XQG *HVDPWRUJDQLVDWLRQ
,,,. ,7-=LHOH XQG ,7-2UJDQLVDWLRQ ,QVWLWXW IU
,9. 6DD6/2Q-SUHPLVH $QZHQGXQJVEHLVSLHO (6FKZHUSXQNWWKHPD) :LUWVFKDIWVLQIRUPDWLN
9. 2XWVRXUFLQJ YRQ ,7- XQG *HVFKlIWVSUR]HVVHQ 6SDQGDXHU 6WU. 1
9,. 7HFKQRORJLHHLQIKUXQJ 10178 %HUOLQ
9,,. *HVFKlIWVHUJHEQLVVH XQG *HVFKlIWVSDUWQHU (:DKOWHLO)
9,,,. .XQGHQVHLWLJH %H]LHKXQJHQ (:DKOWHLO) 7HOHIRQ +49 >30@ 2093-1582
,;. $UEHLWVSODW] XQG 2UJDQLVDWLRQ (:DKOWHLO) 7HOHID[ +49 >30@ 2093-5741

$OOH ,QIRUPDWLRQHQ ZHUGHQ VHOEVWYHUVWlQGOLFK VWUHQJ YHUWUDXOLFK EHKDQGHOW XQG QXU [Link]#[Link]
DQRQ\PLVLHUW DXVJHZHUWHW. 6ROOWHQ 6LH )UDJHQ ]XP $XVIOOHQ GHV )UDJHERJHQV KDEHQ, ZHQGHQ
6LH VLFK ELWWH DQ QHEHQVWHKHQGH .RQWDNWSHUVRQHQ. 6LH N|QQHQ HEHQIDOOV JHUQH GLH 0|JOLFKNHLW
ZDKUQHKPHQ, GHQ )UDJHERJHQ (DXI 'HXWVFK RGHU (QJOLVFK) RQOLQH ]X EHDQWZRUWHQ XQWHU GHU
:HE-$GUHVVH: [Link]
'DV =XJDQJVSDVVZRUW KLHUIU ODXWHW: ELWVWXGLH 9HUWUHWXQJ:
$QQD-/HQD %XMDUHN
$OV 'DQNHVFK|Q IU ,KUH 7HLOQDKPH YHUJHEHQ ZLU 100 +DQG\-1RWIDOO-/DGHJHUlWH DQ GLH HUVWHQ
(LQVHQGHU XQG YHUORVHQ HLQ QHXHV $SSOH L3DG 2 XQWHU DOOHQ 7HLOQHKPHUQ, GLH GHQ )UDJHERJHQ 7HOHIRQ +49 >30@ 2093-5806
RUGQXQJVJHPl DXVJHIOOW KDEHQ. =XGHP VWHOOHQ ZLU ,KQHQ JHUQH GLH (UJHEQLVVH GHU 6WXGLH ]XU 7HOHID[ +49 >30@ 2093-5741
9HUIJXQJ. +LQWHUODVVHQ 6LH KLHUIU ELWWH ,KUH .RQWDNWGDWHQ DXI GHU 5FNVHLWH GHV )UDJHERJHQV.
EXMDUHN#[Link]
:LU VLQG ,KQHQ VHKU GDQNEDU, GDVV 6LH GLH =HLW ]XP $XVIOOHQ GLHVHV )UDJHERJHQV ILQGHQ, XQG
ZLVVHQ ,KUH 0KHQ ]X VFKlW]HQ.

0LW IUHXQGOLFKHQ *U
HQ, :HLWHUH ,QIRUPDWLRQHQ:
[Link]

copying prohibited copying prohibited

3URI. 'U. )UDQFLV %LGDXOW 3URI. 2OLYHU *QWKHU, 3K.'.

(XURSHDQ 6FKRRO RI 0DQDJHPHQW DQG 7HFKQRORJ\ +XPEROGW-8QLYHUVLWlW ]X %HUOLQ

245
Appendices

246
 6 Business and Information Technology Survey (Supplements)

6.2 Questionnaire

Business and Information Technologies

Studie 2011 / Fragebogen

Bitte senden Sie diesen Fragebogen ausgefllt bis zum 15. Mai 2011 an die rckseitig stehende
Adresse. Sie knnen ebenfalls gerne online an der Studie teilnehmen unter: [Link]/uc/bit
Das Zugangspasswort lautet: bitstudie
247
Appendices

Unternehmens- und Teilnehmerprofil

I.a Unternehmensprofil

1. Bitte whlen Sie die Branche Ihres Unternehmens

(Whlen Sie die am besten zutreffende Alternative aus)

Automobilbau & Zulieferer Logistik & Transport

Banken Maschinenbau
Bauwesen & Immobilien Metall & Rohstoffe
Chemie & Pharma ffentliche Verwaltung
Dienstleistungen Telekommunikation & Medien
Einzelhandel Textil & Mode
Elektronik & High-Tech Tourismus & Gastronomie
Energie & Versorger Versicherungen
Gesundheit Andere (bitte nachfolgend angeben)
Lebensmittel & Landwirtschaft

2. Bitte beschreiben Sie die Hauptaktivitten Ihres Unternehmens kurz in eigenen Worten

3. Bitte ordnen Sie Ihr Unternehmen anhand der folgenden Gegensatzpaare ein
(Whlen Sie den Mittelpunkt der Skala, falls beide Begriffe gleichermaen zutreffen)

Fertigungswirtschaft Dienstleistungsbranche
Physische Produkte Informationsprodukte
Business-to-Consumer Business-to-Business

4. Bitte geben Sie die Anzahl an Mitarbeitern gesamt in Ihrem Unternehmen an sowie die Anzahl der Mitarbeiter in der IT
Bei global ttigen Unternehmen / Grokonzernen: Betrachten Sie hier bitte die grtmgliche Einheit des Gesamtunternehmens, zu der Sie zuverlssig
Angaben machen knnen (zum Beispiel: Global, nur Deutschland oder nur ein bestimmter Geschftsbereich). Beziehen Sie Ihre Antworten fr alle
anderen Fragen auf genau diese Betrachtungseinheit als "Ihr Unternehmen".

Mitarbeiter gesamt Mitarbeiter in der IT

5. Bitte schtzen Sie den Umsatz und das IT Budget Ihres Unternehmens fr das Jahr 2010 (in Millionen Euro)
(Bei Banken und Versicherungen geben Sie bitte das Geschftsvolumen bzw. die Beitragssumme statt eines Umsatzes an)

Mio. EUR Umsatz Mio. EUR IT-Budget

I.b Teilnehmerprofil

6. Bitte nennen Sie Ihre Position oder Ttigkeitsbezeichnung in Ihrem Unternehmen

7. Bitte geben Sie an, seit wann Sie in diesem Unternehmen arbeiten (in Jahren)

Jahre

8. Bitte whlen Sie die Position innerhalb Ihres Unternehmens, welche am besten auf Ihre Ttigkeit zutrifft (horizontal und vertikal)
Fachseite IT

Geschftsfhrer / Hchster IT-Entscheider

Geschftsbereichsleiter / Bereichsleiter in der IT
Fhrungskraft in einem Fachbereich / Fhrungskraft in der IT
Mitarbeiter in einem Fachbereich / Mitarbeiter in der IT

248
 6 Business and Information Technology Survey (Supplements)

Geschftsziele und Gesamtorganisation

II. Geschftsziele und Gesamtorganisation

9. Bitte vergleichen Sie Ihr Unternehmen mit den direkten Wettbewerbern im Bezug auf die folgenden Kenngren
(Beispiel: Die IT-Kosten pro Mitarbeiter in Ihrem Unternehmen sind geringer/hher/gleich im Vergleich zu denen der Wettbewerber)

etwas
viel geringer geringer gleich etwas hher hher viel hher
geringer
IT-Kosten pro Mitarbeiter
Eigenkapitalrentabilitt
(Gewinn / Eigenkapital)
Gesamtkapitalrentabilitt
(Gewinn / Gesamtkapital)
Umsatzrendite
(Gewinn / Umsatz)
Umsatzwachstum

10. Bitte charakterisieren Sie die Geschftsstrategie Ihres Unternehmens anhand der folgenden Gegensatzpaare

hnliche Mrkte Unterschiedliche Mrkte

(Spezialisierte Produkte) (Diversifizierte Produkte)
Skaleneffekte Verbundeffekte
(Mengenvorteile) (Cross-Selling-Vorteile)
Organisches Wachstum Anorganisches Wachstum
(Internes Wachstum) (Akquisitionen und Zukufe)

11. Bitte geben Sie an, in welchen Regionen Ihr Unternehmen derzeit ttig ist.
Planen Aufnahme von Keine Geschftsttigkeit
ben derzeit
Geschftsttigkeit innerhalb innerhalb der nchsten 3
Geschftsttigkeit aus
der nchsten 3 Jahre Jahre
Mittel- und Osteuropa (inkl. Deutschland)
Westeuropa (GB, Frankreich, Spanien, usw.)
Nordamerika
Lateinamerika
Afrika
Mittlerer Osten
Sdostasien

12. Bitte stufen Sie die folgenden Gren hinsichtlich der Entscheidungsstrukturen in Ihrer Organisation ein
sehr gering gering gering-mittel mittel mittel-hoch hoch sehr hoch
Der Grad zu dem Geschftsentscheidungen zentral
vom Management getroffen werden
Der Grad an Freiheit, den verschiedene
Geschftsbereiche in Entscheidungen haben
Die Autonomie, die leitenden Verantwortlichen bei
operativen Geschftsentscheidungen gegeben wird

13. Bitte schtzen Sie das geschftsspezifische Wissen und die Erfahrung der Mitarbeiter in der IT ein
sehr gering gering gering-mittel mittel mittel-hoch hoch sehr hoch
Das Wissen, das IT-Mitarbeiter ber die
Geschftsprozesse haben ist...
Die Erfahrung, die IT-Mitarbeiter in Aktivitten des
Kerngeschfts haben, ist...

14. Bitte schtzen Sie nun das IT-Wissen und IT-Erfahrung der Mitarbeiter in den Fachbereichen ein
sehr gering gering gering-mittel mittel mittel-hoch hoch sehr hoch
Das Wissen, das der Fachbereich von IT
Management, Technologien und Anwendungen
hat, ist...
Die Erfahrung, die Mitarbeiter im Fachbereich in
IT-Projekten gemacht haben, ist...

249
Appendices

IT-Ziele und IT-Organisation

III. IT-Ziele und IT-Organisation

15. Inwieweit stimmen Sie folgenden Aussagen bezglich der IT-Ziele Ihres Unternehmens zu
stimme
berhaupt stimme nicht stimme eher stimme eher stimme zu stimme sehr
neutral
zu nicht zu zu zu
nicht zu
Die IT sollte die vereinbarten Dienste kosten-
effizient anbieten und liefern
Die IT sollte effizient darin sein, IT-Anwendungen
zu betreiben und zu untersttzen
IT sollte dazu beitragen, die Leistungsfhigkeit
der Geschftsprozesse zu verbessern
IT sollte Prozesskosten und -durchlaufzeiten
senken sowie die Qualitt erhhen
IT sollte helfen, den Markt und die Reichweite
Ihres Unternehmens zu vergrern
IT sollte die Weiterentwicklung der Produkte
Ihres Unternehmens voran bringen

16. Bitte ordnen Sie die IT-Ziele Ihres Unternehmens zusammenfassend anhand der folgenden Gegensatzpaare ein
Gewinn optimieren Wachstum erzielen
Geringe Kosten des Geschftsbetriebs Hohe Geschftsinnovation
Betriebliche Leistungsfhigkeit Strategische Positionierung

17. An wen berichtet der hchste IT-Leiter in Ihrem Unternehmen?

an den Geschftsfhrer (CEO)
an den Finanzvorstand (CFO)

an andere, bitte hier angeben:

18. Falls es IT-Mitarbeiter gibt, die nicht diesem hchsten IT-Leiter unterstehen, sondern an dezentrale IT-Verantwortliche in den
Fachbereichen berichten, wie hoch ist dieser Anteil an "dezentralen" IT-Mitarbeitern?
(Bitte geben Sie eine Zahl zwischen 0 und 100 Prozent an, gemessen an der Gesamtanzahl von IT-Mitarbeiter)

19. Wer trifft in Ihrem Unternehmen generell die wichtigsten IT-Entscheidungen in Bezug auf...
Fachseite mit IT- Fachseite und IT IT mit Beteiligung
Fachseite Zentrale IT
Beteiligung gemeinsam der Fachseite
den Bedarf an IT-Anwendungen
(z.B. neue Anwendungen, funktionale Anforderungen)
die IT-Ausgaben
(z.B. IT-Budget und Priorisierung der IT-Investitionen)
die IT-Architektur
(z.B. eingesetzte Technologien, Anbieterauswahl,
Integrationsfragen)

20. Beurteilen Sie die allgemeine Leistungsfhigkeit der IT-Organisation

niedrig-
sehr niedrig niedrig mittel mittel-hoch hoch sehr hoch
mittel
Die Effizienz der IT-Organisation in der Ausfhrung
ihrer Arbeit ist...
Die Qualitt der von der IT angebotenen Dienste
ist...
Die Untersttzung der Geschftsprozesse durch
die IT ist...
Die Zufriedenheit des Fachbereichs mit der
Arbeit der IT ist...

250
 6 Business and Information Technology Survey (Supplements)

SaaS/On-Premise Anwendungsbeispiel

IV. SaaS/On-Premise Anwendungsbeispiel (Schwerpunktthema)

Definitionen: Wir verstehen Software-as-a-Sevice (SaaS) als jegliche Unternehmenssoftware, die von einem externen Anbieter fr viele
Kundenorganisationen bereitgestellt wird und von den Mitarbeitern Ihres Unternehmens ber das Internet im Web Browser genutzt werden
kann.

Im Gegensatz dazu sprechen wir von einer On-premises-Anwendung, wenn die Unternehmenssoftware auf den Rechnern der
Betriebssttten Ihres eigenen Unternehmens (engl. premises) auf herkmmliche Weise installiert und betrieben wird, und diese Installation
auch ausschlielich von Ihrem Unternehmen genutzt wird.

21. Nutzt ihr Unternehmen Software-as-a-Service?

Ja Nein

Falls ja, dann stellen Sie sich jetzt bitte die SaaS-Anwendung vor, die in Ihrem Unternehmen genutzt wird. Im Fall von mehreren mglichen
Anwendungen, whlen Sie bitte die wichtigste oder die, mit der Sie am besten vertraut sind.

Falls nein, so whlen Sie bitte an eine herkmmliche, on-premises-Anwendung, die in Ihrem Unternehmen im Einsatz ist. Da hier vermutlich
mehrere in Frage kommen, whlen Sie bitte die wichtigste oder die, mit der Sie am besten vertraut sind.

22. Was ist die genaue Bezeichnung fr die von Ihnen gewhlte Anwendung

23. Bitte besttigen Sie hier hier erneut, ob es sich bei dieser Anwendung um SaaS oder On-premises handelt
SaaS On-premises

24. Welcher Anwendungszweck charakterisiert diese Anwendung am besten?

Business Intelligence & Analytics Human Resource Management (HRM)
Communications & Collaboration Office- & Productivity-Anwendungen
Customer Relationship Management (CRM) Production Execution
Digital Content Creation (DCC) Service Management
Engineering & Design Supply Chain Management (SCM)
Enterprise Resource Planning (ERP) Andere

25. Seit wieviel Jahren nutzt Ihr Unternehmen diese Anwendung?

Jahre

26. Wie lange hatte die Einfhrung dieser Anwendung gedauert (von der Anbieterentscheidung bis zum Go-Live, in Monaten)?

Monate

27. Hat diese Anwendung eine vorherige Lsung abgelst oder wurde sie als neue Lsung eingefhrt?
Ablsung einer vorherigen Lsung
Einfhrung als neue Lsung

28. Bitte spezifizieren Sie hinsichtlich der Einfhrung dieser Anwendung...

Fachbereich mit Fachbereich und IT mit Beteiligung
Fachbereich Zentrale IT
IT-Beteiligung IT gemeinsam des Fachbereichs
Aus welchem Teil der Organisation stammte die
initiale Idee, diese Anwendung einzufhren
Welcher Teil der Organisation war treibende Kraft
whrend der Einfhrung dieser Anwendung

29. Bitte schtzen Sie die Anzahl der Nutzer der genannten Anwendung

Nutzer

30. Bitte geben Sie den Umfang der Nutzung dieser Anwendung an
einzelnen mehreren einem Geschfts- mehreren unternehmens-
[Link] [Link] bereich Geschftsbereichen weit
Die Anwendung wird genutzt in...

