Identifying Viable "Need–Solution Pairs": Problem Solving Without Problem Formulation

Author(s): Eric von Hippel and Georg von Krogh

Source: Organization Science, January–February 2016, Vol. 27, No. 1 (January–February
2016), pp. 207-221
Published by: INFORMS
Stable URL: [Link]
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide
range of content in a trusted digital archive. We use information technology and tools to increase productivity and
.facilitate new forms of scholarship. For more information about JSTOR, please contact support@[Link]

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at
[Link]

INFORMS is collaborating with JSTOR to digitize, preserve and extend access to Organization Science

This content downloaded from

[Link] on Wed, 28 Jan 2026 17:03:46 UTC
All use subject to [Link]
Organization Science
Vol. 27, No. 1, January-February 2016, pp. 207-221
ISSN 1047-7039 (print) | ISSN 1526-5455 (online) [Link]
©2016 INFORMS

CROSSROADS

Identifying Viable "Need-Solution Pairs":

Problem Solving Without Problem Formulation
Eric von Hippel
MIT Sloan School of Management, Cambridge, Massachusetts 02141, evhippel@[Link]

Georg von Krogh

ETH Zurich, 8092 Zurich, Switzerland, gvkrogh@[Link]

research and formal problem-solving practice begin with the assumption that a problem has been
identified or formulated for
Problem-solving solving. The problem-solving process then involves a search for a satisfactory or optimal
solution to that problem. In contrast, we propose that, in informal problem solving, a need and a solution are often
discovered together and tested for viability as a "need-solution pair." For example, one may serendipitously discover a
new solution and assess it to be worth adopting although the "problem" it would address had not previously been in mind
as an object of search or even awareness. In such a case, problem identification and formulation, if done at all, come only
after the discovery of the need-solution pair.
We propose the identification of need-solution pairs as an approach to problem solving in which problem formulation is
not required. We argue that discovery of viable need-solution pairs without problem formulation may have advantages over
problem-initiated problem-solving methods under some conditions. First, it removes the often considerable costs associated
with problem formulation. Second, it eliminates the constraints on possible solutions that any problem formulation will
inevitably apply.

Keywords', organizational economics; economics and organization; managerial and organizational cognition; organization
and management theory; decision making and theory of the firm; knowledge production; innovation; information
transfer costs
History: Published online in Articles in Advance December 30, 2015.

1. Introduction and Overview any payroll-processing deficiencies, but now I recognize

Problem-solving research and formal problem-solving that we do and also recognize that this new software

practice begin with the assumption that a problem may resolve them." Alternatively, instead of this type of
has been identified. The problem-solving process then "new-to-me" discovery, the employee may be stimulated

involves a search for a satisfactory or optimal solution to identify a new-to-the-world need-solution pair: "This
to that problem. In this article, we propose something payroll-processing approach gives me an idea to create
rather different. Often, we think, a need and a solution a better design for another situation entirely, one that I
are or can be discovered as a need-solution pair. When was not thinking needed improvement."
this is so, problem identification, if done at all, comes As a real-world example of a new-to-the-world inno
vation conceived of in this manner, consider
only after that discovery. Bernard

To illustrate what we mean, consider a firm employee Sadow's description of need-solution pair identification
walking through a trade show "just to see what is new." when inventing the rolling suitcase (Sadow 1972):
The employee encounters a booth offering new payroll Mr. Bernard D. Sadow, now 85, had his eureka moment
processing software (a solution) that claims to address in 1970 as he lugged two heavy suitcases through an air
certain needs better than earlier software. The employee port while returning from a family vacation in Aruba.
then mentally assesses the need-solution pair on offer Waiting at customs, he said, he observed a worker effort
for utility to the firm. "I wasn't thinking that we had lessly rolling a heavy machine on a wheeled skid.

