KNOWLEDGE MANAGEMENT

Chapter V
Knowledge Transfer and Sharing

Topics Covered in This Chapter

5.1 Knowledge Transfer and Sharing — Overview
5.2 Primary Strategies: Codification vs. Personalisation
5.3 Key Differences Between the Two Strategies
5.4 Benefits of Combining Both Strategies
5.5 Transfer Methods and Categories of Lesson Sharing
5.6 IT Tools Supporting the Personalisation Approach
5.7 Role of the Internet in Knowledge Management
5.8 Knowledge Transfer in the e-World
5.9 KM System Tools — Classification and Inventory
5.10 Neural Networks for KM
5.11 Association Rules for KM
5.12 Classification Trees for KM
5.13 Data Mining and Business Intelligence
5.14 Decision-Making Architecture in KM
5.15 Data Management in KM
5.16 KM Protocols
5.17 Managing Knowledge Workers
5.18 The CRISP-DM Process Model
5.19 Real-World Examples: Successful KM Portals

5.1 Knowledge Transfer and Sharing — Overview

Knowledge sharing is the process through which explicit or tacit knowledge is communicated to
other individuals. Knowledge transfer is the strategic process of linking knowledge requirements
with available knowledge sources. At its core, knowledge management aims to establish an
environment where people are encouraged to learn, share, and use knowledge together for the
benefit of the organisation.

Core Definitions
Knowledge The communication of explicit or tacit knowledge from one individual
Sharing or group to another.

Knowledge The strategic process of satisfying knowledge requirements by linking

Transfer them with available knowledge sources. Transfer is deemed
successful when the recipient reaches the same basic level of
understanding as the sender.
Figure 1. Knowledge transfer: a process model

Knowledge transfer (KT) is the systematic process of sharing and moving expertise,
information, and skills from one part of an organization to another to ensure critical "know-how"
is preserved and utilized. Within the broader field of Knowledge Management (KM), it acts as
the bridge between capturing information and applying it to solve problems or innovate.

The Knowledge Transfer Process

While models vary, the process typically follows these five core stages:

 Identification: Determining what critical knowledge exists and who holds it. This
includes identifying "mission-critical" information that would be lost if a key expert left.

 Capture and Documentation: Converting tacit knowledge (personal experience and

intuition) into explicit knowledge (manuals, SOPs, or video tutorials).

 Knowledge Sharing: The active dissemination of information through methods like

mentorship, training sessions, or digital platforms like Microsoft SharePoint or
Confluence.

 Integration and Application: The recipient absorbs the knowledge and begins using it
in their daily role. This stage often requires follow-up or additional coaching to ensure
the knowledge is correctly implemented.

 Evaluation and Improvement: Measuring the effectiveness of the transfer through

feedback, performance metrics, and Key Performance Indicators (KPIs) like reduced
resolution times or error rates.

Key Strategies and Techniques

Different types of knowledge require different transfer methods:

 Formal Methods: Structured approaches such as classroom training, e-learning

courses, and Standard Operating Procedures (SOPs).
  Informal Methods: Organic interactions like job shadowing, mentorship, and
communities of practice where employees discuss shared problems.

 Tacit Knowledge Focus: Techniques like storytelling (sharing experiences of past

successes/failures) and after-action reviews help capture experiential wisdom that isn't
easily

Common Barriers

Several factors can obstruct effective knowledge flow:

 Inadequate Documentation: Critical information remains "trapped" in the minds of

veteran employees.

 Cultural Challenges: A lack of trust or a "knowledge is power" culture where

employees are reluctant to share their expertise.

 Geographic Dispersion: Siloed teams in different locations struggling with

communication and coordination.

Barriers to Effective Sharing

Creating a sharing environment is challenging because organisational cultures frequently reinforce
behaviours that inhibit knowledge flow. Key barriers include:
• 'Knowledge is Power' — The belief that status or job security depends on keeping knowledge
secret.
• Time Constraints — Staff feel too busy with 'real work' to spend time sharing or reflecting.
• Not-Invented-Here (NIH) Syndrome — Cultural tendency to reject ideas originating outside
one's own department or unit.
• Lack of Trust — Individuals are unlikely to share high-value knowledge unless they trust the
recipient and the system.

Critical Success Factors

• Leadership Support — Top management must act as role models and provide a clear vision
for knowledge sharing.
• Incentives and Rewards — Successful firms recognise and reward both the sharing and the
reuse of knowledge.
• Formal Integration — Making knowledge sharing a formal part of job descriptions and
performance appraisals encourages participation.
• User-Centric Technology — Systems must be easy to use and integrated into employees'
actual work processes.

Organisational Examples
• Xerox — 'Eureka' system allowed service technicians to share repair tips globally, saving
millions in labour and parts.
• World Bank — Transformed into a 'Knowledge Bank' using thematic Communities of Practice
(CoPs) to adapt global expertise to local needs.
• Ford — 'Best Practice Replication' system systematically shares and implements proven
manufacturing improvements across global sites.
• Accenture — 'Knowledge Xchange' (KX) hosts over 7,000 databases of employee capital and
client experiences.
 5.2 Primary Strategies: Codification vs. Personalisation
Most organisations employ one of two main strategies — or a carefully managed hybrid of both —
to facilitate the flow of knowledge. The choice depends on the nature of the knowledge being
managed, organisational culture, and strategic objectives.