251
Appendices

SaaS/On-Premise Anwendungsbeispiel

31. Bitte bewerten Sie folgende Charakteristika der Anwendung (im Vergleich zu anderen Anwendungen in Ihrem Unternehmen)

sehr gering gering gering-mittel mittel mittel-hoch hoch sehr hoch

Der Anteil an Mitarbeitern in Ihrem Unternehmen,
die diese Anwendung nutzen
Die Hufigkeit, mit der diese Anwendung im
Tagesgeschft genutzt wird
Die Menge an Trainings die notwendig waren, um
die Nutzer auf diese Anwendung zu schulen
Der Aufwand, um diese Anwendung (initial und
kontinuierlich) an Ihr Unternehmen anzupassen
Der Grad, zu dem diese Anwendung auf die
Prozesse Ihres Unternehmens angepat wurde
Der Schwierigkeitsgrad, um eine Anpassung an
dieser Anwendung vorzunehmen
Die Anzahl von erforderlichen Mitarbeitern mit
speziellen Qualifikationen, um diese Anwendung zu
betreiben
Der Grad, zu dem diese Anwendung technisch mit
der restlichen Anwendungslandschaft integriert ist

32. Przisieren Sie bezglich der letzen Frage bitte die Anzahl und den Typ von Schnittstellen, die diese Anwendung mit anderen
Systemen hat
Keine Unidirektionale Bidirektionale
Mehrere Hochintegrierte
Schnittstelle Schnittstelle Schnittstelle
Schnittstellen Anwendung
(stand-alone) (one-way) (two-way)
Anzahl und Typ von Schnittstellen

Die nun folgenden Fragen beziehen sich auf die Aufteilung von Entscheidungs- und Aufgabenverantwortlichkeiten zwischen
Fachbereich und IT, um die genannte Anwendung zu betreiben

33. Wer entscheidet ber...

Fachbereich mit Fachbereich und IT mit Beteiligung
Fachbereich Zentrale IT
IT-Beteiligung IT gemeinsam des Fachbereichs
nderungen an der Anwendung
(z.B. Freigaben fr Change Requests oder
Anpassungen)
IT-Ausgaben fr diese Anwendung
(z.B. fr Lizenzen, Erweiterungen, Wartung)
Architekturfragen bezglich dieser Anwendung
(z.B. Integration mit anderen Systemen, genutzte
Infrastrukturkomponenten)

34. Wer fhrt die folgenden Aktivitten durch...

(Hinweis: Falls die jeweilige Aufgabe von einem externen Unternehmen geleistet wird, vermerken Sie dies bitte in der nchsten Frage und whlen hier
denjenigen Teil der Organisation, der fr die Steuerung dieses Dienstleisters verantwortlich ist)

Fachbereich mit Fachbereich und IT mit Beteiligung

Fachbereich Zentrale IT
IT-Beteiligung IT gemeinsam des Fachbereichs
nderungen an der Anwendung
(z.B. Umsetzung von Anpassungen und Customizing)
1st Level-Support fr diese Anwendung
(z.B. Beantwortung von Nutzeranfragen, Incident
Management)
2nd-Level-Support fr diese Anwendung
(z.B. Auffinden technischer Fehler, Problembehebung)

35. Bitte whlen die folgenden Checkboxen an, falls eine der genannten Ttigkeiten durch einen externen Dienstleister erbracht wird
Umsetzung von nderungen durch externen Dienstleister
1st-Level-Support durch externen Dienstleister
2nd-Level-Support durch externen Dienstleister

252
 6 Business and Information Technology Survey (Supplements)

SaaS/On-Premise Anwendungsbeispiel Outsourcing von IT- und Geschftsprozessen

36. Inwieweit wrden Sie den folgenden Aussagen bezglich der Support-Organisation fr die genannte Anwendung zustimmen
(Hinweis: Je nach Beantwortung der vorangegangenen zwei Fragen kann sich der Begriff "IT-Support " sowohl auf Fachbereichs- als auch auf IT-
Mitarbeiter oder Externe beziehen)

stimme
stimme eher stimme zu stimme sehr
berhaupt stimme nicht stimme eher neutral
zu nicht zu zu zu
nicht zu
Der IT-Support fr diese Anwendung ist fr die
Nutzer erreichbar und reagiert schnell
Der IT-Support fr diese Anwendung ist zuverlssig
und kompetent fr die auftretenden Probleme
Der IT-Support versteht die Bedrfnisse der Nutzer
zur Weiterentwicklung der Anwendung
Der IT-Support dieser Anwendung ist in der Lage,
Nutzeranforderungen umzusetzen und neue
Funktionalitt bereit zu stellen
Die Nutzer der Anwendung sind zufrieden mit dem
IT Support fr diese Anwendung

37. Bitte vergleichen Sie die tatschlichen Ergebnisse der Nutzung heute mit den erwarteten Ergebnissen vor der Einfhrung der
genannten Anwendung, im Bezug auf folgende Gren
(Beispiel: Die tatschlichen Kosten heute sind kleiner/grer/gleich wie die zuvor erwarteten Kosten)

etwas etwas
viel kleiner kleiner gleich grer viel grer
kleiner grer
Die Kosten (Implementierung und Betrieb) fr
diese Anwendung
Der betriebliche Nutzen durch diese Anwendung
Die Auswirkungen auf das Umsatzwachstum
durch diese Anwendung

V. Outsourcing von IT- und Geschftsprozessen

38. Bitte schtzen Sie das Budget fr IT-Outsourcing (ITO) insgesamt als Anteil des gesamten IT-Budgets fr den IT-Betrieb
(Geben Sie eine Zahl zwischen 0 und 100 Prozent an)

% von IT-Budget fr ITO (z.B. fr Anwendungsentwicklung, Rechenzentrum, Netzwerk)

39. Bitte schtzen Sie das Budget fr Business Process Outsourcing (BPO), also das Auslagern kompletter Geschftsprozesse,
gemessen am Umsatz Ihres Unternehmens
(Geben Sie eine Zahl zwischen 0 und 100 Prozent an)

% von Umsatz fr BPO (z.B. fr Call Center, Lohnabrechnung, Buchhaltung)

40. Welche der folgenden Bereiche sind zurzeit von Ihrem Unternehmen ausgelagert? Bitte whlen Sie fr jeden Bereich die am ehesten
zutreffende Spalte.
Zur Zeit nicht
ausgelagert,
Zur Zeit nicht Auslagerung Teilweise Weitestgehend
ausgelagert jedoch innerhalb ausgelagert Nicht zutreffend
ausgelagert
der nchsten 3
Jahre geplant
ITO: Anwendungsentwicklung
ITO: Rechenzentrum
ITO: Netzwerk-Management
ITO: Daten-Management
ITO: Nutzerbetreuung (z.B. Help Desk)
BPO: Kundenbetreuung (z.B. Call Center)
BPO: Lohn- und Gehaltsabrechnung
BPO: Marktforschung
BPO: Buchhaltung
BPO: Finanzen
BPO: Auftragsabwicklung
BPO: Ausschreibungs- und Vertragsmanagement

253
Appendices

Technologieeinfhrung

VI. Technologieeinfhrung

41. Beschreiben Sie die Entwicklung Ihres Budgets ber die letzten 3 Jahre fr folgende IT-Investitionen Ihrer Organisation.
(Bitte whlen Sie nicht zutreffend", falls keine der mglichen Antworten zutrifft)

Stark Stark Nicht

Gesunken Unverndert Gestiegen gestiegen zutreffend
gesunken
Service-Vertrge, Serverbetrieb, Application
Integration, etc.
Software-as-a-Service / On-Demand Computing
Auslagerung von Geschftsprozessen (Business
Process Outsourcing)
Software: Anwendungen
Software: Sicherheit
Software: Betriebssysteme & Netzwerke
Intra- und Extranet
Hardware: Speicher
Hardware: Sicherheit
Kabellose Infrastruktur
Notfallwiederherstellung

42. Bitte beschreiben Sie den Status der Einfhrung und Nutzung der folgenden Anwendungen und Technologien in Ihrer Organisation
(Whlen Sie die am ehesten zutreffende Option)
Keine Einfhrung
Einfhrung innerhalb
Gegenwrtig im innerhalb der
der nchsten 3 Jahre Nicht zutreffend
Einsatz nchsten 3 Jahre
geplant
geplant
Storage Area Networks (SAN) und Network
Attached Storage (NAS)
Hard- und Software fr drahtlose Netzwerke
(Wi-Fi, Wireless LAN, etc.)
Authentifizierung und Verifikation durch Dritte
(Versign, etc.)
Identittsmanagement-Lsungen
(z.B. Single Singh-On
berwachungssysteme
Biometrie
Radio Frequency Identification (RFID)
Open Source Software
Enterprise Resource Planning (ERP)
Supply Chain Management (SCM)
Enterprise Application Integration (EAI) und
Middleware (auch: SOA)
Business Process Modeling
Business Intelligence / Data Warehouse
Content Management
E-L earning
Groupware und Organisationssoftware (Lotus
Notes, Exchange, etc.)
Enterprise Instant Messaging (IM)
Kollaborations- und Portalsoftware (Dokumenten-
management, Portale, Kollaboration, etc.)
Webseite und E-Commerce
Social Web (Facebook-Reprsentanz)

Vielen Dank, Sie haben hiermit den Pflichtteil der Umfrage beendet!

Von den nchsten drei Abschnitten, whlen Sie bitte einen Abschnitt, der am besten auf Ihr Unternehmen zutrifft, und fllen nur diesen aus
- VII. Geschftsergebnisse und Geschftspartner (zutreffend z.B. fr Unternehmen der Fertigungswirtschaft) Seite 9
- VIII. Kundenseitige Beziehungen (zutreffend fr z.B. fr Unternehmen, die E-Commerce nutzen) Seite 9 unten
- IX. Arbeitsplatz und Organisation (zutreffend fr alle sonstigen Unternehmen) Seite 11

(Selbstverstndlich drfen Sie auch mehrere Abschnitte bearbeiten.)

254
 6 Business and Information Technology Survey (Supplements)

Dankeschn

Vielen Dank fr Ihre Teilnahme an der BIT-Studie 2011. Als kleines Dankeschn vergeben wir 100 Notfall-Ladegerte frs Handy an die
ersten Einsender und verlosen unter allen Teilnehmern ein Apple iPad 2. Zudem stellen wir Ihnen gerne die Ergebnisse der diesjhrigen
Studie zur Verfgung.

Besttigen Sie hier, falls Sie hiermit einverstanden sind und geben Sie bitte unten Ihre Kontaktdaten an (eine Weitergabe an Dritte ist
selbstverstndlich ausgeschlossen).

Ja, ich mchte an der Verlosung eines Apple iPad 2 teilnehmen

Ja, bitte senden Sie mir bitte die Ergebnisse der diesjhrigen BIT-Studie zu

Name:

Unternehmen:

Adresse:

E-M ail:

Telefon (optional):

Falls Sie Fragen oder Anregungen bezglich der Umfrage haben, knnen Sie uns diese gerne hier mitteilen:

Kontakt:

Till Winkler
Wissenschaftlicher Mitarbeiter

Humboldt-Universitt zu Berlin
Wirtschaftswissenschaftliche Fakultt
Institut fr Wirtschaftsinformatik
Spandauer Str. 1
10178 Berlin

Telefon: +49 (0) 30 2093-1582

Telefax: +49 (0) 30 2093-5741

Email: [Link]@[Link]

255
 6 Business and Information Technology Survey (Supplements)

6.3 Sample Description

Preamble

The quantitative empirical research on Software as a Service conducted in this disser-

tation (i.e., Chapters5.2and5.4) is based on a sample acquired in course of a larger
survey: the Business and Information Technologies (BIT) Project 2011 in Germany. This
Appendix is taken from the BIT project documentation byWinkler, Goebel, Bidault,
and Gnther(2013) and describes the detailed data acquisition procedures and sample
characteristics.

Data Selection and Acquisition

Data for the analyzes stems from a representative survey among German businesses. The
corresponding addresses were purchased from one of the leading publishers of company
information in Germany. We have chosen this particular database because it provides not
only comprehensive firm information but also the opportunity to address the head of IT
directly by name.
We restricted the survey to heads of IT in large private sector organizations. The
Germany Institute for Research on Small and Medium Enterprises defines a large com-
pany as an organization with more than 50 million Euro of revenues and more than
500 employees (IfM2002). A major reason for this restriction was that some of the
questions in the global BIT survey (e.g., those on globalization, e-commerce, governance and
organization) are more likely to apply to large businesses. Industry selection was based
on the 2008 edition of the German Classification of Economic Activities (WZ 2008).
All industry sectors except Public Administration, Defense, Social Insurance (84),
Education (85), Homes and Institutions (87-88), Private Households (97), and Extra-
Territorial Bodies (99) were included. A query to the database yielded in 3,285 results,
which were subsequently extracted.
The obtained contact data was cleaned and corrected. Particularly, in order to avoid
duplicate contacts, multiple entries for the same company were eliminated. Missing
addresses and forms for addressing the participants were completed and rectified. After
the process of data cleaning was completed, we kept 2,886 contacts to be addressed in a
mailing. As we considered this number appropriate to achieve the desired sample size (n
>200), no further random selection or sorting was performed.
Subjects were addressed in a formal mailing containing a cover letter, a paper version
of the questionnaire and a return envelope. Information regarding the survey was com-
plemented by a reference to the web page of the German BIT project. Answers could
be given either paper-based or online. The survey contained standard BIT questions as
well as questions regarding the focus topic and amounted to 10 pages in length in an A4-
booklet format. However, for reasons of convenience, only the focus part has been
declared as mandatory. The overall survey was tested to take approximately 45 minutes
to complete.

257
Appendices

As an incentive, a small gadget to charge a mobile phone (worth 5 Euro) was offered to
the first 100 respondents. Further, participants could enter a lottery for a modern tablet
computer (worth 500 Euro). Therefore participants had to leave their contact details at
the end of the questionnaire, as the survey itself did not contain any such information.
Invitations to participate were sent out mid April 2011, and the survey was closed end
of May 2011 after a two-week extension. Two reminders were sent to the corporate email
addresses of the companies, the first after two weeks and the second (final) reminder just
before the extension period. Those companies that could not be reminded via email (and
had not replied by then) were contacted over the phone. Out of these 1,018 reminder
calls, 257 (25%) of the companies could not be reached after several trials. Another
247 (24%) of the companies stated that they were not able to participate due to time
constraints or corporate policy. The majority of 429 (42%) companies agreed to accept a
formal reminder via email, 73 (7%) stated that they are planning to respond to the
survey (see Figure 2).
The total number of companies that reacted to our requests somehow amounted to 534
companies (19%). 29 of these companies returned the letter due to a recipient that is
unknown or has left the company. 90 of the companies stated via email, letter or fax, that
they were not willing or able to participate in the survey. After data cleaning, we counted
195 (6.8%) data sets from subjects that started the survey online but canceled. The
remaining 220 usable responses represent an overall response rate of 7.6% (see Figure 2).
Out of these usable responses, 89 (40%) came in as paper-based questionnaires which were
subsequently transferred to the online tool for the analysis. 131 (60%) of the
respondends directly used the online survey, out of whom 26 answered anonymously (see
Figure1).
Data cleaning was an important task to differentiate between anonymous and incom-
plete answers. We only kept answers from respondents who at least completed the
mandatory part of the questionnaire with a low number of missing values. 75% of
the remaining data records have less or equal than 3 (q3=3) missing values (median
m=1). Most of these missing values referred to fields with supposedly sensitive or
unknown information which were occasionally left blank, such as IT budget or company
performance.