207

This content downloaded from

[Link] on Wed, 28 Jan 2026 17:03:46 UTC
All use subject to [Link]
 Hippel and Krogh: Identifying Viable "Need-Solution Pairs"
208 Organization Science 27(1), pp. 207-221, ©2016 INFORMS

"I said to my wife, 'You know, that's what we need for 2. Problem Formulation-Related
"
luggage,' recalled. When he got back to
Mr. Sadow Literature
work [Sadow worked at a luggage company], he took In this section we provide a distillation of critical
casters off a wardrobe trunk and mounted them on a big
assumptions and choices in the problem-solving litera
travel suitcase. "I put a strap on the front and pulled it,
ture regarding problem formulation. Our purpose is to
and it worked," he said. (Sharkey 2010)
provide an informed contrast to the need-solution pair
This same process of need-solution pair identification perspective we discuss in this paper.
without problem identification or formulation in advance First, we note that those engaged in both formal prac
of "solving" is used in a great deal of informal personal tice and research on problem solving assume that a prob
lem must first be identified and formulated for solving,
problem solving as well. "Look at that bike baby car
and then addressed via problem-solving processes of
rier in the store window. It looks very sturdy and safe.
various types. This framing is ubiquitous. With respect
I wasn't thinking I needed a baby carrier, but that one
to research, a statement by Volkema (1983, p. 640)
could be a very useful solution element in a new pattern
is representative. He argues that solvers would be bet
of daily activity I just now envision: I could use my bike
ter off identifying an optimal problem formulation at
to carry my baby to preschool instead of using the car."
the outset because of the irreversibility of the problem
In the need-solution discovery and evaluation process
solving process: "Because the amount of information
just illustrated, there is no independent problem identi
needed to change a decision is much greater than the
fication or formulation step at all. Instead, the problem
amount needed to make it initially..., reformulation of
solver discovers or envisions a new and possibly useful
the problem becomes less likely once a particular for
need-solution pair and then compares it with a preex
mulation is selected and pursued. This places added
isting situation to determine if the new one is preferred.
pressure on decision makers to avoid premature clo
One could dub the inferior older arrangement "a prob
sure and to select 'optimal' problem statements." With
lem" once viewed relative to the new arrangement. How
respect to practice, Spradlin (2012, p. 85), for example,
ever, this problem is only specifiable post hoc. In net,
explains that "Indeed, when developing new products,
what we see is a problem-solving process that begins
processes, or even businesses, most companies aren't
with identification of viable need-solution pairs rather
sufficiently rigorous in defining the problems they're
than with the formulation of a problem.
attempting to solve and articulating why those issues are
What we term a "need-solution pair problem-solving
important.... Many organizations need to become bet
process" is presently unknown to problem-solving liter ter at asking the rightquestions so that they tackle the
ature in the management and organization fields as well
right problems." Dictionary definitions are in line with
as formal problem-solving practice. We think it flour such a sequence of events. Thus, in the Oxford Dic
ishes in real life but can offer no evidence beyond illus
tionary, a problem is defined as "a matter or situation
trative examples at this point. In this article, we will as unwelcome or harmful and needing to be
regarded
postulate and discuss the need-solution pair problem dealt with and overcome."1
solving process, proposing that under some conditions The assumption that problem identification or formu
it will be superior to problem-solving processes built lation precedes solving holds both for situations where
around the solving of problems specified ex ante. We a problem is fixed at the beginning of solving and
will conclude by suggesting that these matters merit fur remains unchanged, and situations where the initial
ther investigation.
problem specification is progressively reformulated or
In §2 we briefly review core assumptions in rele re-specified as problem solving proceeds (e.g., Schon
vant literature on problem identification and formulation 1983, Smith and Eppinger 1997, Thomke and Fujimoto
and problem solving. In §3, we introduce the concept 2000, Kurup et al. 2011).
of need and solution landscapes and propose the con
cept of viable need-solution pairing. In §§4 and 5 we 2.1. Identifying the "Underlying" Problem to
discuss the richness and complexity of real-world need Be Solved
and solution landscapes, and the advantages and costs Problems can and often do precede problem solving
of widespread landscape search. In §6 we further illus when they press themselves upon one's attention inde
trate need-solution pair discovery. In §7 we propose pendent of and prior to any solution being evident (e.g.,
next steps for need-solution pair theory development "My head hurts" or "This experiment failed"). When
and research, and we note useful links to other research
problems do present themselves, experts and everyday
streams. In §8 we offer suggestions for improvements solvers alike realize that the troublesome manifestation
to present-day problem-solving practices attainable with may be only a symptom of a cause that lies deeper
increased awareness of the value of need-solution pair (see Emirbayer and Mische 1998). For this reason, in
problem solving. In §9 we briefly conclude the paper. depth exploration seeking underlying causes, i.e., the

This content downloaded from

[Link] on Wed, 28 Jan 2026 17:03:46 UTC
All use subject to [Link]
Hippel and Krogh: Identifying Viable "Need-Solution Pairs"
Organization Science 27(1), pp. 207-221, ©2016 INFORMS 209