Codification — The 'Collecting' Dimension

This strategy focuses on linking people with information. It involves extracting explicit knowledge
(knowledge that can be expressed in words and numbers) from individuals and storing it in
electronic repositories such as databases, manuals, and intranets. Known as a 'people-to-
document' strategy, it allows many users to retrieve knowledge without needing direct contact with
the original developer.
• Best suited for: recurring problems where recorded solutions can be easily retrieved and
reused.
• Strategic goal: efficiency and effectiveness in operationally focused value chains.
• Technology role: central — operates like a library with large document caches and search
engines.
Personalisation — The 'Connecting' Dimension
This strategy focuses on linking people with people. It is designed to facilitate the flow of tacit
knowledge — the highly personal, experiential 'know-how' that is difficult to articulate in writing.
Because tacit knowledge is embedded in individuals, it is shared most effectively through direct
human interaction, social networks, and personal contact.
• Best suited for: high-level strategic problems requiring creative, analytically rigorous advice
from individual experts.
• Strategic goal: innovation, advancement, and fresh insight generation.
• Technology role: supportive — IT is used to help individuals locate each other (Expert Yellow
Pages, skills directories, online forums).
5.3 Key Differences Between Codification and Personalisation

Feature Codification Strategy Personalisation Strategy

Core Focus Linking people with information Linking people with people

Model People-to-document Person-to-person

Knowledge Type Explicit (written, structured) Tacit (experiential, intuitive)

Knowledge Output Knowledge stocks for frequent Knowledge flows through

reuse interaction

Role of IT Central — primary repository Supportive — locates and

connects people

Strategic Goal Efficiency and cost reduction Innovation and unique problem-
solving

Business Fit Low-cost, operational firms Differentiation firms (McKinsey,

(Dell, HP, Delta Airlines) BCG)

Advantage Always available; quick Always up-to-date; preserves

distribution to many deep context

Disadvantage Info overload; documents Inefficient; may reinvent the

become obsolete; loss of wheel; expert-dependent
context

KEY The most successful KM programmes achieve an integrated approach

PRINCIPLE balancing both strategies. A database (codification) can identify a best
 practice, while on-the-job personal contact (personalisation) transfers the
deep knowledge of how to implement it.

5.4 Benefits of Combining Both Strategies

1. Increased Efficiency and Reduced Redundancy

• Avoiding 'Reinventing the Wheel' — Codification ensures previously learned lessons and
recurring solutions are stored and readily accessible, preventing duplication of effort inherent
in purely personalised approaches.
• Managing Information Overload — A strategy focused only on codification can lead to
'information overflow.' Personalisation provides the human interaction needed to navigate vast
document collections and find what is truly relevant.

2. Complementary Knowledge Flow

• Databases as Starting Points — Explicit repositories serve as a useful starting point, but real
value is added when personalisation (CoPs, direct contact) reveals the deeper tacit insights
beneath the explicit information.
• Databases Point to People — Most effective when they identify experts who can transfer
context-specific knowledge that cannot be easily documented.

3. Creating Dynamic and Enriched Knowledge

• Interactive Documentation — Linking collected documents directly to their authors and
including interactive possibilities makes knowledge dynamic rather than static.
• Two-Way Enrichment — When codified knowledge is shared through face-to-face dialogue, it
provides a 'two-way benefit' enriching the knowledge of both conveyor and recipient.
• Shortened Learning Curves — Exit interviews (capturing tacit knowledge before departure)
combined with searchable databases allow new staff to get 'up to speed' more quickly.

4. Strategic Flexibility and Speed

• Coping with New Economy Demands — Personalisation-only strategies often struggle to
match modern speed requirements. An integrated approach allows rapid reuse of codified
assets while maintaining the innovation benefits of human networks.
• Predominant vs. Supporting Roles — Research suggests the most effective organisations
pursue one strategy predominantly (~80%), using the second as a crucial supporting
mechanism (~20%).

80/20 RULE The most successful companies do not try to excel equally at both strategies
— they pursue one dominantly (80%) and use the other as a support (20%).
Attempting both equally risks failing at both.

5.5 Transfer Methods and Categories of Lesson Sharing

Five Categories of Lesson Sharing

Research identifies five distinct categories of lesson sharing within large organisations, differing
based on the intended receiver and the nature of the task:
 • Serial Transfer — Sharing lessons across frequent and routine tasks within the same team
performing the same task in a new context.
• Near Transfer — Explicit knowledge shared between teams performing similar tasks in
different parts of the organisation (e.g., Ford's Best Practice Replication).
• Far Transfer — Tacit knowledge shared between teams in very different contexts, often
facilitated through the physical movement of personnel.
• Strategic Transfer — Collective knowledge shared for high-stakes, non-routine strategic
decisions affecting the entire organisation.
• Expert Transfer — Seeking technical or specialist knowledge for non-routine tasks from
specialized sources (e.g., Buckman Lab's TechForums).
In the knowledge management (KM) process, Lesson Sharing—often referred to as
sharing "Lessons Learned"—is categorized by the nature of the knowledge and the timing
or method of the exchange. These categories ensure that insights from past experiences
are effectively reused to prevent recurring errors and replicate successes.
1. Nature of Knowledge Categories
This classification focuses on how the lesson is articulated and shared:
 Explicit Lessons: These are documented insights that are easy to articulate and
store in a database or repository, such as standard operating procedures (SOPs),
technical manuals, or project checklists.

 Tacit Lessons: These are intuitive, experience-based insights that are difficult to
write down. They are typically shared through direct interaction, such as mentorship,
job shadowing, or storytelling.

2. Process-Based Categories (Methodologies)

Lessons are often shared through specific KM tools or "events" designed for different
stages of a project lifecycle:
 Peer Assist (Before): A "learning before" technique where a team about to start a
task meets with people who have already completed similar work to gain their
insights.

 After Action Review (AAR) (During): A "learning during" technique used during or
immediately after a task to discuss what was supposed to happen, what actually
happened, and why.

 Retrospect (After): A "learning after" session at the end of a project. It is a

structured meeting where the team captures collective knowledge before they
disband.

3. Channel & Format Categories

Lessons can also be categorized by the medium used to disseminate them:
  Written/Digital Repositories: Lessons stored in searchable databases, wikis, or
internal portals like a "Lessons Learned Register".