Sample Characteristics

Respondent profiles As intended, the large majority of the respondents were high-
level IT decision-makers with significant work experience. The median working time
of respondents in their current company is 11.5 years. We surveyed the position of
the respondent on a horizontal (business/ IT) and a vertical dimension (top/ senior/
manager/ staff level). An additional text field regarding the job position allowed for a
further validity check. It revealed that many respondents rated themselves to pertain
to the second or third level, even if they actually held the highest position in IT (e.g.,
Head of IT). Also, there were 11 cases where the respondent left this field blank, for
example business executives responsible for IT (e.g., the Chief Financial Officer), or staff

258
 6 Business and Information Technology Survey (Supplements)

3000

1500
Delivere d
1730

Sent
1000 2886
Not
reached
Not 263
delivered Not
930 interested Unknown
247 recipient 29
500
Not interested 90
Send Canceled
195 Online
email anonymous 26
433 Usable
No email esponse Online 105
Will respond 75 r
0 220 Letter 89
Invitations Reminder emails Reminder calls Reactions Responses
(2886) (2660) (1018) (534) (220)

Figure 1:Survey reminders and response distribution

pertaining to an IT unit within the business organization (e.g., Demand Manager). For
the purpose of evaluation, these cases have been attributed to either side based on the
textual job descriptions. The distribution of work years and job positions are depicted
in Figure2.

Industry Characteristics The respondents industries represent a good sample of the

German private sector. Based on the first two digits of the Germany industry classifica-
tion code (WZ 2008), we clustered the invited companies into 18 industries which had
a relative frequency between 1% and 12% in our primary database (see Table3). The
participants were asked to classify their company based on this list.
Figure4shows the relative industry distribution of invited and respondent companies.
Some deviations are notable, for example for Construction & Real Estate (positive)
and Retail Trade (negative deviation). So one may conclude that these industries are
rather over- or underrepresented in the sample. However, testing the observed and the
expected industry frequency distribution by a chi-square test (=35.86; df =18; p <0.01)
does not reject the null hypothesis that the observed sample is consistent with the
industry distribution from the database. This speaks against a non-response bias of

259
Appendices

Table 3:Industry classification

Classiftcation code range Industry
01; 10 11 Food & Agriculture
13 15 Textile
19 21 Chemicals & Pharma
24 25 Metal & Raw Materials
26 27 Electronics & High-tech
28 Mechanical Engineering
29 30 Automotive
35 38 Energy & Utilities
41 43; 68; 80 81 Construction & Real Estate
45 47 Retail Trade
49 53 Logistics & Transportation
55 56; 79 Tourism & Gastronomy
58 63 Telecommunication & Media
64 Banking
65 66 Insurance
69 78 Professional Services
86 Health Care
All other Other

Automotive
Invitations in % Responses in %
Banking (n=2886) (n=220)

Chemicals & Pharma

Construction & Real Estate

Electronics & High-tech

Energy & Utilities

Food & Agriculture

Health Care

Insurance

Logistics & Transportation

Mechanical Engineering

Metal & Raw Materials

Professional Services

Retail Trade

Telecommunication & Media

Textile

Tourism & Gastronomy

Other

,0 2,0 4,0 6,0 8,0 10,0 12,0 14,0

Table 4:Respondent industries

260
 6 Business and Information Technology Survey (Supplements)

50 Senior Chief
Business Executive
45 Officer
Work years Management
Business 2% 3%
40
Manager
35 3%
Highest
30 IT staff Information
member Officer
25 3% 20%
20 IT Manager
20%
15
Senior IT
10
Management
5 49%

0
<= 2.0 2-5 5-10 10-15 15-20 20-25 25-30 30-35 >35

Figure 2:Respondent work years and position

certain industries in the sample.

Based on three bipolar scales we could further classify the companies according to
their main business activities (Figures4,4, and5). We see that 51% of respondents
are from manufacturing and 41% from service industries. The 8% who do not attribute
themselves to either category stem from companies who clearly have mixed activities,
for example utilities that produce and distribute energy (see Figure3).
The next question about the type of products is strongly correlating with the classifi-
cation in manufacturing versus services (Spearmans rank correlation =0.75). Most
companies in the sample (68%) deal with physical products (like in manufacturing)
rather than informational products (like in service industries). However, there are also
firms producing information products, for example a software vendors, as well as service
businesses who deal with physical products, e.g. retailers (see Figure4).
Finally, we checked whether the firms in the sample rather sell to corporate or private
customers. The majority (67%) focuses on business-to-business (see Figure5). This
is not surprising as most physical consumer products pass multi-tiered supply chains
spanning several businesses and retailers, before reaching the end consumer.

Firm Sizes Regarding firm sizes, most companies represented in the sample are rather
large, which is partly a result of our pre-selection. 91.8% of the companies have more than
500 employees, 95.3% have more than 50 milions Euros of revenues. Still, some companies
fall below these thresholds as we also selected companies from the database which fulfill
only one of these size criteria. The size comparison of invited (expected) and responding
(observed) companies reveals that large companies have responded over-proportionally
(see Figure6). The mean of employees is much higher for respondent ( mresp= 4,783)
than for invited firms (minv= 1,739). This difference is confirmed by a chi-square test
over the employee classes, which rejects the null hypothesis of consistent distributions
(=106.4; df =9; p <0.1). A reason for this may be that respondents themselves have

261
Appendices

100

50

0
Manufacturing industy Mixed Service industry

Figure 3:Manufacturing vs. service industries

150
100
50
0
Physical Mixed Informational
products products

Figure 4:Physical vs. informational products

100

50

0
Business to Mixed Business to
consumer business

Figure 5:Business to consumer vs. business to business

262
 6 Business and Information Technology Survey (Supplements)

50,0%
45,0%
40,0% Invited Responded
35,0%
30,0%
25,0%
20,0%
15,0%
10,0%
5,0%
0,0%
<200 200-500 500-1,000 1,0000-2,000 2,000-5,000 5,000-10,000 10,000-20,000 20,000-50,000 50,000-100,000 >100,000

Figure 6:Employee firm sizes (invited and responded)

70
60 Revenue Assets under management
50 (financial services)
40
30
20
10
0
<10 m EUR 30 - 100 m EUR 100 - 300 m 300 - 1000 m 1 - 3 bn EUR 3 - 10 bn EUR 10 - 30 bn EUR 30 - 100 bn EUR 100 - 300 bn
EUR EUR EUR

Figure 7:Financial firm size (revenues and assets under management)

recognized that the survey was directed to larger firms. Also larger firms may be more
accustomed to take part in such surveys and foresee better procedures to handle such
requests.
Regarding financial measures for firm size (revenues), we had to treat firms in the finan-
cial services sector separately. Since for banks and other financial institutions revenues
are not an appropriate measure of firm size, we asked for the business volume measured
by total assets under management of these companies. The resulting distributions are
depicted in Figure7using pseudo-logarithmic class sizes. Managed assets are usually
much higher than company revenues. The average revenue of companies in the samples is
mRev = 1.442 bn and mean assets under management mAssets= 11.526 bn Euros.

IT Organization Sizes Not surprising, measures describing the IT organization are

strongly correlated with company size. Both, IT employees and IT budget exhibit a
high and significant correlation with company employees (Pearson coefficient r=0.78 for
employees, r= 0.78 for budget). On average, companies employ 2.26% of their staff in
IT, and spend 1.71% of their annual revenues on IT. The distribution of IT employees
and IT budget are given in Figures8.
However, our results indicate that there are other factors explaining the level of IT
investments, as well. A logarithmic regression of company employees on IT employees,
respectively company revenues on IT budget only explains half of the variance in the
observed IT organization sizes (R2empl= 0.52; R2rev = 0.48). Thus, it would be worth
to explore further factors which explain the observed levels of IT employees and IT

263
Appendices

60 70
IT employees IT budget
50 60

50
40

40
30
30

20
20

10 10

0 0
1-5 5-10 10-20 20-50 50-100 100-200 200-500 500-1000 2000-5000 >5000 <= 0.10 0.1 - 0.2 0.2 - 0.5 0.5 - 1 m 1 - 2 m 2 - 5 m 5 - 10 m 10 - 20 m 20 - 50 m 50 - 100 > 100 m
m EUR m EUR EUR EUR EUR EUR EUR EUR m EUR EUR

Figure 8:Histogram of IT employees and IT budget

investment. Such questions of inferential statistics shall be subject of future work.

Positioning of the CIO Finally, we also surveyed the positioning of the highest Infor-
mation Officer (CIO) within the organization. Surprisingly, we find that more than 50%
of the CIOs in our sample report directly to the CEO. This may reflect that at least for
large companies the view has prevailed that IT needs to be governed on board level and
involved in major business decisions. About a third of the CIOs resport to the CFO,
which possibly represents the more conservative companies. The remaining 15% state
to report to other C-level executives such as a Chief Operating Officer (COO), regional
heads or the entire board, see Figure9.

Other
15%

Chief
Chief Executive
Financial Officer (CEO)
Officer (CFO) 54%
31%

Figure 9:Position of the CIO (reports to...)

264
 7 SaaS Adoption Processes (Enlarged)

7 SaaS Adoption Processes (Enlarged)

(turn page)

265
Appendices

B External Business IT
Context and
Pre-Decision Phase
IT organization
Firm History and Corporate
and governance
context Governance
context

Need for CRM

(Use of Consolidation of
Spreadsheed and IT department
word lists)

(1a)

Decision Initiative
Phase Vendor selection (3a)
(long and short (1b)
list)
Absorptive Capacities
(2a)
+ Disburden IT Decision
+ Internationalization Auth. SF responsible
o Cost equal SaaS? (3b) leaves the
- Security (from IT)
company

(2b)

Contract with SF

(5)
Implementation
Phase
SF Specification (4)

Migration of old
Specificity
data

Customizing
SF introduction
(SF consultant) Absorptive
(7) Capacities
(6a,b) (6d)

Interface to ERP
Integration with
(External Testing (6c) Training
ERP / involvement
Consultancy)

(8)
Specificity Dec. Auth. Scope
Assimilation
Further
Phase adapt ation
(11a,b) (10a,b) System use (9)
(decision and
specification)

Implementation of
Task Second level First level
enhancements
Authority support for SF support
(partly external)

(12)

Future Initiative

Introduction of
further SaaS
solutions

Figure 10:SaaS Adoption Process Case B

266
 7 SaaS Adoption Processes (Enlarged)

C Business IT External

Context and Pre-Decision Phase

IT organization
Firm History and Spreadsheet
and governance
context solutions for CRM
context

Corporate Governance
Strategy Change / New IT Strategy
ERP
Business incl. ERP and
implementation
Reengineering Outsourcing

(1a)

Several intents for

implementing Staffing of a CRM
CRM (1b) Manager

(2)
Decision Phase
(3a)
Vendor evaluation
(3b)
Decision Authority (long listing)
+Usability
+Mobility
+Cost
Transparency SaaS? Initiative
Absorptive Capacities
o Security

Short listing and

(3a)
cost evaluation

Contract with SF

Further staffing
(4) CRM Competence
Implementation Phase Center

SF Pilot
implementation
Spain
Workshops.
(5) specification,
(6)
testing, training
(Consultancy)
Global SF Rollout Integration to ERP
Absorptive
Capacities Developing Key Specificity
Users & Change (7)
Management
Assimilation Phase
Task Responsibility (8)

Scope of First Level Second level Third level

(9)
Use (10) support support for SF support
(Helpdesk) (Comp. Center) (SF)
Increasing Use
(Users)
Dec. Spec.
Enhancement
Auth. Implementation of
(consulting,
(11a,b) (12a,b) enhancements
decision,
(13) (testing, training)
specification)

(14a)

Future Corp.
Further process
Gov.
standardization
(14b) (core template
rollout)

Figure 11:SaaS Adoption Process Case C

267
 8 Information Systems Specificity (Items and Cross-loadings)

8 Information Systems Specificity (Items and Cross-loadings)

Table 5:Items and cross-loadings

Please rate on the following FuncSp HumSp TechSp Scope DecAuth TaskResp
characteristics of the application. . .
The degree to which this application has been .949 .722 .493 .272 .055 .015
customized to the companys processes
The effort for adapting this application to the .945 .660 .517 .316 .025 .004
companys needs
The amount of trainings that were provided .705 .931 .546 .312 .006 -.024
to the users
The need for personnel with specialized skills .643 .918 .344 .322 -.022 .007
to manage this application
The degree to which the application is .473 .386 .945 .522 .302 .151
technically integrated with other applications
The number and types of interfaces (none, .534 .533 .943 .504 .128 .027
one-way, two-way, several, highly-integrated)
The scope of use of this application within .097 .111 .328 .823 .505 .396
your company (single unit to company-wide)
The share of people in the company who use .373 .410 .580 .963 .445 .388
this application
Who decides on . . .
Changes to this application (e.g., approvals of -.024 -.037 .154 .476 .844 .563
a change request or customizations)
IT spending for this application (e.g., for .079 .039 .271 .452 .924 .655
licenses. enhancements. Maintenance)
Architecture issues regarding this application .050 -.028 .165 .391 .834 .664
(e.g., integration with other systems.
infrastructure)
Who is responsible for . . .
Changes to this application (e.g.. doing -.030 -.065 .037 .372 .714 .925
customizations. implementing a change)
First level support for this application (e.g., -.005 .020 .096 .445 .649 .927
answering user requests. incident
management)
Second level support for this application .065 .019 .133 .364 .643 .921
(e.g., ftnding technical errors. problem
management)

269
Bibliography

D. Acemoglu. Introduction to economic growth. Journal of Economic Theory, 147(2): 545

550, 2012. (cited on p.1)
R. Agarwal and V. Sambamurthy. Principles and models for organizing the IT function.
MIS Quarterly Executive, 1(1):116, 2002. (cited on pp.1,13,18,20, and24)
R. S. Aguilar-Savn. Business process modelling: Review and framework. International
Journal of Production Economics, 90(2):129149, 2004. (cited on p.98)
S. Al-khamayseh, E. Lawrence, and A. Zmijewska. Towards understanding success factors
in interactive mobile government. In Proceedings of the Second European Conference
on Mobile Government (Euro mGov) Opportunities for eGovernment: Adapting to
Mobile and Ubiquitous Business, pages 35. Citeseer, 2006. (cited on pp.5,60,61,
and93)
S. Albers and L. Hildebrandt. Methodische Probleme bei der Erfolgsfaktorenforschung-
Messfehler, formative versus reflektive Indikatoren und die Wahl des
Strukturgleichungs-Modells. Zeitschrift fr betriebswirtschaftliche Forschung, 58 (1):2
33, 2006. (cited on p.51)
J. Alford. Why Do Public-Sector Clients Coproduce? Administration & Society, 34(1):
3256, 2002. (cited on pp.76,79,80,84, and90)
S. Ali and P. Green. IT governance mechanisms in public sector organisations: An
Australian context. Journal of Global Information Management, 15(4):4163, 2007.
(cited on pp.62,207, and208)
K. V. Anderson and H. Z. Henriksen. The first leg of e-government research: domains and
application areas 1998-2003. International Journal of Electronic Government Research
(IJEGR), 1(4):2644, 2005. (cited on p.5)
R. N. Anthony and V. Govindarajan. Management control systems. Management control
systems. McGraw-Hill/Irwin, 12th edition edition, 2007. (cited on pp.17and20)
P. M. Aoki, R. J. Honicky, A. Mainwaring, C. Myers, E. Paulos, S. Subramanian, and
A. Woodruff. Common sense: Mobile environmental sensing platforms to support
community action and citizen science. In Adjunct Proceedings Ubicomp 2008, pages
Paper 201+. Human Computer Interaction Institute, 2008. (cited on pp.90,92,
and114)
M. Armbrust, A. Fox, R. Griffith, A. D. Joseph, R. H. Katz, A. Konwinski, G. Lee, D. A.
Patterson, A. Rabkin, I. Stoica, and Others. Above the clouds: A Berkeley view of
cloud computing. EECS Department, University of California, Berkeley, Tech. Rep.
UCB/EECS-2009-28, 2009. (cited on pp.117,119,138,147, and166)
M. Armbrust, A. Fox, R. Griffith, A. D. Joseph, R. Katz, A. Konwinski, G. Lee,
D. Patterson, A. Rabkin, I. Stoica, and M. Zaharia. A view of cloud computing.