"real problem," is routine in many fields. For exam search (e.g., Levinthal and Posen 2007, Ghemawat
ple, a skilled doctor visited by a patient complaining of and Levinthal 2008). NK search begins with a well
a headache will routinely consider possible underlying structured problem statement. A precisely describable
causes, rather than simply treating the headache. Engi solution landscape is then searched by first casting
neers and scientists also seek underlying causes: "Yes, rough-mesh digital nets over the whole landscape to dis
that valve is corroded and must be replaced, but we must cover the approximate locations of desirable solution
understand why the valve corroded and solve that under "peaks." This is followed by the casting of progres
lying problem, too." "Yes, that physics experiment failed sively finer-mesh nets over promising areas to discover
to provide the answer we expected: what is the under a more satisfactory or even optimal solution. Rules such
lying reason?" Students of problem solving also under as "breadth-" or "depth-first" for visiting edges and ver
stand this. As Schwenk and Thomas (1983) and others texes in a graph have been formulated and examined
explain, problem formulation is an activity that not only for their ability to reach certain goals or results within
involves problem finding, but also in-depth exploration constraints (for a comprehensive review of algorithmic
for underlying causes for the problem as observed. The search, see Corneil 2004).
underlying cause discovered then becomes the problem In management, a particular emphasis is put on the
to solve. When solved, this will also solve the surface economics of search—the efficiency and effectiveness
manifestation initially observed. of search strategies and procedures (e.g., Fleming and
Practical methods have been developed by many to Sorenson 2004, Laursen and Salter 2006, Garriga et al.
encourage problem formulators to look below surface 2013). Often, in the case of managerial problems, solvers
symptoms (e.g., Ishikawa 1968, 1976; Volkema 1983). do not expect to find the optimal solution, given prob
Thus, Ishikawa developed and promoted cause-and lem complexity and the level of resources available to
effect diagrams (sometimes called "fishbone diagrams") conduct the work. Under these conditions, "satisficing
to help problem solvers to organize their searches into search" algorithms are used to identify any solution
deeper causes that might underlie a problem. His tech deemed satisfactory, where no distinction is being made
nique and variations have been widely used in industrial among alternative satisfactory solutions (Simon 1978,
process design and elsewhere. In cases when failures Greiner 1996, Greiner et al. 2006).
(i.e., problems) of a specific type would be predictably Also within the management literature, researchers
extremely costly, as in a nuclear reactor failure or air have studied how problem formulation activities are
plane malfunction, analytic methods have been devel and should be organized (Lyles and Mitroff 1980, Baer
oped to search for likely underlying causes of failure et al. 2013, Ben-Menahem et al. 2015). Thus, Mintzberg
before any actual failure occurs. Fault tree analysis is et al. (1976) found that, under conditions of high
an example of such a method (e.g., International Elec uncertainty in decision making, problem solvers work
trotechnical Commission 2006). very incrementally, iteratively formulating problems and
rapidly following up with search for solutions. When
2.2. Formulating Problems for Solving the identified problems are very uncertain or ill struc
Once a problem has been identified, it is then commonly tured, Pich et al. (2002) argue, problem solvers should
"formulated" for solving. In this regard, Simon (1973) explore a variety of solution paths independently and
drew a crucial distinction between well-structured and then select the best path after careful ex post analysis.
ill-structured problems, noting that only well-structured More generally, in what might be called literature on the
problems were suitable for algorithmic solution. He also "management of the unknown," Pich et al. (2002) and
pointed out that problems that are initially ill structured Abbott (2005), among others, suggest ways of coping
are converted into well-structured ones by the efforts with unforeseen influences that inevitably arise as novel
of problem formulators (Simon 1977, p. 309): "In gen problem-solving projects unfold.
eral, the problems presented to problem solvers by the
world are best regarded as ill-structured problems. They 2.3. The Difficulties of Problem Formulation
become well-structured problems only in the process of Obtaining problem solutions that function successfully
being prepared for the problem solvers. It is not exag when applied requires that problems as formulated
gerating much to say that there are no well-structured adequately represent the real-world situation. Scholars
problems, only ill-structured problems that have been understand that problem formulation is difficult, and that
formulated for problem solvers." formulations used may often be defective. Frequently,
The core of problem solving via algorithmic search this is because the formulations do not contain all rele
is creating a problem statement that contains a vant problem-specific and contextual information. When
well-structured problem specification. Once that has this is so, the result can be failure when the solution
been accomplished, algorithmic search can be applied. developed is applied to the intended real-world contexts.
An example of an algorithmic method applicable Incomplete or inaccurate information can affect a
to many well-structured problems is NK landscape problem-solving process when a problem is initially