 Social & Narrative: Lessons shared via storytelling, podcasts, or community "lunch
and learns." Stories are often more impactful than text-based imperatives.

 Embedded in Workflow: Lessons that are automatically prompted within a system,

such as a project management tool asking for a lesson before a phase can be
closed.
Specific Transfer Methods and Tools

Best Practice Replication (BPR)

A formal process where proven improvements are identified, codified in structured templates, and
'pushed' to other relevant sites for adoption. Ford's BPR system is a leading example, having
delivered over US$1.25 billion in value since 1995.

After Action Reviews (AARs) and Retrospects

Structured tools used to capture lessons learned both during and immediately after a project.
Follows four key questions: What was supposed to happen? What actually happened? Why did it
happen that way? What can be learned?

Peer Assists
A meeting where a team working on a project seeks insights and knowledge from other teams with
relevant experience before embarking on the activity, preventing duplication and accelerating
learning.

Communities of Practice (CoPs)

Organic, self-organised groups informally bound together by shared expertise who interact regularly
to develop and share knowledge around specific themes. The World Bank and Xerox are prominent
examples.

Knowledge Harvesting
A manual or software-facilitated technique used to capture the knowledge of experts — particularly
tacit know-how — and make it available to others before it is lost through retirement or departure.

Storytelling
Using narratives with characters, plot, and a 'hero' protagonist to capture both knowledge content
and the social context of 'how things are done around here.' Stories are 'sticky' and memorable,
conveying tacit knowledge in its rich, original form that a manual cannot replicate.

Knowledge-Based Systems
Rule-based or Case-Based Reasoning (CBR) systems that preserve and apply human expertise,
allowing users to solve new problems based on solutions to similar past cases.

How Storytelling and AARs Differ

Feature After Action Reviews (AARs) Storytelling

Form Structured discussion / meeting Narrative / story with plot and

 characters

Trigger Completion of a specific action Ongoing; used for change,

or phase culture, or community

Key Output Specific 'lessons learned' and Holistic understanding and

repeatable advice shared values

Language Style Analytical and objective Authentic, personal, emotional

Knowledge Focus Turning tacit insights into explicit Sharing tacit knowledge through
points rich context

Impact Structured blame-free learning Memorable, inspiring, builds

event community trust

Technological Enablers for Transfer

• Communication Technologies — Email and videoconferencing enable real-time and
asynchronous transfer across geographical boundaries.
• Collaboration Tools — Virtual whiteboards, shared document workspaces, and groupware
(e.g., Lotus Notes) facilitate cooperative work.
• Knowledge Portals — Single gateways to integrated and personalised information, using 'Best
Bets' to direct users to the most relevant content.

Selection Criteria
The appropriate form of transfer depends on context. Face-to-face communication is the most
valuable but most time-consuming method — best for complex issues like R&D. Simple issues can
be transferred effectively using ICT tools. Selection requires a suitability analysis of business
requirements, infrastructure, and culture.

5.6 IT Tools Supporting the Personalisation Approach

The personalisation approach uses information technology as a supportive tool to help individuals
locate one another and communicate directly, rather than as a primary repository. The following
categories of IT tools best support this strategy.

1. Expertise Locator Systems and Electronic White Pages

Widely considered the most essential IT tools for personalisation — they help people find 'who
knows what' within an organisation.
• Expertise Locators / Knowledge Yellow Pages — Catalog employee competencies, skills, and
experience. Examples: AskMe's Employee Knowledge Network, Kamoon.
• Electronic White Pages — Include knowledge, skills, interests, and personal photographs,
maintained by individuals themselves.
• Discovery Software — Tools like Tacit or Lotus Discovery Server act as 'introduction brokers,'
identifying users searching for similar intranet topics and suggesting they collaborate.

2. Collaborative Tools and Groupware

Provide the digital workspace for people to interact once they have found each other.
• Groupware — Lotus Notes, Novell GroupWise, Microsoft Exchange — allow electronic
discussions, common document editing, and activity coordination.
• Intranets and Discussion Boards — Simplify interaction across locations through email,
discussion boards, and one-stop-shop portals.
 • Wikis and Blogs — Enterprise 2.0 tools allowing bottom-up, peer-to-peer content creation and
informal networking (e.g., Cisco's internal 'Idea Zone' wiki).

3. Communication Technologies
• Videoconferencing and Teleconferencing — Convey body language and tone of voice, critical
for understanding complex tacit information across distances.
• Instant Messaging and Chat — Tools like Sametime or BlueReach provide real-time
consultation, connecting seekers with available experts instantly.
• Virtual Working Tools — Sophisticated systems allow expertise in one location to be applied in
another in real-time (e.g., remote surgical procedures).

4. Social Networking and Mapping Tools

• Social Networking Software — Internal platforms similar to Facebook (e.g., Cisco's Directory
3.0) strengthen existing networks and facilitate new connections based on expertise.
• Social Network Analysis (SNA) Tools — Map and visualize relationship data from email logs
and interaction patterns to identify actual subject matter experts and community leaders.
• NASA POPS ELS — Uses Semantic Web technologies to display social networks between
users, projects, and skill sets, including identifying intermediaries to help people connect.
In Knowledge Management (KM), IT tools leverage personalization to ensure that the right
information reaches the right person at the right time. This transition from generic document
storage to tailored delivery is primarily driven by Artificial Intelligence (AI) and Machine
Learning (ML).
Key Personalization Tools in KM
 AI-Powered Knowledge Bases: Modern platforms like Guru and Document360 use
AI to deliver verified answers directly into existing workflows (e.g., Slack or Microsoft
Teams), personalizing responses based on the user's current task.

 Personalization Engines & Recommendation Systems: Tools such

as Coveo and Bloomfire analyze individual browsing habits, roles, and past
interactions to suggest relevant articles and experts.

 Enterprise Search Tools: Systems like Microsoft Search tailor search results based
on a user's organizational role and historical activity across platforms like Office 365.