271
Bibliography

Communications of the ACM, 53(4):5058, 2010. (cited on pp.7and8)

C. P. Armstrong and V. Sambamurthy. Information technology assimilation in firms: The
influence of senior leadership and IT infrastructures. Information Systems Research,
10(4):304327, 1999. (cited on pp.127,129, and169)
B. A. Aubert, S. Rivard, and M. Patry. A transaction cost model of IT outsourcing.
Information & Management, 41(7):921932, 2004. (cited on pp.181,186,187, and199)
K. Backhaus, B. Erichson, and W. Plinke. Multivariate Analysemethoden: Eine anwen-
dungsorientierte Einfhrung. Springer, 2003. (cited on pp.38and56)
J. Ballantine and S. Stray. Financial appraisal and the IS/IT investment decision making
process. Journal of Information Technology, 13(1):314, 1998. (cited on pp.94,95,
and114)
R. Baskerville, J. Pries-Heje, and J. Venable. Soft design science research: Extending
the boundaries of evaluation in design science research. DESRIST 2007, pages 1838,
2007. (cited on p.96)
F. M. Bass. A New Product Growth for Model Consumer Durables. Management Science,
15(5):215227, 1969. (cited on pp.100,111, and113)
G. Bassellier and I. Benbasat. Business competence of information technology profession-
als: conceptual development and influence on it-business partnerships. MIS Quarterly,
28:673694, 2004. (cited on pp.148,149,150, and161)
G. Bassellier, I. Benbasat, and B. H. Reich. The influence of business managers IT
competence on championing IT. Information Systems Research, 14(4):317, 2003. (cited
on pp.127,129,147,150, and151)
R. Behara, C. D. Huang, and Q. Hu. A System Dynamics Model of Information Security
Investments. Proceedings of the 15th European Conference on Information Systems
(ECIS), 2007. (cited on p.95)
I. Benbasat, D. K. Goldstein, and M. Mead. The Case Research Strategy in Studies
of Information Systems. MIS Quarterly, 11(3):369386, 1987. (cited on pp.34,64,
and213)
A. Benlian and T. Hess. Chancen und Risiken des Einsatzes von SaaS Die Sicht der
Anwender. In A. Benlian, T. Hess, and P. Buxmann, editors, Software-as-a-Service,
pages 173187. Gabler, 2010a. (cited on p.7)
A. Benlian and T. Hess. The Risks of Sourcing Software as a ServiceAn Empirical
Analysis of Adopters and Non-Adopters. In ECIS 2010 Proceedings, 2010b. (cited on
pp.118,119, and138)
A. Benlian and T. Hess. Opportunities and risks of software-as-a-service: Findings from
a survey of IT executives. Decision Support Systems, 52(1):232246, 2011. (cited on
p.8)
A. Benlian, T. Hess, and P. Buxmann. Drivers of SaaS-Adoption An Empirical Study
of Different Application Types. Business & Information Systems Engineering, 1(5):
357369, 2009. (cited on pp.8,43,117,118,119,121,127,128,129,138,146,150,
151,162,164,181,186,187,189,190, and199)
A. Benlian, M. Koufaris, and T. Hess. The Role of SaaS Service Quality for Continued
SaaS Use: Empirical Insights from SaaS Using Firms. ICIS 2010 Proceedings, 2010.

272
 Bibliography

(cited on pp.118and119)
A. L. Bento and R. Bento. Cloud Computing: A New Phase in Information Technology
Management. Journal of Information Technology, XXII(1):3946, 2011. (cited on
pp.8,26,181, and199)
R. Bernnat, O. Acker, N. Bieber, and M. Johnson. Friendly Takeover The Consumer-
ization of Corporate IT. booz&co, 2010.
[Link] (accessed
7.6.2012). (cited on pp.6and26)
C. P. Bezemer and A. Zaidman. Multi-tenant SaaS applications: maintenance dream
or nightmare? In Proceedings of the Joint ERCIM Workshop on Software Evolution
(EVOL) and International Workshop on Principles of Software Evolution (IWPSE),
IWPSE-EVOL 10, pages 8892, New York, NY, USA, 2010. ACM. (cited on pp.187
and189)
S. Biazzo. Process mapping techniques and organisational analysis: Lessons from so-
ciotechnical system theory. Business Process Management Journal, 8(1):4252, 2002.
(cited on p.103)
Bitkom. Internet per Handy erobert den Massenmarkt, 2010.
[Link] (accessed 7.6.2012). (cited
on p.41)
C. Borucki, S. Arat, and I. Kushchu. Mobile government and organizational effectiveness.
In Proceedings of the First European Conference on Mobile Government, Brighton,
UK: Mobile Government Consortium International, pages 5666, 2005. (cited on
pp.60,61,62, and94)
M. C. Boudreau and D. Robey. Organizational transition to enterprise resource planning
systems: theoretical choices for process research. In Proceedings of the 20th inter-
national conference on Information Systems, ICIS 99, pages 291299, Atlanta, GA,
USA, 1999. Association for Information Systems. (cited on pp.169and170)
M.-C. Boudreau, D. Gefen, and D. W. Straub. Validation in Information Systems
Research: A State-of-the-Art Assessment. MIS Quarterly, 25(1):1+, 2001. (cited
on p.30)
F. Bousquet and C. Le Page. Multi-agent simulations and ecosystem management: a
review. Ecological Modelling, 176(3-4):313332, 2004. (cited on pp.39and40)
A. C. Boynton, G. C. Jacobs, and R. W. Zmud. Whose responsibility is IT management.
Sloan Management Review, 33(4):3238, 1992. (cited on pp.15and20)
S. C. Brailsford, P. R. Harper, B. Patel, and M. Pitt. An analysis of the academic
literature on simulation and modelling in health care. Journal of Simulation, 3(3): 130
140, 2009. (cited on p.40)
A. E. Brown and G. G. Grant. Framing the frameworks: A review of IT governance
research. Communications of the AIS, 15:696712, 2005. (cited on pp.2,3,14,15,20,
24,149,198,206, and208)
C. V. Brown. Examining the Emergence of Hybrid IS Governance Solutions: Evidence
From a Single Case Site. Information Systems Research, 8(1):6994, 1997. (cited on
pp.118,138,142,144,150,184,198,206, and208)

273
Bibliography

C. V. Brown. Horizontal Mechanisms under Differing IS Organization Contexts. MIS

Quarterly, 23(3):421454, 1999. (cited on pp.16,17,20,21,62,143,144,166, and167)
C. V. Brown. The IT Organization of the Future. In J. N. Luftman, editor, Competing
in the Information Age: Align in the Sand, 2nd ed. Oxford University Press, 2003.
(cited on pp.13,18,20,21,22, and181)
C. V. Brown and S. L. Magill. Alignment of the IS functions with the enterprise: toward
a model of antecedents. MIS Quarterly, 18(4):371403, 1994. (cited on pp.2,13,14,
15,20,117,118,120,125,127,129,130,138,142,143,148,149,150,151,180,184,
190,198,204, and208)
A. Bryman. Social Research Methods. Oxford University Press, 2004. (cited on p.30)
E. Brynjolfsson. The productivity paradox of information technology. Communications
of the ACM, 36(12):6677, 1993. (cited on p.1)
E. Brynjolfsson and L. M. Hitt. Information Technology and Organizational Design:
Evidence from Micro Data. Technical report, MIT, 1998. (cited on pp.2,15,20,
and204)
E. Brynjolfsson and L. M. Hitt. Beyond Computation: Information Technology, Or-
ganizational Transformation and Business Performance. The Journal of Economic
Perspectives, 14(4), 2000. (cited on p.1)
E. Brynjolfsson, Y. J. Hu, and M. D. Smith. From Niches to Riches: Anatomy of the
Long Tail. Social Science Research Network Working Paper Series, 2006. (cited on
p.161)
J. Burke, D. Estrin, M. Hansen, A. Parker, N. Ramanathan, S. Reddy, and M. B.
Srivastava. Participatory sensing. In World Sensor Web Workshop, pages 15. Citeseer,
2006. (cited on pp.76,80,84,90,91, and92)
T. Burns and G. M. Stalker. The management of innovation. Social science paperbacks.
Tavistock Publications, 1966. (cited on p.1)
A. Burton-Jones and D. W. Straub. Reconceptualizing System Usage: An Approach and
Empirical Test. Information Systems Research, 17(3):228246, 2006. (cited on p.189)
N. G. Carr. IT doesnt matter. IEEE Engineering Management Review, 32(1):2432,
2004. (cited on p.7)
N. G. Carr. The end of corporate computing. MIT Sloan Management Review, 46(3): 67
73, 2005. (cited on pp.139,163, and206)
L. Carter and F. Blanger. The utilization of e-government services: citizen trust,
innovation and acceptance factors. Information Systems Journal, 15(1):525, 2005.
(cited on pp.76,81,87, and88)
L. D. Carter. Political Participation in a Digital Age: An Integrated Perspective on the
Impacts of the Internet on Voter Turnout. PhD thesis, Virginia Polytechnic Institute
and State University, 2006. (cited on pp.80and84)
R. T. Cenfetelli and G. Bassellier. Interpretation of formative measurement in informa-
tion systems research. MIS Quarterly, 33(4):689707, 2009. (cited on p.37)
Y. E. Chan. Why havent we mastered alignment? The importance of the informal
organization structure. MIS Quarterly Executive, 1(2):97112, 2002. (cited on pp.17
and20)

274
 Bibliography

K. Charmaz. Grounded theory. In N. K. Denzin and Y. S. Lincoln, editors, Strategies of

qualitative inquiry, volume 2, page 249. Sage Publications, Inc, 2003. (cited on pp.32
and33)
I. J. Chen and K. Popovich. Understanding customer relationship management (CRM):
People, process and technology. Business Process Management Journal, 9(5):672688,
2003. (cited on p.1)
W. W. Chin. Commentary: Issues and opinion on structural equation modeling. MIS
Quarterly, 1998a. (cited on p.36)
W. W. Chin. The partial least squares approach to structural equation modeling. In G. A.
Marcoulides, editor, Modern methods for business research, pages 295336. Lawrence
Erlbaum, Mahwah, NJ, 1998b. (cited on pp.37,51,52,83,84,154,157,158,191,
192, and193)
F. Chong and G. Carraro. Architecture strategies for catching the long tail. MSDN
Library, Microsoft Corporation, pages 910, 2006. (cited on pp.138and161)
V. Choudhary. Comparison of Software Quality Under Perpetual Licensing and Software
as a Service. Journal of Management Information Systems, 24(2):141165, 2007a.
(cited on pp.119,138,144, and150)
V. Choudhary. Software as a Service: Implications for Investment in Software Devel-
opment. 40th Annual Hawaii International Conference on System Sciences (HICSS),
page 209a, 2007b. (cited on pp.7,8,138, and166)
G. A. Churchill. A Paradigm for Developing Better Measures of Marketing Constructs.
Journal of Marketing Research, 16(1):64+, 1979. (cited on pp.35and190)
G. A. Churchill and D. Iacobucci. Marketing research: methodological foundations. South-
Western Pub, 2002. (cited on pp.35and36)
CIO. State of the CIO 2008 Data Shows CIO Salaries, Influence Rising, 2008.
[Link]
CIO_Salaries_Influence_Rising (accessed 7.6.2012). (cited on p.210)
J. Cohen. Statistical power analysis for the behavioral sciences. Lawrence Erlbaum, 1988.
(cited on p.194)
W. M. Cohen and D. A. Levinthal. Absorptive capacity: a new perspective on learning
and innovation. Administrative science quarterly, 35(1), 1990. (cited on pp.119
and127)
J. R. Commons. Institutional economics. The American economic review, pages 648657,
1931. (cited on pp.141and182)
Computerwoche. Zahl der mobilen Internetnutzer wird sich bis 2012 verdreifachen, 2009.
[Link] (accessed
7.6.2012). (cited on p.41)
E. M. Cronrath and A. Zock. Forecasting the Diffusion of Innovations by Analogies:
Examples of the Mobile Telecommunication Market. In lnternational Conference of the
System Dynamics Society, pages 116, 2007. (cited on p.105)
D. Cuff, M. Hansen, and J. Kang. Urban sensing: out of the woods. Communications
of the ACM, 51(3):2433, 2008. (cited on pp.76,90, and91)
M. Cusumano. Cloud computing and SaaS as new computing platforms. Communications

275
Bibliography

of the ACM, 53(4):2729, 2010. (cited on pp.7,117,119,138,139,146,166, and181)

R. L. Daft. Organization theory and design. South-Western Pub, 2009. (cited on pp.2,
14,16, and20)
T. H. Davenport. Process innovation: reengineering work through information technology.
Harvard Business School Press, Boston, MA, USA, 1993. (cited on pp.1and6)
T. H. Davenport and J. E. Short. Information technology and business process redesign,
pages 127. Routledge, 2003. (cited on pp.2and148)
R. J. David and S.-K. Han. A systematic assessment of the empirical support for
transaction cost economics. Strat. Mgmt. J., 25(1):3958, 2004. (cited on p.183)
F. D. Davis. Perceived Usefulness, Perceived Ease of Use, and User Acceptance of
Information Technology. MIS Quarterly, 13(3):319340, 1989. (cited on pp.43,77,
78,80,81, and84)
F. D. Davis, R. P. Bagozzi, and P. R. Warshaw. User acceptance of computer technology:
a comparison of two theoretical models. Management science, pages 9821003, 1989.
(cited on pp.77,79,86, and106)
S. De Haes and W. Van Grembergen. IT Governance and its Mechanisms. Information
Systems Control Journal, 1:2733, 2004. (cited on p.166)
S. De Haes and W. Van Grembergen. An Exploratory Study into IT Governance
Implementations and its Impact on Business/IT Alignment. Information Systems
Management, 26(2):123137, 2009. (cited on p.140)
P. C. Deans. The Impact of Social Media on C-level Roles. MIS Quarterly Executive, 10
(4):187200, 2011. (cited on pp.180and200)
H. Demirkan, H. Cheng, and S. Bandyopadhyay. Coordination Strategies in an SaaS
Supply Chain. Journal of Management Information Systems, 26(4):119143, 2010.
(cited on p.138)
J. S. Denford and Y. E. Chan. Reconciling IS strategic alignment approaches. In
Proceedings of the 13th Americas Conference on Information Systems, pages 912,
2007. (cited on p.45)
R. I. Desourdis. Emerging Public Safety Wireless Communication Systems. Artech House
Mobile Communications Series. Artech House, 2002. (cited on p.6)
A. Diamantopoulos, M. Sarstedt, C. Fuchs, P. Wilczynski, and S. Kaiser. Guidelines
for choosing between multi-item and single-item scales for construct measurement: a
predictive validity perspective. Journal of the Academy of Marketing Science, 40(3): 1
16, 2012. (cited on pp.36,163, and201)
J. Dibbern, W. W. Chin, and A. Heinzl. The Impact of Human Asset Specificity on
the Sourcing of Application Services. In European Conference on Information Systems
(ECIS), page 114, 2005. (cited on pp.181,187,188,190, and199)
D. V. Dimitrova and Y. C. Chen. Profiling the adopters of e-government information
and services. Social Science Computer Review, 24(2):172188, 2006. (cited on pp.76,
80, and84)
Y. Dittrich, S. Vaucouleur, and S. Giff. ERP Customization as Software Engineering:
Knowledge Sharing and Cooperation. Software, IEEE, 26(6):4147, 2009. (cited on
pp.146,147, and150)