This content downloaded from

[Link] on Wed, 28 Jan 2026 17:03:46 UTC
All use subject to [Link]
 Hippel and Krogh: Identifying Viable "Need-Solution Pairs"
210 Organization Science 27(1), pp. 207^221, ©2016 INFORMS

identified or formulated, and/or when it is reformulated for problem-solving tournaments, where a fixed prob
to convert it from an ill-structured to a well-structured lem is posed by a "problem owner" to external solvers,
form. Simon (1977, pp. 239-240), explains: and awards are given for the best solution to that fixed

Consider the Missionaries and Cannibals problem (Wallin et al. 2013).

problem.
[This is a classic problem used in Artificial Intelligence
Finally, note that some "wicked" problems cannot
research. The puzzle follows: three missionaries be formulated at all because those with a stake in
posed
and three cannibals must cross a river using a boat that the formulation have nonoverlapping criteria for accept
can carry at most two people. If missionaries are present able solutions. This situation is especially acute when
on a river bank, they cannot be outnumbered by can solutions acceptable to one set of solution users cre
nibals. Also, the boat cannot cross the river by itself ate inescapable negative externalities to other potential
without people on board.] As it is usually discussed in users. For example, the enjoyment of national forests by
artificial intelligence research, it is a well-defined prob individuals practicing motor sports can create a serious
lem. The objects are specified, together with the legiti noise pollution problem for those seeking quiet enjoy
mate moves, constraints, and starting situation. The state
ment. If externalities are severe, and they can be when
space is, in fact, very small, consisting of only 10 allow
able states...
conflicting uses both require access to the same common
Contrast this "laboratory" Missionaries and Cannibals resource, there may simply be no acceptable formulation
that can be agreed upon (Rittel and Webber 1973).
problem with a real-life problem. In the real-life problem,
the number of persons the boat will hold may only be
determinable by trying it, or at least examining it; and
the fact that the boat leaks may only be discovered after 3. Viable Need-Solution Pairs
it is loaded. The real environment may permit alternative We have seen from our review of core assumptions

ways of crossing the river—by a ford a few miles further that the study of problem solving in the management
up, on a bridge, or by airplane. The actors may be joined and organization fields generally assumes that formu
by several more missionaries (or cannibals!) while the lation of a problem to be solved precedes the search
action is going on. for a solution. Recall that this holds both for situations
In real life there is not a single, static, well-defined where a problem is fixed at the beginning of solving
problem, but a constantly changing problem whose def and remains unchanged, and situations where the ini
inition is being altered by information that the actors
tial problem specification is progressively reformulated
recover from memory and by other information obtained
or re-specified as problem solving proceeds (e.g., Schon
from the environment's response to actions taken...
1983, Smith and Eppinger 1997, Thomke and Fujimoto
von Hippel and Tyre (1996b), and Tyre and von Hip 2000, Kurup et al. 2011). We have also seen that prob
pel (1997) demonstrated the consequences of problem lem formulation can be problematic.
formulations that transmitted inadequate information to In this section, we open the way to a new view of
problem solvers. They studied the two novel process the problem-solving process that may avoid problem
machines designed and constructed in response to fixed formulation altogether. To do this we must first make
specifications/fixed problem statements. They found that, our terminology clear. In psychology, economics, and
when the machines developed according to these specifi marketing, needs are clearly distinguished from wants.
cations were placed into factory service in the real world, Needs reflect a state of felt deprivation of some basic
they failed repeatedly and required design modifications requirement for living such as food, shelter, belonging;
to function as intended. The most common underlying wants are desires for satisfiers of needs. As the underly
cause for the field failures was found to be a lack of full ing motivator for wants is needs, we use that term in our
information about the real-world operating environment discussions here. Although needs are often considered an
in the original problem statement. As Simon noted, this individual-level construct, they are potentially collective
problem is not easily rectifiable by making a fixed prob at the levels of groups, communities, and organizations
lem statement more complete. Information needed by (see Hamilton 2003). For example, the need for trans
problem solvers that a problem statement "should" con portation from home to work can be individual as well
tain often changes as problem-solving work progresses. as collective.
Wallin et al. (2013) and Sieg (2012) show that the We propose that it is conceptually useful to think of
problems associated with inflexible initial problem state a pool of need-related information as the contents of a
ments grow more severe when solution seekers are sep need landscape that, along with the contents of a solu
arated from solvers by organizational barriers. This was tion landscape, are drawn upon for problem formulation
the case in the situation explored by Tyre and von Hippel and solving. As has been noted, and as we will further
(1997) just discussed: Machine function specifications see, these pools of information can be very rich and
(problems) were formulated by a production group in complex. They may contain tacit elements as well as
a machine user firm and outsourced to an independent encoded information held in the minds of individuals.
machine design firm. This also is commonly the case They may contain information about a setting or context