 Adaptive Learning Systems: Platforms like Absorb LMS personalize internal

training by adjusting the pace and content of educational materials to match an
employee's specific learning style and professional needs.

Core Personalization Techniques

 Contextual Semantic Search: AI interprets the intent behind a query rather than
just matching keywords. It considers factors like the user's department, geographic
location, and current project.
  Automated Content Tagging: Machine learning algorithms automatically
categorize and tag information with relevant metadata, ensuring that content is
discoverable by the specific users who need it.

 Predictive Analytics: By identifying patterns in data usage, KM systems can

anticipate future knowledge needs and proactively surface information before a user
explicitly searches for it.

 Role-Based Access & Views: Systems like Salesforce Knowledge ensure

employees only see content relevant to their specific role, reducing information
overload and protecting sensitive data.

Benefits of Personalized KM
 Increased Productivity: Reduces "search time" (which can take employees over 5
hours a week) by surfacing the most pertinent documents instantly.

 Enhanced Decision-Making: Provides real-time, context-aware insights, giving

leaders a "360-degree view" before making critical choices.

 Improved Knowledge Retention: Tailored delivery helps preserve "tacit

knowledge" (experiential expertise) and makes it easier for new employees to find
role-specific mentors.

5.7 Role of the Internet in Knowledge Management

The Internet serves as the foundational backbone of the Knowledge Economy, acting as a vast web
of conversations and exchanges that has enabled KM to achieve possibilities previously considered
unfeasible. Its role is both infrastructural and catalytic.

The Internet as Catalyst and Infrastructure

The availability of the World Wide Web was instrumental in catalysing the KM movement. While
earlier information technologies were better suited to managing structured data, the Web enabled
the handling of unstructured text and graphic information. Without the connectivity, simplicity, and
commonality of an Internet-based interface, organisational ability to leverage knowledge would
remain 'stuck in the Stone Age.' Technically, the Internet permits communication between
heterogeneous networks, providing secure and reliable transport of information packets.

Facilitating Global Collaboration and Communication

• Asynchronous Communication — Email has become the most popular way to share
knowledge globally, allowing exchange without both parties being present simultaneously.
• Real-Time Interaction — Videoconferencing and instant messaging enable collaboration
across geographical boundaries and time zones.
• Intranets — 'Private Internets' provide a 'one-stop knowledge shop,' simplifying interaction
through discussion boards and shared document workspaces.

Enterprise Portals: The 'Face' of KM

The Internet facilitates the creation of enterprise portals — personalised gateways or single points
of interaction for employees, partners, and customers. These portals integrate various sources of
information into one location to avoid information overload, turning information into knowledge by
facilitating navigation, visualisation, and heuristic interaction.

Knowledge Discovery and Web Mining

• Expertise Location — Web content mining techniques automatically 'read' natural language
documents and associate employee names with specific skills.
• Competitive Intelligence — Intelligent agents surf the Web to analyse retrievable documents,
providing competitive advantages by reviewing far more information than a human analyst
could manage.

The Shift to Web 2.0 and Collective Intelligence

The emergence of Web 2.0 represents a revolution where the Internet acts as a platform for
harnessing collective intelligence:
• Wikis and Blogs — Allow many authors to contribute and edit content collaboratively, fostering
cooperative culture and enabling bottom-up authoring.
• Social Networking — Platforms like LinkedIn and internal 'enterprise Facebooks' (e.g., Cisco's
Directory 3.0) strengthen existing networks.
• Crowdsourcing — Organisations outsource tasks to large, undefined communities through
open calls, leveraging value creation beyond traditional boundaries.

Knowledge Mobilisation
The Internet combined with wireless and mobile technologies enables 'knowledge mobilisation' —
allowing 'road warriors' and field workers to access intellectual assets and human expertise
'anytime, anywhere, and on any device.' Mobile accessible information puts knowledge to work
directly at the point of demand.

5.8 Knowledge Transfer in the e-World

In the modern e-world, the emergence of Internet and intranet technology has enabled knowledge
transfer to achieve formidable possibilities previously considered unfeasible. Transfer is deemed
successful when the recipient reaches the same basic level of understanding as the sender.

Two Dimensions of Electronic Transfer

• The 'Collecting' Dimension (Linking People with Information) — Capturing and disseminating
explicit knowledge through computer networks, repositories, and document management
systems.
• The 'Connecting' Dimension (Linking People with People) — Enhancing tacit knowledge flow
through expertise directories, online Communities of Practice, and video conferencing.

Key Technologies and Tools

• Enterprise Portals — The 'face' of KM; personalised gateways integrating collaboration,
transaction, and information management into a single point of interaction.
• Expertise Locator Systems (ELS) — Systems such as NASA's Expert Seeker or IBM's
BlueReach use digital profiles and data mining to help employees find experts for immediate
consultation.
• E-Learning Environments — Provide 'knowledge scaffolding' by delivering training
electronically to the desktop, allowing learners in different locations to collaborate on
assignments.
 • Knowledge Mobilisation — Wireless technologies (cell phones, PDAs, WLAN) allow 'road
warriors' to access intellectual assets 'anytime, anywhere, and on any device.'
• Enterprise 2.0 — Wikis and blogs facilitate bottom-up, peer-to-peer authoring and harness
collective intelligence, improving databases as more people use them.

Strategic Categories of Digital Transfer

• Near Transfer — Explicit knowledge shared between similar teams, e.g., Ford's Best Practice
Replication where plants receive proven improvements via intranet.
• Far Transfer — Tacit knowledge shared between teams in different contexts, often via online
forums or physical movement of people.
• Expert Transfer — Seeking technical knowledge for non-routine tasks through systems like
Buckman Lab's TechForums.