276
 Bibliography

C. Drury. Management and cost accounting. Cengage Learning EMEA, 2007. (cited on
p.107)
J. R. Edwards. Multidimensional Constructs in Organizational Behavior Research: An
Integrative Analytical Framework. Organizational Research Methods, 4(2):144192,
2001. (cited on p.185)
P. Ein-Dor and E. Segev. Organizational Context and MIS Structure: Some Empirical
Evidence. MIS Quarterly, 6(3):5568, 1982. (cited on pp.126,127, and129)
K. M. Eisenhardt. Agency Theory: An Assessment and Review. The Academy of
Management Review, 14(1):5774, 1989a. (cited on p.140)
K. M. Eisenhardt. Building theories from case study research. Academy of management
review, pages 532550, 1989b. (cited on p.34)
T. El-Kiki and E. Lawrence. Emerging Mobile Government Services: Strategies for
Success. BLED 2007 Proceedings, page 5, 2007. (cited on pp.60,61,93, and205)
P. Ernst. Ubiquitous City Dienste in deutschen Kommunen Attraktivitt, Real-
isierungsstand, Treiber, Hindernisse und Ziele aus kommunaler Sicht. Masters thesis,
Wirtschaftswissenschaftliche Fakultt, Humboldt-Universitt zu Berlin, 2010. (cited
on p.50)
M. Essaidi. ODBIS: towards a platform for on-demand business intelligence services. In
Proceedings of the 2010 EDBT/ICDT Workshops, EDBT 10, New York, NY, USA,
2010. ACM. (cited on pp.146,147, and150)
EU. The Role of eGovernment for Europes future [COM(2003)567]. European
Commission, 2003.
[Link]
l24226b_en.htm (accessed 7.6.2012). (cited on p.5)
J. R. Evaristo, K. C. Desouza, and K. Hollister. Centralization momentum: the pendulum
swings back again. Communications of the ACM, 48(2):6671, 2005. (cited on pp.2,
13, and204)
E. F. Fama and M. C. Jensen. Separation of Ownership and Control. Journal of Law
and Economics, 26(2):301325, 1983. (cited on pp.143,145,150,162, and185)
B. Farbey, F. Land, and D. Targett. Moving IS evaluation forward: learning themes and
research issues. The Journal of Strategic Information Systems, 8(2):189207, 1999.
(cited on pp.94,95, and114)
D. F. Feeny and L. P. Willcocks. Core IS capabilities for exploiting information technol-
ogy. Sloan Management Review, 39(3):921, 1998. (cited on pp.18,19, and20)
W. D. Fernndez. 5. The grounded theory method and case study data in IS research:
issues and design. In D. N. Hart, editor, Information systems foundations: constructing and
criticising, pages 44+. ANU E Press, 2005. (cited on pp.33and213)
F. Fiedler and Others. A Contingency Model of Leadership Effectiveness. Advances in
experimental social psychology, 1:149190, 1964. (cited on p.120)
M. Fishbein and I. Ajzen. Belief, attitude, intention and behavior: An introduction to
theory and research. Addison-Wesley Reading, MA, 1975. (cited on pp.43,77,78,
and79)
[Link]. Five Misunderstandings About Case-Study Research. Qualitative Inquiry,

277
Bibliography

12(2):219245, 2006. (cited on p.34)

C. Fornell and D. Larcker. A second generation of multivariate analysis: Classification
of methods and implications for marketing research. Review of marketing, 51:407450,
1987. (cited on p.36)
C. Fornell and D. F. Larcker. Evaluating Structural Equation Models with Unobservable
Variables and Measurement Error. Journal of Marketing Research, 18(1):3950, 1981.
(cited on pp.53,84, and192)
J. W. Forrester. Industrial Dynamics. M.I.T. Press, Cambridge, Massachusetts, 1961.
(cited on p.95)
J. W. Forrester and P. M. Senge. Tests for building confidence in system dynamics models.
System Dynamics Group, Sloan School of Management, Massachusetts Institute of
Technology, 1978. (cited on p.40)
M. Fowler. Patterns of enterprise application architecture. Addison-Wesley Professional,
2003. (cited on pp.142and200)
P. Fraternali, G. Rossi, and F. Snchez-Figueroa. Rich Internet Applications. Internet
Computing, IEEE, 14(3):912, 2010. (cited on pp.7and139)
G. G. Gable. Integrating case study and survey research methods: an example in
information systems. European Journal of Information Systems, 3(2):112126, 1994.
(cited on pp.210and211)
J. M. Gallaugher and Y.-M. Wang. Understanding Network Effects in Software Markets:
Evidence from Web Server Pricing. MIS Quarterly, 26(4):303+, 2002. (cited on pp.146
and150)
J. Gardner. Leading into the Future. Leading the IT Workforce in Higher Education,
page 133, 2008. (cited on p.118)
N. Gartner. Gartner says worldwide saas revenue to grow 18 percent in 2009, 2009.
[Link] (accessed 7.6.2012). (cited on
p.8)
D. Gefen, E. Karahanna, and D. W. Straub. Trust and TAM in Online Shopping: An
Integrated Model. MIS Quarterly, 27(1):5190, 2003. (cited on pp.77,78,79,80,81,
82, and84)
F. Gens. IT Cloud Services Forecast 2008, 2012: A Key Driver of New Growth. IDC
exchange, 2008.
[Link] (accessed 7.6.2012). (cited on p.8)
B. Glaser and A. Strauss. The discovery of grounded theory: Strategies for qualitative
research. Aldine de Gruyter, New York, 1967. (cited on pp.30,32,33,121, and213)
B. G. Glaser. Theoretical sensitivity: Advances in the methodology of grounded theory,
volume 2. Sociology Press, 1978. (cited on p.33)
B. G. Glaser. Basics of grounded theory analysis. Sociology Press Mill Valley, CA, 1992.
(cited on pp.63and122)
R. L. Glass. Enterprise resource planningbreakthrough and/or term problem? SIGMIS
Database, 29(2):1316, 1998. (cited on p.187)
B. Golden. Cloud computing will cause three it revolutions its going to be hard to
justify this man in the middle staffing. [Link], 2010.

278
 Bibliography

[Link]
it_revolutions/ (accessed 7.6.2012). (cited on pp.165and167)
R. M. Grant. Toward a knowledge-based theory of the firm. Strategic management
journal, 17:109122, 1996. (cited on p.141)
S. Gregor. The nature of theory in information systems. MIS Quarterly, 30(3):611642,
2006. (cited on pp.30and200)
C. Gresov. Exploring Fit and Misfit with Multiple Contingencies. Administrative Science
Quarterly, 34(3), 1989. (cited on pp.118,120,123, and133)
A. Grnlund and T. A. Horan. Introducing e-gov: history, definitions, and issues.
Communications of the Association for Information Systems, 15(2004):713729, 2004.
(cited on pp.5and6)
M. Grottke and S. Rssler. On occupancy, coupon collectors and committee problems.
Technical report, Chairs of Statistics, University of Erlangen-Nuremberg, 2003. (cited
on p.102)
B. Gu, L. Xue, and R. Ray. IT Governance and IT Investment Performance: An
Empirical Analysis. ICIS 2008 Proceedings, page 30, 2008. (cited on pp.117,120,
138,143, and166)
O. Gnther, G. Tamm, L. Hansen, and T. Meseg. Application Service Providers: Ange-
bot, Nachfrage und langfristige Perspektiven: Elektronische Marktpltze und Supply
Chain Management. Wirtschaftsinformatik, 43(6):555567, 2001. (cited on p.7)
J. Hair, C. Ringle, and M. Sarstedt. PLS-SEM: Indeed a Silver Bullet. The Journal of
Marketing Theory and Practice, 19(2):139152, 2011. (cited on pp.37,38,154,191,
and192)
H. H. Harman. Modern factor analysis. University of Chicago Press, 1976. (cited on
p.54)
J. M. Hartman, J. W. Forsen, M. S. Wallace, and J. G. Neely. Tutorials in clinical
research: part IV: recognizing and controlling bias. The Laryngoscope, 112(1):2331,
2002. (cited on p.36)
A.F. Hayes. Beyond Baron and Kenny: Statistical Mediation Analysis in the New
Millennium. Communication Monographs, 76(4):408420, 2009. (cited on pp.37,194,
and195)
B. Hayes. Cloud computing. Communications of the ACM, 51(7):911, 2008. (cited on
p.7)
R. Heeks and S. Bailur. Analyzing e-government research: Perspectives, philosophies,
theories, methods, and practice. Government Information Quarterly, 24(2):243265,
2007. (cited on p.5)
G. Heijkoop and S. Cunningham. Using System Dynamics for Modeling Benefit Realiza-
tion in the Adoption of New Business Software. In Proceedings of the 2007 International
Conference of the System Dynamics Society, 2007. (cited on p.95)
J.-F. Hennart. Explaining the Swollen Middle: Why Most Transactions Are a Mix of
Market and Hierarchy. Organization Science, 4(4):529547, 1993. (cited on pp.182
and199)
J. Henseler and W. W. Chin. A Comparison of Approaches for the Analysis of Inter-

279
Bibliography

action Effects Between Latent Variables Using Partial Least Squares Path Modeling.
Structural Equation Modeling: A Multidisciplinary Journal, 17(1):82109, 2010. (cited
on pp.37and191)
J. Henseler, C. M. Ringle, and R. R. Sinkovics. The use of partial least squares pa th
modeling in international marketing. Advances in International Marketing, 20(2009):
277319, 2009. (cited on pp.39and157)
A. R. Hevner, S. T. March, J. Park, and S. Ram. Design science in information systems
research. MIS Quarterly, 28(1):75105, 2004. (cited on pp.30,96, and212)
D. Hoch and M. Payn. Establishing good IT governance in the public sector. McKinsey
on Government, pages 4555, 2008. (cited on p.62)
G. Homburg and A. Giering. Konzeptualisierung und Operationalisierung Komplexer
Konstrukte - Ein Leitfaden fr die Marketing Forschung. Marketing ZfP, 18(1):524,
1996. (cited on pp.52,53, and55)
P. Hoyningen-Huene. Paul Feyerabend und Thomas Kuhn. Journal for General Philos-
ophy of Science, 33(1):6183, 2002. (cited on p.30)
R. Huang, R. Zmud, and R. Price. IT Governance Practices in Small and Medium-
Sized Enterprises: Recommendations from an Empirical Study. In G. Dhillon, B. C.
Stahl, and R. Baskerville, editors, Information Systems Creativity and Innovation in
Small and Medium-Sized Enterprises, volume 301, chapter 12, pages 158179. Springer,
Berlin, Heidelberg, 2009. (cited on pp.20,21, and24)
J. Hughes and S. Jones. Reflections on the use of grounded theory in interpretive
information systems research. Electronic Journal of Information Systems Evaluation,
6(1), 2003. (cited on pp.33and213)
P. Hui, K. Xu, V. O. K. Li, J. Crowcroft, V. Latora, and P. Lio. Selfishness, Altruism
and Message Spreading in Mobile Social Networks. In INFOCOM Workshops 2009,
IEEE, pages 16. IEEE, 2009. (cited on p.80)
B. Hull, V. Bychkovsky, Y. Zhang, K. Chen, M. Goraczko, A. Miu, E. Shih, H. Balakr-
ishnan, and S. Madden. CarTel: a distributed mobile sensor computing system. In
Proceedings of the 4th international conference on Embedded networked sensor systems,
SenSys 06, pages 125138, New York, NY, USA, 2006. ACM. (cited on pp.90,92,
and114)
S. D. Hunt, R. D. Sparkman, and J. B. Wilcox. The Pre-Test in Survey Research: Issues
and Preliminary Findings. Journal of Marketing Research, 19:26973, 1982. (cited on
pp.36and49)
M. G. Hunter. The Duality of Information Technology Roles. International Journal of
Strategic Information Technology and Applications (IJSITA), 2(1):3748, 2011. (cited
on p.23)
IfM. Kmu-definition des ifm bonn. Institut fr Mittelstandsforschung, 2002.
[Link] (accessed 7.6.2012). (cited on p.257)
M. Indulska, D. S. Hovorka, and J. Recker. Quantitative approaches to content analysis:
identifying conceptual drift across publication outlets. European Journal of Informa-
tion Systems, 21(1):4969, 2012. (cited on p.32)
Z. Irani. Information systems evaluation: navigating through the problem domain.

280
 Bibliography

Information & Management, 40(1), 2002. (cited on pp.94and114)

Z. Irani, P. E. D. Love, and Others. Developing a frame of reference for ex-ante IT/IS
investment evaluation. European Journal of Information Systems, 11(1), 2002. (cited
on pp.94and114)
Z. Irani, P. E. D. Love, and S. Jones. Learning lessons from evaluating eGovernment:
reflective case experiences that support transformational government. The Journal of
Strategic Information Systems, 17(2):155164, 2008. (cited on pp.1,61, and205)
R. A. Irvin and J. Stansbury. Citizen Participation in Decision Making: Is It Worth the
Effort? Public Administration Review, 64(1):5565, 2004. (cited on p.76)
ITU. Measuring the Information Society 2010. International Telecommunication Union
(ITU), Place des Nations, Geneva Switzerland, 2010. (cited on pp.5,60,105,108,
and111)
B. Iyer and J. C. Henderson. Business Value from Clouds: Learning from Users. MIS
Quarterly Executive, 11(1):5160, 2012. (cited on p.27)
M. Janssen and A. Joha. Challenges for Adopting Cloud-based Software as a Service
(SaaS) in the Public Sector. In ECIS 2011 Proceedings, pages Paper 80+, 2011. (cited
on pp.8and210)
M. Janssen and W. S. Shu. Transformational government: basics and key issues. In
Proceedings of the 2nd International Conference on Theory and Practice of Electronic
Governance, ICEGOV 08, pages 117122, New York, NY, USA, 2008. ACM. (cited
on pp.60,61, and205)
J. S. Jasperson, P. E. Carter, and R. W. Zmud. A comprehensive conceptualization of post-
adoptive behaviors associated with information technology enabled work systems. MIS
Quarterly, 29(3):525557, 2005. (cited on p.180)
K. Jedidi, H. S. Jagpal, and W. S. DeSarbo. Finite-mixture structural equation models
for response-based segmentation and unobserved heterogeneity. Marketing Science,
pages 3959, 1997. (cited on p.39)
M. Jensen and W. Meckling. Specific and general knowledge and organizational structure.
In L. Werin and H. Wijkander, editors, Contract Economics, pages 251274. Blackwell,
Oxford, 1998. (cited on pp.141and142)
M. C. Jensen. Organization Theory and Methodology. Social Science Research Network
Working Paper Series, 1998. (cited on pp.145,150, and162)
P. Johnson, A. Kapadia, D. Kotz, N. Triandopoulos, and N. H. Hanover. People-centric
urban sensing: Security challenges for the new paradigm. Technical report, Dartmouth
College, Computer Science, Hanover, NH, 2007. (cited on pp.76and81)
R. Johnson and D. Wichern. Applied Multivariate Statistical Analysis. Pearson Prentice
Hall, 2007. (cited on pp.38and39)
K. G. Jreskog and H. Wold. The ML and PLS techniques for modeling with latent
variables: historical and comparative aspects. In K. G. Jreskog and H. Wold, editors,
Systems Under Indirect Observation, Part 1, pages 263270. Amsterdam: North-
Holland, North-Holland, Amsterdam, 1982. (cited on p.37)
M. M. Kamal. IT innovation adoption in the government sector: identifying the critical
success factors. Journal of Enterprise Information Management, 19(2):192222, 2006.