This content downloaded from

[Link] on Wed, 28 Jan 2026 17:03:46 UTC
All use subject to [Link]
Hippel and Krogh: Identifying Viable "Need-Solution Pairs"
Organization Science 27(1), pp. 207-221, © 2016 INFORMS 211

of use not known to anyone, such as the state of seawor individual's need in the case of urgent thirst, but get
thiness of a boat. For example, the contents of a need ting a less-delicious glass of water from the kitchen tap
landscape for a surgeon would include the working envi may be.
ronment plus professional and personal and Next we observe that "formulating a need" and "for
experiences
views and information, all content that could potentially mulating a problem" are really a single process. The
be used in problem solving. distinction is between content that falls within the need
The metaphor of "landscape" makes it intuitive to landscape and content that falls within the solution land
represent both need and solution landscapes as three
scape as a consequence of the formulation choices made.
dimensional surfaces on which the position of a point Consider the following example of successive need or
solution statements:
on the x and y axes specifies the location of informa
tion related to a specific formulatable need or solution. I need
special lumber for my construction project.
On the need landscape, the z-axis "height" of a point The lumber I need is available in Helsinki.
I need to arrange a lumber-buying trip to Helsinki.
represents the benefit to be attained by satisfaction of
the need at position x, y. On the solution landscape, the I will arrange to travel to Helsinki.
z-axis "height" of a point represents the cost required to I will arrange to buy the lumber.
provide the solution located at position x, y. I will arrange to transport it to my construction site.
Problem solving then consists of making a link In this sequence as given, the partition between need
between a specific point on a need landscape and a spe and solution landscapes (i.e., the gap in the sequence
cific point on a solution landscape. We term these linked of statements above) is the formulation of the need
points a "need-solution pair." This pair is viable if the to arrange a lumber-buying trip to Helsinki. Problem
benefit derived from a solution is equal to or higher than
solving steps then follow. However, the division can
the cost of providing it in that pairing. For example, equally well be located between any other two adja
consider the thick arrow in Figure 1 that connects a spe cent statements. Thus, we could also set our formulated

cific point on the need landscape with a specific point need statement to be the first item in the listed sequence:
on the solution landscape. Suppose the selected location "I need special lumber for my construction project." Suc
on the need landscape represents an individual's state cessive steps in the problem-solving process are then
of thirst, and the selected location on the solution land carried out on the solution landscape: "I find I can buy
a source of drinkable water. the lumber I need in Helsinki. I must travel to Helsinki
scape represents specific
Then that need-solution pair is viable if and only if the to arrange the purchase. I will arrange travel to Helsinki.
reward associated with that particular point on the indi I will buy the lumber, and transport it to my construc
vidual's need landscape (i.e., the intensity of the need tion site."

to slake thirst) is equal to or higher than the cost of

providing the solution (i.e., potable water) to the individ 4. Value of Rich Landscape Search
ual at issue from the particular source pointed to on the As we have said, the traditional practice of separat
solution landscape. Traveling to drink pure water from ing need- or problem-formulation steps from problem
a mountain spring may not be a viable pairing to the solving steps converts what could be a single continuous