Critical Success Factors in the e-World

While technology is a crucial enabler, it is not a panacea: effective KM is 80% related to
organisational culture and human factors and only 20% related to technology. Key risks include:
• Technology must 'fit' the organisation's people and processes — otherwise it will not be used.
• Email overload and 'info-glut' can paradoxically become barriers, making it difficult to find
specific knowledge.
• A culture of openness and trust is essential — individuals will not share high-value knowledge
electronically without trusting the recipient and the system.

5.9 KM System Tools — Classification and Inventory

KM system tools are the integrated technologies and mechanisms developed to support an
organisation's efforts to discover, capture, share, and apply knowledge. They form one leg of the
'three-legged stool' of people, processes, and technology — with people considered the most
critical component.

Classification Framework 1: Functional Families

One framework groups tools into three families based on their core functional focus:
• Content Family — Static document and information realm: document management systems
(DMS), taxonomies, document templates, best practice repositories.
• Collaboration Family — Messaging and workflow coordination: groupware (Lotus Notes,
Microsoft Exchange), messaging, whiteboarding, P2P platforms.
• Computation Family — Algorithm-intensive tasks: search engines, clustering, Social Network
Analysis (SNA), business intelligence tools.

Classification Framework 2: Process-Oriented KM Systems

A second major framework categorises tools based on the four primary KM processes:
• Knowledge Discovery Systems — Support development of new knowledge: data mining,
brainstorming retreats, socialization mechanisms.
• Knowledge Capture Systems — Elicit and preserve expert knowledge: storytelling, concept
maps (CmapTools), RFID technology, lessons learned databases.
• Knowledge Sharing Systems — Organise and distribute knowledge: Enterprise Portals,
Expertise Locator Systems (White Pages), collaborative computing environments.
• Knowledge Application Systems — Help individuals apply others' knowledge without learning
it themselves: expert systems, Case-Based Reasoning (CBR), decision-support systems, help
desks.
Standard KM Toolbox
• After Action Reviews (AARs) — Structured tool to capture lessons during/after a project.
• Communities of Practice (CoPs) — Networks linked by a common theme to develop and
share knowledge.
• Expertise Locator Systems (White Pages) — Online directories to find 'who knows what.'
• Lessons Learned Systems (LLS) — Repositories providing insights from previous similar
experiences.
• Peer Assists — Meetings where a team seeks knowledge from others before starting a
project.
• Social Network Analysis (SNA) — Technique to map and visualize knowledge flows and
relationships.
• Storytelling — Narrative techniques to capture tacit knowledge and organisational culture.

Emerging Web 2.0 Tools

• Blogs (k-logs) — Narratives and searchable archives of 'think-aloud' behaviour.
• Wikis — Collaborative authoring tools allowing users to publish and edit content easily.
• Knowledge Dashboards — Personalised interfaces for tracking key knowledge assets and
real-time activities.
• Knowledge Mobilisation Tools — Wireless technologies providing access to expertise
anytime, anywhere.

Critical Success Factors for KM Tool Usage

• Fit the culture — Systems will not be used if they do not integrate with how people actually
work.
• Align with strategy — Tool selection should be driven by business needs (low-cost vs.
differentiation strategy).
• Avoid over-engineering — Start small and simple with proven tools before scaling up.
• Incentivize usage — Public recognition, reward schemes (e.g., Infosys Knowledge Currency
Units), and formal performance evaluations encourage contribution.

5.10 Neural Networks for Knowledge Management

Artificial Neural Networks (ANNs) are an AI mechanism that attempts to simulate human
intelligence by recreating the brain's connective physiology. They are designed to learn how to map
information between two vector spaces: input and output. In KM, they serve as a foundational
technology for knowledge discovery systems.

How Neural Networks Work

• Architecture — Consist of an input layer, output layer, and one or more hidden layers of
processing neurons.
• Weights — Connections between neurons have 'weights' signifying the strength of
relationships; these are adjusted during learning.
• Supervised Learning (Back Propagation) — Network is presented with examples and
expected outputs; calculates errors and iteratively adjusts weights until it accurately predicts
new data.
• Unsupervised Learning — Hidden neurons self-organise by competing to respond to input
patterns without pre-defined sample outputs.

Primary KM Applications
 • Predictive Data Mining — Develop models predicting outcomes based on input variables; e.g.,
market segmentation and customer 'churn' prediction in telecommunications and insurance.
• Data Clustering — Kohonen networks use unsupervised learning to group complex data into
natural clusters, helping managers visualize relationships between knowledge domains.
• Behaviour and Tactical Modeling — Integrated into Context-Based Reasoning (CxBR)
systems to capture expert tactical knowledge (e.g., how an expert driver reacts to a traffic
light).
• Expertise Location — NASA used data mining algorithms to build expert profiles by
automatically mining information published by employees on internal webpages.

Advantages and Limitations

• Advantage — Can solve a large variety of problems difficult to define with traditional rules;
successfully used to predict stock market variations and synthesize speech from text.
• Limitation — Often described as 'black boxes' — it is difficult for humans to understand exactly
how the network reaches a conclusion (unlike 'white box' methods like decision trees).
• Limitation — Not well-suited for capturing motor skills expertise (physical performance), which
is highly tacit and difficult to replicate electronically.

5.11 Association Rules for Knowledge Management

Association rules (also called affinity analysis) are a data mining technique used within knowledge
discovery systems to identify items or variables that are closely linked within a dataset. The method
focuses on uncovering previously unknown relationships — specifically which items are likely to
occur or be purchased together. It is primarily a descriptive model, describing natural groupings in
data rather than predicting a specific outcome variable.

Key Algorithms
• Apriori — Association rule algorithm requiring discrete or logical input fields (e.g., true/false,
1/0). Generally faster to train; used to identify products likely to be purchased together.
• Generalized Rule Induction (GRI) — More flexible; handles continuous, numeric, or logical
data. Produces rules describing associations between attributes and a symbolic target.