281
Bibliography

(cited on pp.90and94)
B. Kaplan and D. Duchon. Combining Qualitative and Quantitative Methods in Infor-
mation Systems Research: A Case Study. MIS Quarterly, 12(4):571586, 1988. (cited
on pp.210,211, and212)
P. Kappelhoff. Multidimensionale skalierung. Technical report, Lehrstuhl fr empirische
Wirtschafts- und Sozialforschung, Schumpeter School of Business and Economics, Ber-
gische Universitt Wuppertal, 2001. (cited on pp.38,195, and196)
F. Karimi-Alaghehband, S. Rivard, S. Wu, and S. Goyette. An assessment of the use
of Transaction Cost Theory in information technology outsourcing. The Journal of
Strategic Information Systems, 20(2):125138, 2011. (cited on pp.181,183,186,187,
189, and199)
P. G. W. Keen. Information systems and organizational change. Communications of the
ACM, 24(1):2433, 1981. (cited on p.1)
M. Keil, B. C. Y. Tan, K. K. Wei, T. Saarinen, V. Tuunainen, and A. Wassenaar. A Cross-
Cultural Study on Escalation of Commitment Behavior in Software Projects. MIS
Quarterly, 24(2), 2000. (cited on pp.39and157)
A. Khajeh-Hosseini, I. Sommerville, and I. Sriram. Research Challenges for Enterprise
Cloud Computing.
[Link] (accessed 7.6.2012),
2010. (cited on pp.8,181, and199)
V. Khatri and C. V. Brown. Designing data governance. Communications of the ACM,
53(1):148152, 2010. (cited on pp.27,138,166,184, and208)
J. Kim. Causation. In R. Audi, editor, The Cambridge Dictionary of Philosophy (2nd
edition), pages 125127. Cambridge University Press, Cambridge, UK, 1999. (cited on
pp.169and170)
J. L. King and K. Lyytinen. Information Systems: The State of the Field (John Wiley
Series in Information Systems). John Wiley & Sons, 2006. (cited on pp.8,180,
and200)
S. Kinkade and K. Verclas. Wireless Technology for Social Change: Trends in Mobile Use
by NGOs. Technical report, UN Foundation: Vodafone Partnership, Washington, DC
and Berkshire, UK, 2008. (cited on p.76)
N. Kock. Using WarpPLS in e-collaboration studies: An overview of five main analysis
steps. International Journal of e-Collaboration (IJeC), 6(4):111, 2010. (cited on
p.36)
P. Krmmergaard and J. Rose. Managerial Competences for ERP Journeys. Information
Systems Frontiers, 4:199211, 2002. (cited on pp.180and200)
H. Krasnova and N. F. Veltri. Privacy Calculus on Social Networking Sites: Explorative
Evidence from Germany and USA. In System Sciences (HICSS), 2010 43rd Hawaii
International Conference on, pages 110. IEEE, 2010. (cited on pp.81and88)
P. Kreutter and G. Stadtmann. The Captives End: Lebenszyklusmuster in der Entwick-
lung der deutschen IT-Outsourcing-Industrie. Research notes working paper series //
Deutsche Bank Research, 2009.
[Link] (accessed 7.6.2012). (cited on pp.19and23)

282
 Bibliography

H. Kromrey. Empirische Sozialforschung. Utb, 2006. (cited on p.49)

J. Kruskal and M. Wish. Multidimensional Scaling. Quantitative Applications in the
Social Sciences. Sage Publications, 1978. (cited on pp.38and195)
M. Kumar and O. P. Sinha. M-governmentmobile technology for e-government. In
International conference on e-government, India, 2007. (cited on pp.43,60,61,90,
and93)
N. Kumar and R. Vragov. Active citizen participation using ICT tools. Communications
of the ACM, 52(1):118121, 2009. (cited on pp.76and88)
V. Kumar, S. Sunder, and B. Ramaseshan. Analyzing the Diffusion of Global Cus-
tomer Relationship Management: A Cross-Regional Modeling Framework. Journal of
International Marketing, 19(1):2339, 2011. (cited on p.1)
I. Kushchu. Mobile Government: An Emerging Direction in E-government. IGI Publish-
ing, Hershey, PA, USA, 2007. (cited on pp.44,45, and93)
I. Kushchu and H. Kuscu. From E-government to M-government: Facing the Inevitable.
In the 3rd European Conference on e-Government, pages 253260. Citeseer, 2003.
(cited on pp.5,6,42,60,61,78, and93)
I. Kushchu, H. Kuscu, and B. Yu. Introducing Mobile Government. Mobile government: an
emerging direction in e-government, page 1, 2007. (cited on pp.60and61)
S. Kuznetsov and E. Paulos. Participatory sensing in public spaces: activating urban
surfaces with sensor probes. In Proceedings of the 8th ACM Conference on Designing
Interactive Systems, DIS 10, pages 2130, New York, NY, USA, 2010. ACM. (cited
on p.76)
M. C. Lacity, S. A. Khan, and L. P. Willcocks. A review of the IT outsourcing literature:
Insights for practice. The Journal of Strategic Information Systems, 18(3):130146,
2009. (cited on pp.18,20,182, and185)
M. C. Lacity, L. P. Willcocks, and S. Khan. Beyond Transaction Cost Economics:
Towards an endogenous theory of Information Technology Outsourcing. The Journal
of Strategic Information Systems, 20(2):139157, 2011. (cited on pp.181,187, and199)
J. R. Landis and G. G. Koch. The Measurement of Observer Agreement for Categorical
Data. Biometrics, 33(1):159174, 1977. (cited on p.123)
C. Landmark. CRM Software-as-a-Service Industry, 2011.
[Link] (accessed 7.6.2012). (cited on
pp.117and121)
N. D. Lane, S. B. Eisenman, M. Musolesi, E. Miluzzo, and A. T. Campbell. Urban
sensing systems: opportunistic or participatory? In Proceedings of the 9th workshop
on Mobile computing systems and applications, HotMobile 08, pages 1116, New York,
NY, USA, 2008. ACM. (cited on p.92)
J. N. Lee, M. Q. Huynh, R. C. W. Kwok, and S. M. Pi. IT outsourcing evolution:
past, present, and future. Communications of the ACM, 46(5):8489, 2003. (cited on
pp.7,117,119, and124)
J. N. Lee, S. M. Miranda, and Y. M. Kim. IT Outsourcing Strategies: Universalis-
tic, Contingency, and Configurational Explanations of Success. Information Systems
Research, 15(2):110131, 2004. (cited on pp.18and20)

283
Bibliography

H. Legewie and B. Schervier-Legewie. Im gesprch: Anselm strauss. Journal fr Psy-

chologie, 3(1):6475, 1995. (cited on p.32)
P. Legris, J. Ingham, and P. Collerette. Why do people use information technology? A
critical review of the technology acceptance model. Information & Management, 40
(3):191204, 2003. (cited on pp.78and86)
S. Lehmann, T. Draisbach, C. Koll, P. Buxmann, and H. Diefenbach. Preisgestaltung
fr Software-as-a-Service. Multikonferenz Wirtschaftsinformatik 2010: Gttingen, 23.-
25. Februar 2010 Kurzfassungen der Beitrge, pages 505516, 2010. (cited on pp.7
and119)
J. M. Leimeister, M. Bhm, and P. Yetton. Managing IT in a Business Unit Divestiture.
MIS Quarterly Executive, 11:3748, 2012. (cited on p.22)
S. Leimeister, C. Riedl, M. Bhm, and H. Krcmar. The business perspective of cloud
computing: Actors, roles, and value networks. In Proceedings of 18th European Con-
ference on Information Systems ECIS 2010, 2010. (cited on p.27)
A. Lenk, M. Klems, J. Nimis, S. Tai, and T. Sandholm. Whats inside the Cloud? An
architectural map of the Cloud landscape. In Proceedings of the 2009 ICSE Workshop
on Software Engineering Challenges of Cloud Computing, volume 0 of CLOUD 09,
pages 2331, Washington, DC, USA, 2009. IEEE Computer Society. (cited on p.8)
H. Liang, N. Saraf, Q. Hu, and Y. Xue. Assimilation of enterprise systems: the effect of
institutional pressures and the mediating role of top management. MIS Quarterly, 31
(1):5987, 2007. (cited on pp.85,151,154,157, and193)
R. Likert. A technique for the measurement of attitudes. Archives of Psychology, 140
(140):155, 1932. (cited on p.35)
C. Loebbecke and C. Huyskens. What drives netsourcing decisions? An empirical anal-
ysis. European Journal of Information Systems, 15(4):415423, 2006. (cited on pp.7,
181,186,187, and188)
J. N. Luftman. Managing IT Human Resources: Considerations for Organizations and
Personnel. IGI Global, 2011. (cited on pp.19and20)
J. M. Lyneis, K. G. Cooper, and S. A. Els. Strategic management of complex projects:
a case study using system dynamics. System Dynamics Review, 17(3):237260, 2001.
(cited on p.95)
K. Lyytinen and M. Newman. Explaining information systems change: a punctuated
socio-technical change model. European Journal of Information Systems, 17(6):589
613, 2008. (cited on pp.1,168, and170)
P. Maartensson and A. S. Lee. Dialogical action research at omega corporation. MIS
Quarterly, pages 507536, 2004. (cited on p.34)
L. Maciaszek. Requirements analysis and system design. Pearson Education, 2007. (cited
on p.109)
J. Macionis and L. Gerber. Sociology. Pearson Canada, 2011. (cited on p.30)
S. B. MacKenzie, P. M. Podsakoff, and N. P. Podsakoff. Construct measurement and
validation procedures in MIS and behavioral research: integrating new and existing
techniques. MIS Quarterly, 35(2):293334, 2011. (cited on pp.37and192)
V. Mahajan, E. Muller, and F. M. Bass. Diffusion of new products: Empirical general-

284
 Bibliography

izations and managerial uses. Marketing Science, 14(3 supplement):G79G88, 1995.

(cited on pp.105and108)
R. P. Mahowald, A. Konary, and C. G. Sullivan. Market Analysis Perspective: Worldwide
SaaS & Cloud Services, 2011: New Models for Delivering Software. IDC, 2011.
[Link] (accessed 7.6.2012). (cited
on p.7)
N. Maisonneuve, M. Stevens, and B. Ochab. Participatory noise pollution monitoring
using mobile phones. Information Polity, 15(1):5171, 2010. (cited on pp.76,90,92,
and114)
S. T. March and G. F. Smith. Design and natural science research on information
technology. Decision Support Systems, 15(4):251266, 1995. (cited on pp.30and96)
G. A. Marcoulides. A Critical Look at Partial Least Squares Modeling. MIS Quarterly,
33(1):171175, 2009. (cited on pp.37and51)
G. A. Marcoulides and C. Saunders. Editors comments: PLS: a silver bullet? MIS
Quarterly, 30(2):iiiix, 2006. (cited on p.37)
[Link] and D. P. Rau. Splitting demand from supply in IT. The McKinsey Quarterly,
(Fall):2229, 2006. (cited on pp.15and20)
E. A. Marks. Service-oriented Architecture, Soa, Governance for the Services Driven
Enterprise. Wiley, 2008. (cited on p.136)
M. L. Markus and C. Tanis. The enterprise systems experience-from adoption to success.
Framing the domains of IT research: Glimpsing the future through the past, 173:207
173, 2000. (cited on pp.168and169)
M. Marrone and L. M. Kolbe. Impact of IT Service Management Frameworks on the IT
Organization. Business & Information Systems Engineering, 3(1):518, 2011. (cited
on pp.14and26)
R. Matavire and I. Brown. Investigating the use of "Grounded Theory" in information
systems research. In Proceedings of the 2008 annual research conference of the South
African Institute of Computer Scientists and Information Technologists on IT research
in developing countries: riding the wave of technology, SAICSIT 08, pages 139147,
New York, NY, USA, 2008. ACM. (cited on pp.33,121,122, and213)
R. McAdam and A. Galloway. Enterprise resource planning and organisational inno-
vation: a management perspective. Industrial Management & Data Systems, 105(3):
280290, 2005. (cited on pp.13and21)
M. P. McDonald. The enterprise capability organization: a future for IT. MIS Quarterly
Executive, 6(3):179192, 2007. (cited on p.26)
N. Melo and M. Pidd. A conceptual framework for understanding business processes
and business process modelling. Information Systems Journal, 10(2):105129, 2000.
(cited on p.98)
S. B. Merriam. Qualitative Research and Case Study Applications in Education. Jossey-
Bass Publishers, San Francisco, CA, 1998. (cited on pp.34,64, and75)
B. Miller, E. Kinikin, and L. Herbert. SAP Customer Competence Center. Forrester
Best Practices, 2004.
[Link]

285
Bibliography

[Link] (accessed 7.6.2012). (cited on p.180)

J. Mingers. Combining IS research methods: towards a pluralist methodology. Informa-
tion systems research, 12(3):240259, 2001. (cited on pp.210and211)
J. Mingers. The paucity of multimethod research: a review of the information systems
literature. Information Systems Journal, 13(3):233249, 2003. (cited on pp.30,34,
40,210, and211)
L. Mohr. Explaining organizational behavior. Jossey-Bass social and behavioral science
series. Jossey-Bass, 1982. (cited on p.168)
M. J. Moon. The Evolution of E-Government among Municipalities: Rhetoric or Reality?
Public Administration Review, 62(4):424433, 2002. (cited on p.76)
M. J. Moon and D. F. Norris. Does managerial orientation matter? The adoption of
reinventing government and e-government at the municipal level. Information Systems
Journal, 15(1):4360, 2005. (cited on pp.44and54)
M. Moore and S. Khagram. On Creating Public Value: What Business Might Learn
from Government about Strategic Management, 2004. (cited on p.209)
M. H. Moore. Creating public value: Strategic management in government. Harvard
University Press, 1995. (cited on pp.90and209)
D. L. Morgan and Others. Qualitative content analysis: A guide to paths not taken.
Qualitative health research, 3(1):112, 1993. (cited on p.32)
V. Morwitz. Why consumers dont always accurately predict their own future behavior.
Marketing Letters, 8(1):5770, 1997. (cited on pp.77and106)
M. Mun, S. Reddy, K. Shilton, N. Yau, J. Burke, D. Estrin, M. Hansen, E. Howard,
R. West, and P. Boda. PEIR, the personal environmental impact report, as a platform
for participatory sensing systems research. In Proceedings of the 7th international
conference on Mobile systems, applications, and services, MobiSys 09, pages 5568,
New York, NY, USA, 2009. ACM. (cited on pp.76,90,92, and114)
M. D. Myers. Qualitative Research in Information Systems. MIS Quarterly, 21(2):
241242, 1997. (cited on p.31)
M. D. Myers and M. Newman. The qualitative interview in IS research: Examining the
craft. Information and Organization, 17(1):226, 2007. (cited on p.31)
P. Nagpal. Use of Transaction Cost Economics Framework to Study IT Sourcing: Over-
Application or Under-Theorizing. Sprouts: Working Papers on Information Environ-
ments, Systems and Organizations, 4(2):98110, 2004. (cited on pp.141,147,150,
and164)
F. F. H. Nah, J. L. S. Lau, and J. Kuang. Critical factors for successful implementation of
enterprise systems. Business Process Management Journal, 7(3):285296, 2001. (cited
on pp.128and129)
A. E. Napoleon and S. H. Bhuiyan. Contemporary Research on Mobile Government. In
Scandinavian Workshop on e-Government (SVEG), 2010. (cited on pp.61and93)
K. A. Neuendorf. The content analysis guidebook. Sage Publications, Inc, 2002. (cited on
pp.32,46,47, and63)
M. Newman and D. Robey. A Social Process Model of User-Analyst Relationships. MIS
Quarterly, 16(2), 1992. (cited on pp.30,31,165,168,170, and214)