Figure 1 Need and Solution Landscapes Connected by Need-Solution Pairs

Need landscape

Need-solution pairs

Solution landscape

This content downloaded from

[Link] on Wed, 28 Jan 2026 17:03:46 UTC
All use subject to [Link]
 Hippel and Krogh: Identifying Viable "Need-Solution Pairs"
212 Organization Science 27(1), pp. 207-221, © 2016 INFORMS

activity into two modules, separated by a problem state software. They studied the outcomes of problem-solving

ment. This practice does have value as a coordination challenges (i.e.,specific formulated problems) issued by
mechanism: A problem statement can be used to trans the crowdsourcing firm InnoCentive. They found that
fer a task to other entities (i.e., individuals or separate many problems that this firm's clients had found diffi
groups of solvers) supplying them with a fixed goal and cult to solve using their standard, in-house R&D staff
important related information for their problem-solving were easily solved by some individuals among many
work (March and Simon 1958, Baldwin and Clark 2000, recruited to search the solution landscape with diverse
Terwiesch et al. 2002). This is common in firms, for expertise sets. Each diverse solver deeply knew, in other
example, when individuals in the marketing research words, a somewhat different section of the solution land
department formulate a need for a new product, create a scape. The result was that a larger fraction of a rich
problem statement, and then transfer the work of devel solution landscape was searched (nonalgorithmically) by
oping a responsive product to developers. a diverse group of solvers than had been done by in
However, formulating a single fixed problem also house staff, and viable need-solution pairs were found
has major costs: it cuts problem solvers off from as a consequence (Jeppesen and Lakhani 2010; also see
much need landscape richness. As a result, what Penin and Burger-Helmchen 2012).
could be better solutions for the solution user become
inaccessible—because they are "on the other side" of
the problem-formulation/problem-solution divide. For 5. Complexity of Rich Landscape Search
example, consider a firm that uses chemical X in its Given the evidence of efficacy described above, one
production process. The firm might formulate a prob might think that a straightforward solution to the diffi
lem requiring solution thus: "We need a way to produce culties created by present methods of problem formula
chemical X for significantly less than our current cost tion would be to simply widen landscape search enough
of $Y per kilogram." Solvers can then conduct a search to provide access to all possible, viable need-solution
on the solution landscape thus defined, for points that pairs. However, the richness of information contained in
create a viable pairing with that fixed problem. How both need and solution landscapes can often make this
ever, the problem could also be formulated so as to shift impractical.
some need landscape information into the solution land As illustration of need and solution landscape rich
scape, as in the following: "Reevaluate our production ness, and the impracticality of "full search," consider
process and the role of chemical X in it." The enriched a problem-formulation method involving a search for
solution landscape might then include options like elim underlying causes. The "5 whys" method was developed
inating the need for chemical X entirely by developing by Sakichi Toyoda for use in the Toyota Motor Corpo
an altogether different and cheaper production process. ration ("5 Whys"). Toyoda specifies that one must ask
The benefits of rich landscape search have been dis "why" 5 times to get to what he terms the root cause of
cussed and documented by several scholars to date. a problem.2 An example follows:
Raymond (1999, p. 41) explains the value of rich search My auto will not start, (the initial problem state
in both need and solution landscapes in his proposal of ment)
a Linus's law of software debugging: "Given enough 1. Why ?—The battery is dead, (first why)

eyeballs, all bugs are shallow." In software, discovering 2. Why?—The alternator (a battery-recharging mech
(need landscape) and repairing (solution landscape) sub anism used in autos) is not functioning, (second why)
tle code errors or bugs can be very costly (Brooks 1979). 3. Why ?—The rubber belt driving the alternator has
However, Raymond argued, the same task can be greatly broken, (third why)
reduced in cost and also made faster and more effective 4. Why?—The alternator belt was well beyond its use
when it is opened up to a large community of software ful service life and was not replaced, (fourth why)
users. Raymond explains: "Each [user] approaches the 5. Why ?—The car was not maintained according to
task of bug characterization with a slightly different per the recommended service schedule, (fifth why, a root
ceptual set and analytical toolkit, a different angle on the cause) ("5 Whys")
problem.... So adding more beta-testers ... increases Via this illustrative example, one can see that each
the probability that someone's toolkit will be matched level of problem statement points to a different location
to the problem in such a way that the bug is shallow on the need landscape. If one focuses only on the first
to that person" (Raymond 1999, pp. 43-44). In other level problem statement, for example, it reflects a need
words, each diverse solver has information regarding a to start the car. The solution would clearly be to recharge
somewhat different section of the need and/or solution or replace the auto battery. Solving that formulation of
landscapes. As a result, a richer landscape search can be the problem would indeed create a functioning car in the
conducted via the aggregated efforts of the group. short term, but the answer to the second why reveals that
Jeppesen and Lakhani (2010) documented the bene this solution would not be viable for very long. Without
fits of rich solution landscape search in fields beyond the alternator running, the battery would soon discharge