Practical Applications
• Market Basket Analysis — Most well-known application; used by retailers to determine which
products customers frequently buy together in a single transaction, enabling targeted
marketing campaigns.
• Telecommunications — Identify which services or products customers are likely to bundle
(e.g., specific calling plans or hardware packages).
• Education — Identify which courses students are likely to take together, aiding curriculum
planning and scheduling.
• Healthcare — Link analysis identifies complex relationships between networks of physicians
and the specific prescriptions they write.

Role in Analytical CRM

Association rules are a vital part of analytical Customer Relationship Management (CRM). By
revealing previously unknown connections within data, businesses can perform market
segmentation and 'narrowcast' targeted emails or offers to specific customer groups based on likely
interests, applying collective organisational experience to improve services and increase revenue.
 5.12 Classification Trees for Knowledge Management
Classification trees (also called decision trees or induction trees) are hierarchical, acyclical
networks of nodes and links used to represent relationships and patterns within data. They are a
foundational tool for knowledge discovery systems, enabling organisations to uncover previously
unknown relationships and develop predictive or descriptive models.

Core Characteristics
• Structure — Root node (top level), intermediate nodes (decisions/tests on attributes), and leaf
nodes (final classifications or solutions).
• Inductive Learning — Generated through rule induction: a system learns associations by
sequentially examining example datasets, converting a table of known cases into a decision
tree and then into IF-THEN rules.
• 'White Box' Nature — Unlike neural networks, classification trees are transparent. Users can
easily visualize and understand exactly how the system reached a specific conclusion.

Applications in Knowledge Management

• Automated Content Classification — Automatically file unstructured documents into an
organised taxonomy or classification scheme, helping users navigate information.
• Problem-Solving and Diagnosis — Display alternative decisions and their implications; useful
in help desk systems and fault diagnosis.
• Customer Relationship Management (CRM) — Predict customer behaviour such as 'churn'
likelihood or which demographic combinations yield the highest probability of a sale.

Notable Algorithms and Case Studies

• C5.0 — Extracts rules from data by first building an induction tree; one of the most popular
algorithms.
• CART (Classification and Regression Trees) — Standard technique used to split data into
binary subgroups for prediction.
• Real Estate Appraisal — DBI (Difference-Based Induction) used to determine the incremental
worth of specific house features (e.g., value of a third bedroom vs. a fifth).
• Financial Services — Banks use classification trees to determine credit risk by identifying
features of past borrowers who defaulted.
• Medical Diagnosis — QUEST (Quick, Unbiased, Efficient, Statistical Tree) predicts medical
outcomes such as identifying patients who may need additional care post-surgery.

5.13 Data Mining and Business Intelligence

Data Mining (DM) and Business Intelligence (BI) are critical technological components of KM,
particularly within knowledge discovery systems. While often used interchangeably, they serve
distinct but complementary roles in identifying patterns and creating new explicit knowledge from
data.

Core Definitions
Business Focuses on providing decision-makers with valuable information and
Intelligence (BI) knowledge by discovering relationships across a variety of data
sources. Results in insights obtained by revealing previously unknown
connections, often presented via dashboards.
 Data Mining (DM) A class of database applications that look for hidden patterns in
groups of data to describe past activities or predict future behaviour.
A foundational technology for knowledge discovery in databases
(KDD).

The Overlap The primary area of overlap between KM and BI is knowledge

discovery. KM is broader, incorporating social mechanisms and both
tacit and explicit knowledge, whereas BI is primarily technical and
focused on explicit knowledge from data.

Objectives: Two Core Questions

• Predictive DM — 'What can happen?' Uses past history to predict future behaviour through
classification (sorting into predefined categories) and regression (estimating continuous
values).
• Descriptive DM — 'What happened?' Seeks patterns in past activities through clustering
(identifying natural groupings) or affinity analysis (Market Basket Analysis).

Three Primary Data Mining Techniques

• Symbolic / Inductive — Decision trees or rule induction; learns associations from examples.
Produces transparent 'white box' models easy for humans to understand.
• Connectionist — Artificial Neural Networks (ANNs); simulates human intelligence. Often
considered 'black boxes' — internal reasoning is difficult to visualize.
• Statistical — Linear regression, K-means clustering, discriminant analysis; finds correlations
and natural groupings in numerical data.

5.14 Decision-Making Architecture in Knowledge Management

In the context of KM, decision-making architecture is evolving from a traditional linear model into a
multifaceted paradigm designed to address increasingly complex and 'wicked' organisational
problems.

The New Paradigm: Four Perspectives

The new architecture proposes developing multiple perspectives before proceeding to analysis:
• Technical Perspective — Objective analysis of alternatives and implementation of chosen
solutions. The primary focus of traditional decision support and KM systems.
• Personal and Individual Perspective — Accounts for the fact that different actors see problems
through the lens of their unique experiences, personalities, and attitudes toward risk.
• Organisational and Social Perspective — Recognises that different organisations and groups
may view the same problem differently based on their specific interests.
• Ethics and Aesthetics Perspective — Addresses high-level business ethics where no simple
true/false solution exists, requiring direct involvement of key stakeholders.

Collaborative Decision-Making
A significant architectural shift involves moving away from consensus-based models toward
collaborative decision-making. As demonstrated by Hewlett-Packard (HP), this model allows a
single person to remain the decision-maker while tapping into the 'power of many' (depth of
employee knowledge) and the 'power of now' (instant access to quality information), reducing risk
and increasing execution speed.

Knowledge-Based Problem-Solving Architecture

Solving problems through KM typically follows a structured three-step sequence:
• Analysis — Identifying the gap where the Target situation does not equal the Actual situation.
• Structuring — Prioritising the problem and specifying the exact knowledge needed for a
solution.
• Realisation — Implementing a solution to reach the point where Target equals Actual.