286
 Bibliography

J. Nielsen. Usability engineering. Morgan Kaufmann, 1994. (cited on p.79)

C. Nitzl. Eine anwenderorientierte Einfhrung in die Partial Least Square (PLS)-
Methode. Technical report, Institut fr Industrielles Management, Universitt Ham-
burg, 2010. (cited on p.51)
NRW. Gemeindefinanzbericht 2009 des stdtetages nordrhein-westfalen. Eildienst Infor-
mationen fr Rat und Verwaltung, 11:315328, 2009. (cited on p.41)
J. A. OBrien. Introduction to information systems. McGraw-Hill/Irwin, 2005. (cited on
pp.142,180, and187)
W. J. Orlikowski. CASE Tools as Organizational Change: Investigating Incremental and
Radical Changes in Systems Development. MIS Quarterly, 17(3):309340, 1993. (cited
on pp.30and33)
W. J. Orlikowski and J. J. Baroudi. Studying Information Technology in Organizations:
Research Approaches and Assumptions. Information Systems Research, 2(1):128,
1991. (cited on p.210)
W. J. Orlikowski and C. S. Iacono. Research Commentary: Desperately Seeking the "IT"
in IT ResearchA Call to Theorizing the IT Artifact. Information Systems Research,
12(2):121134, 2001. (cited on pp.3,8,142,162,180,181, and199)
G. Panten and S. Boow-Thies. Analyse kausaler Wirkungszusammenhnge mit Hilfe
von Partial Least Squares (PLS). Methodik der empirischen Forschung, pages 311326,
2007. (cited on pp.51,52,53, and54)
C. Pardo, F. J. Pino, F. Garca, M. Piattini Velthius, and M. T. Baldassarre. Trends in
Harmonization of Multiple Reference Models Evaluation of Novel Approaches to
Software Engineering. In L. A. Maciaszek and P. Loucopoulos, editors, Evaluation of
Novel Approaches to Software Engineering, volume 230 of Communications in Com-
puter and Information Science, chapter 5, pages 6173. Springer Berlin Heidelberg,
Berlin, Heidelberg, 2011. (cited on pp.14and26)
D. F. Parkhill. The challenge of the computer utility. Number S. 246 in The Challenge
of the Computer Utility. Addison-Wesley Pub. Co., 1966. (cited on p.7)
H. Patel and D. Jacobson. Factors Influencing Citizen Adoption of E-Government: A
Review and Critical Assessment. 16th European Conference on Information Systems,
2008. (cited on pp.44and45)
M. Q. Patton and Others. Qualitative evaluation and research methods. Sage publications
Newbury Park, CA, 1990. (cited on p.34)
D. Peak and M. H. Azadmanesh. Centralization/decentralization cycles in computing:
Market evidence. Information & Management, 31(6):303317, 1997. (cited on pp.2,
13,21,145,150, and204)
J. Peppard and J. Ward. Beyond strategic information systems: towards an IS capability.
The Journal of Strategic Information Systems, 13(2):167194, 2004. (cited on pp.19
and20)
R. R. Peterson, R. OCallaghan, and P. M. A. Ribbers. Information technology gover-
nance by design: investigating hybrid configurations and integration mechanisms. In
International Conference on Information Systems (ICIS), pages 435452, 2000. (cited
on pp.16,17,20,23,138,140,166,167, and184)

287
Bibliography

P. M. Podsakoff and D. W. Organ. Self-Reports in Organizational Research: Problems

and Prospects. Journal of Management, 12(4):531544, 1986. (cited on pp.36,54,85,
154, and193)
S. Posthumus and R. Solms. Agency Theory: Can it be Used to Strengthen IT Gover-
nance? In S. Jajodia, P. Samarati, and S. Cimato, editors, Proceedings of The Ifip
Tc 11 23rd International Information Security Conference, volume 278 of IFIP Inter-
national Federation for Information Processing, chapter 46, pages 687691. Springer
Boston, Boston, MA, 2008. (cited on p.140)
B. Power. Look to IT for Process Innovation? HBR Blog Network, 2012.
[Link] (ac-
cessed 7.6.2012). (cited on p.2)
N. Prakash, M. P. Jaiswal, and U. Gulla. Measurement of Public Value of Enterprise
Applications in Government and Public Sector. Methodology, 2005:03, 2001. (cited on
pp.94and114)
G. Punj and D. W. Stewart. Cluster Analysis in Marketing Research: Review and
Suggestions for Application. Journal of Marketing Research, 20(2):134148, 1983.
(cited on pp.38and156)
W. P. Putsis. Temporal aggregation in diffusion models of first-time purchase: Does
choice of frequency matter? Technological Forecasting and Social Change, 51(3):265
279, 1996. (cited on pp.105and108)
N. Raden. Shedding Light on Shadow IT: Is Excel Running Your Business? Hired
Brains, Inc, 2005.
[Link] (accessed
7.6.2012). (cited on p.21)
W. Reinartz, M. Haenlein, and J. Henseler. An empirical comparison of the efficacy
of covariance-based and variance-based SEM. International Journal of Research in
Marketing, 26(4):332344, 2009. (cited on pp.37and83)
W. L. Richman, S. Kiesler, S. Weisband, and F. Drasgow. A meta-analytic study
of social desirability distortion in computer-administered questionnaires, traditional
questionnaires, and interviews. Journal of Applied Psychology, 84(5):754, 1999. (cited
on p.106)
C. M. Ringle, S. Wende, and S. Will. SmartPLS 2.0 (M3) Beta.
[Link] 2005. (cited on pp.50and83)
N. Roberts, P. S. Galluch, M. Dinger, and V. Grover. Absorptive Capacity and Informa-
tion Systems Research: Review, Synthesis, and Directions for Future Research. MIS
Quarterly, 36(2):625648, 2011. (cited on pp.119and127)
D. Robey and M. C. Boudreau. Accounting for the contradictory organizational conse-
quences of information technology: Theoretical directions and methodological impli-
cations. Information Systems Research, 10(2):167, 1999. (cited on p.165)
B. Rocheleau and L. Wu. Public Versus Private Information Systems. The American
Review of Public Administration, 32(4):379397, 2002. (cited on pp.209and215)
R. Roepke, R. Agarwal, and T. W. Ferratt. Aligning the IT Human Resource with
Business Vision: The Leadership Initiative at 3M. MIS Quarterly, 24(2):327+, 2000.

288
 Bibliography

(cited on pp.149,150, and161)

E. M. Rogers. Diffusion of innovation. The Free Press New York, 1962. (cited on pp.3,
43,58,100,107,113, and204)
J. W. Ross and M. R. Vitale. The ERP Revolution: Surviving vs. Thriving. Information
Systems Frontiers, 2(2):233241, 2000. (cited on pp.168and169)
J. W. Ross, P. Weill, and D. Robertson. Enterprise architecture as strategy: Creating a
foundation for business execution. Harvard Business Press, 2006. (cited on p.24)
P. Rossel, M. Finger, and G. Misuraca. Mobile e-Government Options: Between
Technology-driven and User-centric. The electronic Journal of e-Government, 4(2): 79
86, 2006. (cited on pp.5and93)
J. R. Rossiter. The C-OAR-SE procedure for scale development in marketing. Interna-
tional Journal of Research in Marketing, 19(4):305335, 2002. (cited on p.36)
R. Ryan and T. Raducha-Grace. IT Financial Management: The Business of IT. IBM
Press, 2009. (cited on pp.18,20, and22)
R. Sabherwal and Y. E. Chan. Alignment between business and IS strategies: A study
of prospectors, analyzers, and defenders. Information Systems Research, 12(1):1133,
2001. (cited on pp.45,127,129, and130)
V. Sambamurthy and R. W. Zmud. Arrangements for Information Technology Gov-
ernance: A Theory of Multiple Contingencies. MIS Quarterly, 23(2):261290, 1999.
(cited on pp.2,13,14,20,24,62,117,118,120,123,126,127,129,133,134,136,138,
142,144,147,148,150,166,167,180,184,198,206, and208)
V. Sambamurthy and R. W. Zmud. Research Commentary: The Organizing Logic for
an Enterprises IT Activities in the Digital EraA Prognosis of Practice and a Call for
Research. Information Systems Research, 11(2):105114, 2000. (cited on pp.143,149,
and151)
G. A. Sandy and S. McMillan. A Success Factors Model For M-Government. In EURO
mGOV, pages 349358, 2005. (cited on pp.60,61,93, and94)
M. A. Schilling. Toward a General Modular Systems Theory and Its Application to
Interfirm Product Modularity. The Academy of Management Review, 25(2):312334,
2000. (cited on pp.140,142,144, and150)
H. J. Scholl. Is E-Government Research a Flash in the Pan or Here for the Long Shot? In
M. A. Wimmer, H. J. Scholl, r. Grnlund, and K. V. Andersen, editors, Lecture Notes
in Computer Science, volume 4084, chapter 2, pages 1324. Springer Berlin Heidelberg,
Berlin, Heidelberg, 2006. (cited on p.6)
G. Schryen. Preserving Knowledge on IS Business Value. Business & Information
Systems Engineering, 2(4):233244, 2010. (cited on p.94)
V. Schulz, A. Hochstein, F. Uebernickel, and W. Brenner. Definition and Classification of
IT-shared-service-center. In Americas Conference on Information Systems (AMCIS),
2009. (cited on pp.13and22)
J. A. Schumpeter. Theorie der wirtschaftlichen Entwicklung, 2nd Ed. (orig. from 1911).
Duncker & Humblot, Mnchen, Leipzig, 1926. (cited on p.1)
A. Schwarz and R. Hirschheim. An extended platform logic perspective of IT governance:
managing perceptions and activities of IT. The Journal of Strategic Information

289
Bibliography

Systems, 12(2):129166, 2003. (cited on p.2)

M. K. Sein, O. Henfridsson, S. Purao, M. Rossi, and R. Lindgren. Action design research.
MIS Quarterly, 35(1):3756, 2011. (cited on pp.31,34,91,96,97,113,114, and212)
T. Sethibe, J. Campbell, and C. McDonald. IT Governance in Public and Private Sector
Organisations: Examining the Differences and Defining Future Research Directions.
In University of Canberra, 18th Australasian Conference on Information Systems,
Toowoomba, Australia, 2007. (cited on pp.3,26,62,207, and208)
SFIA. Skills framework for the information age. SFIA Foundation, 2012.
[Link] (accessed 7.6.2012). (cited on p.19)
P. Sheeran. Intention-behavior relations: A conceptual and empirical review. European
review of social psychology, 12(1):136, 2002. (cited on pp.106and108)
B. H. Sheppard, J. Hartwick, and P. R. Warshaw. The Theory of Reasoned Action: A Meta-
Analysis of Past Research with Recommendations for Modifications and Future
Research. Journal of Consumer Research, 15(3):325343, 1988. (cited on pp.77and79)
P. E. Shrout and N. Bolger. Mediation in experimental and nonexperimental studies:
new procedures and recommendations. Psychological Methods; Psychological Methods,
7(4):422, 2002. (cited on p.194)
S. K. Sia and C. Soh. An assessment of Package-Organisation Misalignment: Institutional
and Ontological Structures. European Journal of Information Systems, 16(5):568583,
2007. (cited on pp.1,180,187, and207)
K. Siau, E. P. Lim, and Z. Shen. Mobile Commerce: Promises, Challenges and Research
Agenda. Journal of Database Management, 12(3):413, 2001. (cited on pp.9and60)
M. E. Sobel. Asymptotic Confidence Intervals for Indirect Effects in Structural Equation
Models. Sociological Methodology, 13:290312, 1982. (cited on pp.193and194)
C. Soh, S. S. Kien, and J. T. Yap. Enterprise resource planning: cultural fits and misfits: is
ERP a universal solution? Communications of the ACM, 43(4):4751, 2000. (cited on
pp.1,180,187, and207)
K. A. Stave. Using system dynamics to improve public participation in environmental
decisions. System Dynamics Review, 18(2):139167, 2002. (cited on pp.39,95,114,
and115)
J. Sterman. Business Dynamics: Systems Thinking and Modeling for a Complex World.
Irwin/McGraw-Hill, 2000. (cited on p.100)
H. Strig. Kleine Weltgeschichte der Philosophie. Fischer, Frankfurt, 6th edition, 2011.
(cited on p.29)
A. Strauss and J. M. Corbin. Basics of Qualitative Research : Techniques and Procedures
for Developing Grounded Theory. SAGE Publications, 1998. (cited on [Link],32,33,
63, and64)
D. M. Strong and O. Volkoff. Understanding organization-enterprise system fit: a path
to theorizing the information technology artifact. MIS Quarterly, 34(4):731756, 2010.
(cited on pp.1,33,180,181,187,200, and207)
W. Sun, K. Zhang, S.-K. Chen, X. Zhang, and H. Liang. Software as a Service: An
Integration Perspective. In B. Krmer, K.-J. Lin, and P. Narasimhan, editors, Service-
Oriented Computing ICSOC 2007, volume 4749 of Lecture Notes in Computer Sci-

290
 Bibliography

ence, chapter 52, pages 558569. Springer Berlin / Heidelberg, Berlin, Heidelberg,
2007. (cited on pp.7,128,129, and189)
W. Sun, X. Zhang, C. J. Guo, P. Sun, and H. Su. Software as a Service: Configuration
and Customization Perspectives. In 2008 IEEE Congress on Services Part II (services- 2
2008), pages 1825. IEEE, 2008. (cited on pp.125,128,129,139,144,146,148,150,
167, and187)
A. Susarla, A. Barua, and A. B. Whinston. Understanding the Service Component
of Application Service Provision: An Empirical Analysis of Satisfaction with ASP
Services. MIS Quarterly, 27(1), 2003. (cited on p.7)
A. Susarla, A. Barua, and A. Whinston. A Transaction Cost Perspective of the "Software
as a Service" Business Model. Journal of Management Information Systems, 26(2): 205
240, 2009. (cited on pp.7,138,141, and150)
A. Susarla, A. Barua, and A. Whinston. Multitask Agency, Modular Architecture, and
Task Disaggregation in SaaS. Journal of Management Information Systems, 26(4): 87
118, 2010. (cited on pp.138,141,144,149,150, and186)
H. Tavakolian. Linking the information technology structure with organizational compet-
itive strategy: a survey. MIS Quarterly, 13:309317, 1989. (cited on pp.138and149)
M. Tenenhaus, V. E. Vinzi, Y. M. Chatelin, and C. Lauro. PLS path modeling. Com-
putational Statistics & Data Analysis, 48(1):159205, 2005. (cited on p.37)
T. Thiadens. Manage IT!: organizing IT demand and IT supply. Springer, 2005. (cited
on pp.15and20)
A. Tiwana. Governance-Knowledge Fit in Systems Development Projects. Information
Systems Research, 20(2):180197, 2009. (cited on pp.27,138,141,143,148,150,184,
185, and208)
A. Tiwana and B. Konsynski. Complementarities Between Organizational IT Architec-
ture and Governance Structure. Information Systems Research, 21(2):288304, 2010.
(cited on pp.24,140,144,147,150,198, and208)
L. G. Tornatzky and K. J. Klein. Innovation characteristics and innovation adoption-
implementation: A meta-analysis of findings. IEEE Transactions on engineering
management, 29(1):2845, 1982. (cited on p.45)
S. Trimi and H. Sheng. Emerging trends in M-government. Communications of the ACM,
51(5):5358, 2008. (cited on pp.5,6,41,42,60, and61)
J. E. Triplett. The Solow productivity paradox: what do computers do to productivity?
The Canadian Journal of Economics/Revue canadienne dEconomique, 32(2):309334,
1999. (cited on p.1)
TSO. ITIL Lifecycle Suite 2011: ITIL Service Strategy 2011 / ITIL Service Design
2011 / ITIL Service Transition 2011 / ITIL Service Operation 2011 / ITIL Continual
Service Improvement 2011. The Stationery Office Ltd, 2011. (cited on pp.124,125,
and126)
N. Umanath. The concept of contingency beyond It depends: illustrations from IS
research stream. Information & Management, 40(6):551562, 2003. (cited on p.136)
UN. E-Government Survey 2012: E-Government for the People. United Nations, De-
partment of Economic and Social Affairs, 2012. (cited on pp.1and6)