This content downloaded from

[Link] on Wed, 28 Jan 2026 17:03:46 UTC
All use subject to [Link]
Hippel and Krogh: Identifying Viable "Need-Solution Pairs"
Organization Science 27(1), pp. 207-221, ©2016 INFORMS 213

again, and the car would again not start. As the third problem-solving paradigm that are common in manage
why reveals, the alternator cannot be restored to function ment and organizational theory, as well as in practice.
without installing a new drive belt, and so forth down In what we think is a novel contribution to the problem
through the levels. solving literature, we propose that one way to engage in
Of course, arbitrarily deciding that five successive need and solution landscape search is to not formulate
whys as presented in this example is sufficient to get to a problem before search. Instead, one may simply scan
a root cause or real underlying need greatly understates one's mind and/or the environment for need-solution
both the depth and breadth of possibly useful inquiry.
pairs that might fit one's own context. Next, these pairs
With respect to depth of inquiry, one might, for exam can be tested against need and solution landscape infor
ple, continue via a sixth why to ask why the car was mation for viability. In this case, problem formulation
designed using a failure-prone belt to drive the alternator comes only after discovery of c| potentially viable need
in the first place. At that level, for the first time possible solution pair, if, indeed, problem formulation is done at
needs to redesign the car and/or the corporate innovation all. (When the need and/or solution component of pairs
process become part of the need landscape considered. has tacit elements, it can be impractical to formulate the
With respect to the potential breadth of both need and
problem or solution judged viable and implemented in
solution landscapes, consider that we can begin with the encoded form even if one wishes to.) As we said in §1,
same initial observation ("My car will not start") and we think that this process is frequently used in practice,
carry the identification of needs and related solutions in but it has not yet been described as a problem-solving
an entirely different direction. We may, for example, ask
process to our knowledge.
a series of why questions related to the function the car
Recall our §1 example of a firm employee walk
is to provide. Why do I need to use the car at all on this
ing through a trade show who discovers new payroll
occasion, and what alternatives might there be? Why do
I need the car to start? I need to drive to the city center. processing software (solution) that claims to solve cer
tain problems (needs) relative to earlier software. The
Why? I have to attend a meeting at work. Why do I need
to take a car to the meeting?... This line of diagnostic employee then mentally assesses that need-solution pair
for viability within the firm after discovering that pair.
inquiry within the need landscape leads one to discover
"I wasn't thinking that we had any payroll-processing
a totally different set of solution options, having to do
deficiencies but now I recognize that we do (thus rec
with alternate ways of substituting for the function of
the ailing car. Perhaps one can borrow a neighbor's car. ognizing a problem worth solving only after encounter
ing a potential solution worth implementing). Further,
Perhaps one can take a train or a cab or a bike or go on
this new software may solve the problem. It looks bet
foot. Perhaps one can phone in to the meeting instead
ter than the system we are using now." Additionally or
of attending in person, and so forth.
Note that each of the many ways a given problem can alternatively, the employee may be stimulated to identify
be specified and solved may require different informa a new-to-the-world need-solution pair: "This payroll
tion that will be useful for problem solvers depending processing approach gives me an idea to create a better
upon the need inquiry and solution path taken. If, in design for another situation entirely, one that I was not
the example given just above, problem formulators or thinking of as problematic."
Problem solving via need-solution pairs may involve
problem solvers decide to redesign the automobile alter
nator to do away with wear-prone alternator drive belts, a problem statement that is, however, treated as a dispos

detailed auto design information will be very valuable able, alterable part of a need-solution pair rather than as
to problem solvers. On the other hand, if problem for a fixed objective. For example, suppose you say to me,
mulators or solvers decide to focus on alternative ways
"I would like to go out for dinner tonight. What restau
to get to the meeting independent of the nonfunctioning rant shall we choose?" I might respond, "Let's go to
auto, then information on local train schedules will be the skating rink instead. If we get hungry, we can grab
very useful, and auto design information will be irrele something at the snack bar." Note that in my response
vant. In effect, an individual or firm with a need cannot I offered you a complete need-solution pair and note
be sure to supply potentially relevant information about also that my response implicitly changed your awareness
the need landscape to solvers without first knowing the and problem statement to a different one. I broadened
"best" solution path, or the one that problem solvers will the "Where shall we go for dinner?" problem you ini
in fact take. Additionally, as noted earlier, these matters tially posed to "What shall we do tonight?" In other
can be anticipated to only a varying and generally lim words, I regarded your initial problem statement as a
ited extent before actual problem solving (Simon 1977, variable. You would not be astonished or offended that
von Hippel and Tyre 1996b). I did this: it is within the bounds of ordinary social
behavior. Indeed, you might flexibly counter by offering
6. Direct Discovery of Need-Solution Pairs still another need-solution pair for assessment, saying:
To this point we have explored some practical diffi "Well, the skating rink will be crowded tonight. Since it
culties with the problem formulation followed by the is Friday, let's drive up to the mountains instead and go