Supporting KM Tools
• Decision Trees — Provide a structured framework for displaying and evaluating alternative
decisions and their potential implications.
• Decision Processing Portals — Advanced enterprise portals (e.g., Business Objects, Cognos)
integrate BI to support self-service decision-making.
• Social Network Analysis (SNA) — Used as a diagnostic tool to identify and reinforce actual
decision-making paths and patterns of trust.
• Knowledge Application Systems — Utilise 'direction' (expert directs action without transferring
underlying knowledge) and 'routines' (embedded procedures) to guide decisions.

The 'Power of Now'

Modern decision-making is heavily influenced by the intensified speed of responsiveness. Rapid
technological shifts have shortened the 'comfort zone' for stakeholders, often requiring mission-
critical decisions in minutes rather than days. Because the required knowledge cannot reside in a
single individual, KM systems are essential for providing necessary inputs to managers exactly
when and where they are needed.

5.15 Data Management in Knowledge Management

In KM, data management refers to the systems and processes used to handle the most basic level
of the knowledge hierarchy. Data comprises facts, observations, or perceptions representing raw
numbers or assertions. Because data are typically devoid of context, meaning, or intent, they serve
as the 'building blocks' for information and knowledge.

The Data Level vs. The Knowledge Level

• Data Level — All documented knowledge stored in systems capable of incorporating,
processing, and exchanging data: databases, printed documents, libraries. Provides essential
input for business processes.
• Knowledge Level — Made up of the justified beliefs and experiences of individual members
and their interactions with one another; constitutes the 'organisational knowledge base.'

Data Management in KM Systems

• Databases and Data Warehouses — Databases are large collections of data organized for
rapid search; data warehouses aggregate data from multiple enterprise sources to support
strategic decision-making.
• Metadata — 'Data about data'; describes how, when, and by whom a set of data was
collected. Essential for understanding and utilizing information stored in data warehouses.
• Data Preparation — Cleaning and structuring data (e.g., eliminating 'outlier points') to make it
suitable for data mining.
 • Taxonomies and Vocabulary — Hierarchical frameworks for classifying knowledge; allow
users to navigate vast collections. Maintaining these is a critical, often underestimated, cost.

Strategic Considerations
KM is not merely about data management. A common pitfall is the creation of 'data tombs' — data
stores where information is deposited but never accessed again, resulting in lost opportunities for
knowledge discovery. Successful KM ensures data management technology is an enabler,
integrated with people and processes to ensure the right information reaches the right person at the
right time.

5.16 Knowledge Management Protocols

In KM, protocols refer to formalised sets of rules, procedures, or policy frameworks that guide how
knowledge is captured, validated, and applied within an organisation. They range from specific
healthcare claim processing to broad organisational standards.

Automated and Managed Protocols

In specialised environments like healthcare and insurance, protocols act as actionable knowledge
objects used for decision support:
• Discovery and Approval — New protocols can be discovered via automated learning
algorithms, but a 'person in the loop' is required to review and approve before deployment.
• Protocol Manager — A specific KM role using a GUI to browse, search, create, and maintain
protocols based on documented and tacit knowledge.
• Application — Approved protocols are stored in repositories, accessed by algorithms to
process claims and predict the need for additional documentation.

KM Policy Frameworks as Protocols

Broad organisational protocols are often captured in a Knowledge Management Policy Framework.
The UK's Office of the e-Envoy proposed a framework based on ten key areas:
• Knowledge Capture — Policies for identifying and capturing explicit and tacit knowledge.
• Knowledge Transfer — Processes for moving knowledge between sources and forms.
• Knowledge Retention — Strategies for preserving knowledge during organisational changes.
• Content Management — Efficiently managing the organisational knowledge base.
• Enabling Communities — Supporting knowledge-based community working (CoPs).

Industry Standards and Governance

• ITIL (Information Technology Infrastructure Library) — A public-domain methodology for
managing IT infrastructure as a unified set of integrated processes (e.g., availability and
capacity management).
• Australian Standard (AS 5037) — One of the first national KM standards; provides a
methodological approach for identifying and harnessing knowledge through the interaction of
technology, process, content, and people.
• Business Rules — Clear protocols should be defined for document publication and
communication, specifically regarding alerting, escalation, and urgent queries.

Governance and Security Protocols

• Access and Validation — Systems like the Rolls-Royce KM Lessons Learned Log use detailed
peer review procedures to ensure only high-quality information is circulated.
 • Confidentiality — Organisations must balance open knowledge exchange with protection of
company secrets through nondisclosure agreements and IT safeguards to prevent 'knowledge
leakage.'

5.17 Managing Knowledge Workers

Managing knowledge workers — formally trained experts who primarily develop, apply, and share
knowledge — requires a fundamental shift from traditional Industrial Age 'command and control'
management. Because knowledge resides in people's heads, managing the knowledge itself is
neither possible nor desirable; management must focus on creating an environment where people
are encouraged to create, learn, share, and use knowledge together.

The Shift in Management Philosophy

• Managing as 'Volunteers' — Peter Drucker suggests knowledge workers should be managed
more like volunteers to maintain loyalty, as they can easily take their expertise elsewhere.
• Integration over Control — Management now consists of guiding and integrating the
autonomous but interconnected work of highly skilled people, not minute-to-minute
supervision.
• Outcome Focus — Managers are responsible for the application and performance of
knowledge rather than supervising tasks.

Creating a Knowledge-Sharing Culture

A major challenge is that many knowledge workers have been trained to believe 'knowledge is
power' — that sharing it may make them dispensable. Managers must:
• Build Trust — People will only build and share knowledge in an environment of trust and
appreciation.
• Eliminate 'Blame and Shame' — A culture that inspires innovation must allow learning from
mistakes rather than penalising them.
• Promote Openness — Moving toward flatter organisational structures reduces status barriers
that impede the exchange of ideas.