291
Bibliography

J. M. Utterback and W. J. Abernathy. A dynamic model of process and product inno -

vation. Omega, 3(6):639656, 1975. (cited on p.1)
C. Van den Bulte and G. L. Lilien. Bias and systematic change in the parameter estimates
of macro-level diffusion models. Marketing Science, pages 338353, 1997. (cited on
pp.112and115)
R. van der Pols. ASL: a framework for Application management. The Stationery Office,
2004. (cited on pp.185and186)
W. Van Grembergen. Strategies for information technology governance. Igi Global, 2004.
(cited on pp.2,16,20,23, and184)
J. C. van Niekerk and J. D. Roode. Glaserian and Straussian grounded theory: similar
or completely different? In Proceedings of the 2009 Annual Research Conference
of the South African Institute of Computer Scientists and Information Technologists,
SAICSIT 09, pages 96103, New York, NY, USA, 2009. ACM. (cited on pp.63
and121)
D. Veit, N. Parasie, and F. Schopp. Brgernahes regieren. In Multikonferenz Wirt-
schaftsinformatik (MKWI) 2010, pages 13431355, 2010. (cited on pp.76,80, and84)
V. Venkatesh and F. D. Davis. A Theoretical Extension of the Technology Acceptance
Model: Four Longitudinal Field Studies. Management Science, 46(2):186204, 2000.
(cited on pp.77,78, and81)
V. Venkatesh, M. G. Morris, G. B. Davis, and F. D. Davis. User Acceptance of Informa-
tion Technology: Toward a Unified View. MIS Quarterly, 27(3):425478, 2003. (cited
on pp.77,78,79,80,81,84,86,87, and88)
N. Venkatraman. Beyond outsourcing: managing IT resources as a value center. Sloan
Management Review, 38(3):5164, 1997. (cited on pp.17and20)
N. Venkatraman. IT-enabled Business Transformation: From Automation to Business
Scope Redefinition. Operations management: a strategic approach, page 145, 2005.
(cited on pp.1and209)
S. Vilvovsky. Differences between public and private IT outsourcing: common themes in
the literature. In Proceedings of the 2008 international conference on Digital govern-
ment research, dg.o 08, pages 337346. Digital Government Society of North America,
2008. (cited on p.209)
Vitako. Mobile Government Die groe Freiheit. (3), 2011.
[Link] 203-2011 20-
[Link] (accessed 7.6.2012). (cited on pp.5,6, and77)
E. von Hippel. The sources of innovation. Oxford University Press, USA, 1988. (cited on
p.1)
F. von Kutschera. Wissenschaftstheorie: Grundzge der allgemeinen Methodologie der
empirischen Wissenschaften. Number Bd. 2 in Uni-Taschenbcher. W. Fink, 1972.
(cited on p.29)
E. M. Von Simson. The "Centrally Decentralized" IS Organization. Harvard Business
Review, 68(4):158162, 1990. (cited on pp.13,21, and22)
M. Wade and J. Hulland. Review: the resource-based view and information systems
research: review, extension, and suggestions for future research. MIS Quarterly, 28

292
 Bibliography

(1):107142, 2004. (cited on pp.19and20)

S. Wallace. ePMbook - The Electronic Project Management Book, 2007.
[Link] (accessed 7.6.2012). (cited on p.94)
G. Walsham. Interpretive case studies in IS research: nature and method. European
Journal of information systems, 4(2):7481, 1995. (cited on p.34)
Y. Wand and R. Weber. Toward a theory of the deep structure of information systems.
In International Conference on Information Systems, pages 6171, 1990. (cited on
p.200)
Y. S. Wang, H. H. Lin, and P. Luarn. Predicting consumer intention to use mobile
service. Information Systems Journal, 16(2):157179, 2006. (cited on pp.76,81,82,
and84)
J. Ward, E. Daniel, and J. Peppard. Building better business cases for IT investments.
MIS Quarterly Executive, 7(1), 2008. (cited on pp.94and114)
C. A. B. Warren. Qualitative interviewing. Handbook of interview research: Context and
method, pages 83101, 2002. (cited on p.31)
P. Webb, C. Pollard, and G. Ridley. Attempting to Define IT Governance: Wisdom or
Folly? In Proceedings of the 39th Annual Hawaii International Conference on System
Sciences - Volume 08, Washington, DC, USA, 2006. IEEE Computer Society. (cited
on p.120)
V. Weerakkody, Y. K. Dwivedi, G. Dhillon, and M. D. Williams. Realising t-government:
a UK local authority perspective. In Proceedings of the Fifth International Conference on
Electronic Government. Inderscience Publishers, Geneva. Citeseer, 2007. (cited on
pp.60and205)
V. Weerakkody, G. Dhillon, Y. Dwivedi, and W. Currie. Realising Transformational
Stage E-Government: Challenges, Issues and Complexities. American Conference on
Information Systems (AMCIS) 2008, page 181, 2008. (cited on p.61)
P. Weill. Effective IT Governance, 2007. (cited on p.62)
P. Weill and J. Ross. IT Governance: How Top Performers Manage IT Decision Rights for
Superior Results. Harvard Business Press, 1 edition, 2004a. (cited on pp.2,3,13,
14,15,20,21,24,26,62,73,117,120,124,125,126,127,129,135,138,140,142,143,
144,149,150,151,152,166,167,171,180,184,185,190,198,207, and208)
P. Weill and J. W. Ross. IT Governance on One Page. CISR WP, 349, 2004b. (cited on
pp.2,185,205, and208)
M. Weiser. The computer for the 21 century. Scientific American, 256(3):94104, 1991.
(cited on p.91)
M. Wetzels, G. O. Schrder, and C. van Oppen. Using PLS path modeling for assessing
hierarchial construct models: guidelines and impirical illustration. MIS Quarterly, 33
(1):177195, 2009. (cited on p.37)
D. A. Whetten, T. Felin, and B. G. King. The Practice of Theory Borrowing in Orga-
nizational Studies: Current Issues and Future Directions. Journal of Management, 35
(3):537563, 2009. (cited on p.182)
G. P. Whitaker. Coproduction: Citizen Participation in Service Delivery. Public Admin-
istration Review, 40(3):240246, 1980. (cited on pp.76,81, and90)

293
Bibliography

E. A. Whitley and L. P. Willcocks. Achieving step-change in outsourcing maturity:

toward collaborative innovation. MIS Quarterly Executive, 10(3):95109, 2011. (cited
on pp.18and20)
L. Willcocks, D. Feeny, and N. Olson. Implementing Core IS Capabilities: Feeny-
Willcocks IT Governance and Management Framework Revisited. European Man-
agement Journal, 24(1):2837, 2006. (cited on pp.144and150)
L. P. Willcocks and C. Griffiths. The crucial role of middle management in outsourcing.
MIS Quarterly executive, 9(3):177193, 2010. (cited on pp.18and20)
O. E. Williamson. Transaction-Cost Economics: The Governance of Contractual Rela-
tions. Journal of Law and Economics, 22(2), 1979. (cited on pp.181,183,186,188,
and199)
O. E. Williamson. The Economics of Organization: The Transaction Cost Approach.
American Journal of Sociology, 87(3), 1981. (cited on pp.181,183,186,188, and199)
O. E. Williamson. The Economic Institutions of Capitalism: Firms, Markets, Relational
Contracting. The Free Press, Collier Macmillan, New York, 1985. (cited on pp.141,
181,182,183, and199)
O. E. Williamson. Comparative Economic Organization: The Analysis of Discrete Struc-
tural Alternatives. Administrative Science Quarterly, 36(2):269296, 1991. (cited on
pp.181,182, and199)
O. E. Williamson. The mechanisms of governance. Oxford University Press, USA, 1996.
(cited on pp.181,186, and199)
O. E. Williamson. The new institutional economics: taking stock, looking ahead. Journal
of economic literature, 38(3):595613, 2000. (cited on p.140)
T. J. Winkler. Kommunale Mobilmachung Studie Innovationen im M-Government.
Vitako Aktuell, 3:68, 2011. (cited on p.213)
T. J. Winkler. Transforming Municipalities: Drivers, Inhibitors and Strategies in Mobile
Government Diffusion. In Y. K. Dwivedi and M. A. Shareef, editors, Transformational
Government Through E-Government: Socio-Economic, Cultural, And Technological
Issues, chapter 14, pages 293314. Emerald, 2012. (cited on p.60)
T. J. Winkler and A. Benlian. The dual role of is specificity in governing software as a
service. In (to appear) International Conference on Information Systems (ICIS) 2012
Proceedings, 2012. (cited on pp.11and180)
T. J. Winkler and C. V. Brown. Investigating horizontal authority for on-premises
applications and software as a service the end of "IT as usual"? Working Paper,
2013a. (cited on pp.11and138)
T. J. Winkler and C. V. Brown. Organizing and Configuring the IT Function. In
H. Topi and A. Tucker, editors, (to appear) Computer Science Handbook, Third Edition
- Information Systems and Information Technology - Volume 2, chapter 8. Taylor &
Francis, 2013b. (cited on pp.11and13)
T. J. Winkler and P. Ernst. Innovationen im Mobile Government Eine Analyse von
Dienstattraktivitten und Motivationen von deutschen Kommunen. In Internationale
Tagung Wirtschaftsinformatik (WI 2011) Proceedings, 2011. paper 2. (cited on pp.11,
41,60,61,62,64,72,90, and94)

294
 Bibliography

T. J. Winkler and O. Gnther. Explaining the Governance of Software as a Service

Applications: A Process View. In Multikonferenz der Wirtschaftsinformatik (MKWI)
2012 Proceedings, 2012. (cited on pp.11,165,187,189, and190)
T. J. Winkler, M. Teuber, and I. Finck. Breaking the Cycles of Innovation - A Holis-
tic View on Business Transformation. In M. Niederkorn, A. Barth, A. Becker, and
L. Schulze, editors, Managing Challenges in a Globalized World - Festschrift zur Emer-
itierung von Prof. Manfred Perlitz, pages 237248. Mannheim University Press, 2008.
(cited on p.1)
T. J. Winkler, K. Koroleva, and G. Tamm. Wireless City Initiatives in Europe Towards
a Service-Oriented Approach. In Proceedings of the 2009 conference on Techniques
and Applications for Mobile Commerce: Proceedings of TAMoCo 2009, pages 3647,
Amsterdam, The Netherlands, The Netherlands, 2009. IOS Press. (cited on pp.43
and50)
T. J. Winkler, C. Goebel, A. Benlian, F. Bidault, and O. Gnther. The Impact of
Software as a Service on IS Authority A Contingency Perspective. In International
Conference on Information Systems (ICIS) 2011 Proceedings, 2011a. (cited on pp.11,
26,27,117,139,145,149,150,151,156,161,165,167,168,170,174,178,181,184,
185,186, and190)
T. J. Winkler, N. Lvova, and O. Gnther. Towards Transformational IT Go vernance
The Case of Mobile Government Adoption. In European Conference on Information
Systems 2011 (ECIS) Proceedings, 2011b. (cited on pp.11,60,88,90,93,94,113,
and114)
T. J. Winkler, H. Hirsch, G. Trouvilliez, and O. Gnther. Participatory Urban Sens-
ing: Citizens Acceptance of a Mobile Reporting Service. In European Conference on
Information Systems 2012 (ECIS) Proceedings, 2012a. (cited on pp.11,76, and94)
T. J. Winkler, H. Ziekow, and M. Weinberg. Municipal benefits of participatory urban
sensing: A simulation approach and case validation. (to appear) Journal of Theo-
retical and Applied Electronic Commerce Research (JTAER) Special Issue on smart
applications for smart cities, 4, 2012b. (cited on pp.11and90)
T. J. Winkler, C. Goebel, F. Bidault, and O. Gnther. Information Technology and
Business Practices in Germany: Results from the 2011 BIT Survey. In U. Karmarkar
and V. Mangal, editors, (to appear) The UCLA Anderson Business and Information
Technologies (BIT) Project: A Global Study of Business Practice. World Scientific
Publishing Company, 2013. (cited on p.257)
S. J. Winter, K. M. Chudoba, and B. A. Gutek. Misplaced resources? Factors associated
with computer literacy among end-users. Information & Management, 32(1):2942,
1997. (cited on pp.81and84)
S. Wood. Comprehensive laboratory informatics: A multilayer approach. American
Laboratory, 2007.
[Link]
1436-Comprehensive-Laboratory-Informatics-A-Multilayer-Approach/
(accessed 7.6.2012). (cited on p.206)
M. Wright, C. Upritchard, and T. Lewis. A validation of the Bass new product diffusion

295
Bibliography

model in New Zealand. Marketing Bulletin, 8:1529, 1997. (cited on pp.105and108) J.-
H. Wu and S.-C. Wang. What drives mobile commerce? Information & Management,
42(5):719729, 2005. (cited on pp.76,81, and82)
M. Xin and N. Levina. Software-as-a Service Model: Elaborating Client-Side Adoption
Factors. ICIS 2008 Proceedings, page 86, 2008. (cited on pp.8,118,119,128,129,
138,139,144,150, and187)
H. Xu, H.-H. Teo, B. Tan, and R. Agarwal. The Role of Push-Pull Technology in Privacy
Calculus: The Case of Location-Based Services. Journal of Management Information
Systems, 26(3):135174, 2009. (cited on pp.76and81)
L. Xue, G. Ray, and B. Gu. Environmental Uncertainty and IT Infrastructure Gov-
ernance: A Curvilinear Relationship. Information Systems Research, 22(2):389399,
2011. (cited on pp.27,138,184, and208)
Y. Xue, H. Liang, and W. R. Boulton. Information technology governance in information
technology investment decision processes: the impact of investment characteristics,
external environment, and internal context. MIS Quarterly, 32:6796, 2008. (cited on
pp.139,144,145,150,162,207, and208)
R. Yanosky. From Users to Choosers: The Cloud and the Changing Shape of Enterprise
Authority. In R. N. Katz, editor, The Tower and The Cloud, pages 126136. Educause E-
Book, Campanile, University of California, Berkeley, 2008. (cited on pp.118,139,
148,149,150,165, and167)
R. K. Yin. Case study research: design and methods. Sage Publications, 3rd edition, 2003.
(cited on pp.30,34,64,121,212, and213)
S. Y. Yousafzai, G. R. Foxall, and J. G. Pallister. Technology acceptance: a meta-analysis
of the TAM: Part 2. Journal of Modelling in Management, 2(3):281304, 2007. (cited
on pp.78,79, and86)
R. Zarnekow, L. M. Kolbe, and N. H. Schmidt. Predictors of Green IT Adoption:
Implications from an Empirical Investigation. AMCIS 2010 Proceedings, page 367,
2010. (cited on p.129)
P. Zwieg, K. Kaiser, C. Beath, C. Bullen, K. Gallagher, T. Goles, J. Howland, J. C.
Simon, P. Abbott, T. Abraham, and Others. The information technology workforce:
Trends and implications 2005-2008. MIS Quarterly Executive, 5(2):4754, 2006. (cited
on pp.19and20)

296
Selbstndigkeitserklrung

Ich bezeuge durch meine Unterschrift, dass meine Angaben ber die bei der Abfassung
meiner Dissertation benutzten Hilfsmittel, ber die mir zuteil gewordene Hilfe sowie ber
frhere Begutachtungen meiner Dissertation in jeder Hinsicht der Wahrheit entsprechen.

Berlin, den 7.6.2012 Till J. Winkler

297