Key Strategies for Managing Knowledge Work

• Formal Integration — Knowledge sharing should be a formal part of job descriptions and
performance appraisals, not a separate add-on.
• Aligning Rewards — Incentives must reward not just individual effort but also knowledge
sharing and reuse of existing knowledge.
• Providing 'Slack Time' — Managers must allow employees time to seek knowledge, reflect on
their work, and have 'knowledge accidents' through informal interactions.
• Recognising Different Worker Types — Research identifies nomads (highly interactive and
mobile) and analysts (independent, frequent IT users). Design environments ('ba') tailored to
each style.

Retention and Knowledge Harvesting

• Exit Interviews — Leverage departures to capture key insights before a worker leaves.
• Knowledge Harvesting — Systematically elicit and document the tacit 'know-how' of experts,
making it available to the rest of the organisation.
• Mentoring and Apprenticeships — Allow for the socialisation of tacit knowledge, helping new
employees shorten their learning curves.
 KEY The successful management of knowledge workers depends more on human
PRINCIPLE factors — leadership support, culture, and motivation — than on the
implementation of any specific technology.

5.18 The CRISP-DM Process Model

The Cross-Industry Standard Process for Data Mining (CRISP-DM) is the industry-standard
framework for structuring knowledge discovery projects. It consists of six iterative, cyclical steps —
the process is not linear; later phases often require returning to earlier ones.

# Phase Key Activities

1 Business Define clear business objectives and assess the current situation.
Understanding Ensure the study has a specific goal, such as increasing the ROI
of a marketing campaign.

2 Data Understanding Collect initial data, describe its contents (unique values, ranges),
and verify data quality to identify missing or inconsistent
information.

3 Data Preparation Select relevant predictor variables, clean the dataset, and
transform variables (e.g., grouping continuous data into categories
like 'high' or 'low') for analysis.

4 Modeling Select and apply various data mining techniques (neural networks,
rule induction, etc.) to build models that predict future behaviour or
categorize information.

5 Evaluation Validate the developed model using a separate data subset to

determine if predictive accuracy is significant and if the model truly
meets the business goals from Step 1.

6 Deployment Implement the model into the organisation's live environment to aid
decision-making, and monitor ongoing performance to ensure
continued effectiveness.

IMPORTANT CRISP-DM is highly iterative — later phases frequently require returning to

earlier ones. For example, findings during Evaluation may require revisiting
Data Preparation or even redefining Business Understanding.

5.19 Real-World Examples: Successful KM Portals

Several organisations have implemented successful KM portals — personalised gateways or 'one-
stop-shops' that integrate information, collaboration tools, and work processes into a single
interface.

Organisation Portal Name Key Features

Accenture Knowledge Xchange Hosts content from proposals to expert links;

(KX) ~400,000 knowledge capital orders monthly for
 70,000 professionals.

Cable & Wireless Phoenix Central intranet 'K ecosystem'; real-time project
cockpit for all stakeholders; integrates document
management, e-learning, and discussion.

Ernst & Young KnowledgeWeb Access to 4,600+ databases of business

(KWeb) intelligence, global news, and internal research;
complements the Ernst & Young Online extranet.

KPMG KWorld Single gateway to all knowledge resources

including regional intranets; 'Skills & Experience'
system locates experts; KClient shares work-in-
progress securely with clients.

World Bank YourNet HR-centric portal empowering staff with a

common 'idiom'; decentralised expert content
creation; ten-fold increase in click-throughs vs.
previous site.

IBM W3 Rated employees' most credible source;

displaced managers and co-workers as primary
info source; specific 'portlets' save employees
valuable time.

Eli Lilly myELVIS Virtual Information Service; subject guides,

CommunitySpace, and a learning portal;
launched in a focused 90-day effort to link people
globally.

Siemens ShareNet Helps 7,000 sales/marketing staff share project

knowledge; uses structured questionnaires and
chat rooms; taps into local innovation across
global markets.

Infosys KnowledgeShop 'Learn Once, Use Anywhere'; helps refine

software and reuse code; unique 'Knowledge
Currency Units' (KCUs) incentive scheme for
rating quality assets.

Veterans Health Virtual Learning Internet-accessible portal for 219,000 employees

Admin. Centre (VLC) sharing informal knowledge and best practices;
significantly reduces 'reinventing the wheel'
across medical centres.

Ford The Hub Evolved from a basic search engine into a

gateway where each division can develop its own
'community' sub-portal with specific tools and
expertise locators.

Bayer KIBIT Hosted on BayNet intranet; facilitates download

of over five million documents monthly.

SUMMARY Effective KM portals act as personalised, single-point-of-access gateways

that integrate information, collaboration, e-learning, and expertise location —
making the right knowledge available to the right person at the right time.
EXAM QUESTION BANK

2 MARK QUESTIONS
 Define Knowledge Transfer
 Define Knowledge Sharing
 What is tacit knowledge?
 What is codification strategy?
 What is personalisation strategy?
 What is AAR?
 Define CRISP-DM
 What is data mining?
 What is neural network?

5 MARK QUESTIONS
 Explain barriers to knowledge sharing
 Explain codification strategy with example
 Explain personalisation strategy
 Differentiate AAR and storytelling
 Explain types of lesson sharing
 Role of Internet in KM
 Explain KM tools classification
 Explain neural networks in KM

10 MARK QUESTIONS
 Explain Knowledge Transfer process with diagram
 Compare Codification vs Personalisation (table + diagram)
 Explain benefits of combining both strategies Codification and Personalisation
 Explain lesson sharing methods with examples
 Explain IT tools supporting personalisation
 Explain KM system tools classification
 Explain Decision-making architecture in KM
 Explain CRISP-DM model with diagram
 Discuss Knowledge Transfer & Sharing in detail with barriers and success factors
 Explain integrated KM strategy with 80/20 rule
 Explain Data Mining and Business Intelligence in KM
 Explain Neural Networks, Association Rules, and Classification Trees
 Explain Knowledge Worker management strategies