This content downloaded from

[Link] on Wed, 28 Jan 2026 17:03:46 UTC
All use subject to [Link]
 Hippel and Krogh: Identifying Viable "Need-Solution Pairs"
214 Organization Science 27(1), pp. 207-221, © 2016 INFORMS

hiking tomorrow. The weather is supposed to be beau conjecture should be rejected. Second, if need-solution
tiful this weekend." By doing this, you, too, regarded pair solving is being practiced, it would be worth study
my problem statement as a variable, implicitly changing ing how it is being done in detail, and how it can be
"What shall we do tonight?" to "What shall we do this done better. Pursuant to both objectives, we must iden
weekend?" and offering still another need-solution pair tify and draw upon extant streams of theory and research
for evaluation. that can offer assistance.
A great advantage of a need-solution pair problem
solving process is that the potentially relevant need and 7.1. Determining the Extent and Value of
potentially useful solution come packaged together. One Need-Solution Pair Problem Solving
need not then invest effort in formulating a problem or Problem solving via identification of viable need
searching either a complete need or solution landscape. solution pairs, as we have described it, has a clear as
One simply tests the proposed need-solution pairing for sociated marker: identification of such a pair without
viability against the points in the relevant landscapes that prior formulation of a problem. To determine how often
the identified need-solution pair happens to touch. this practice is followed will require studies specifically
designed to identify that marker. Consider that, often,
studies of problem solving begin with a list of formu
7. Next Research Steps lated problems: for example, how does department X in
At the most general level, we think the field of problem firm Y prioritize and solve a listed set of problems that
solving should reconsider what we see as a currently they have specified or that has been assigned to them
pervasive assumption that the problem-solving process for solution in period Z. Clearly, problem solving with
begins with the formulation or identification of a prob out problem formulation is likely to go undetected or
lem. It may be difficult to do this: the very term "prob to be undercounted in such studies. As an alternative
lem solving" seems to contain an implicit assumption approach, we propose basing studies upon samples of
that one starts with a problem. However, in this paper we solutions (changes) implemented rather than problems
have argued that there can also be great value in such a formulated. For example, one could begin with a list of
reconsideration. We have proposed the direct discovery the solved problem outputs produced by department X
of need-solution pairs as an approach in which prob over a period of time. Then one could work back from
lem formulation may not be required, and there may be each solved problem to explore the solving method used
others. to devise or discover what was eventually regarded by
Note that we are not proposing that the identifica department X as a viable need-solution pair.
tion of need-solution pairs is a panacea. Indeed, one Today, there are increasingly economical ways to
might argue that the conventional procedure of exhaus examine how people explore or browse because of the
tive search ofa well-defined solution landscape to solve availability of data on searches conducted digitally. Such
a prespecified problem will be much richer than the data cover the path of movements people make when

possibly serendipitous discovery of viable need-solution searching in a physical or virtual landscape, e.g., traces
pairs. However, on the basis of our earlier discussions, people leave in content browsing on the Internet or eye
we suggest that the conventional sequence only gives an
tracking patterns they display when they conduct their
illusion of richness. As we have seen, the actual solution search (Cutrell and Guan 2007). One may also ask peo
landscape segment selected for search via a formulated ple at the start of their searches why they are searching,
problem, relative to the full landscape, is likely to be or what they are searching for and then, by following
very narrow indeed. In net, given the probably typical the evolution of the search, identify whether and when
complexity of need and solution landscapes, all problem novel need-solution pairs emerge, with related purchases
solving approaches are likely to leave many potentially then documenting their viability from the perspective of
viable need-solution pairs undiscovered. the searcher. Scholars have found that browsing patterns
To begin research on the concept of the need-solution do or can predict future purchasing behavior, a possible
pair as a problem-solving strategy, we suggest two indicator that this kind of research can work (see Hui
related research directions. First, it will be important et al. 2009).
to explore whether present-day problem-solving practice With respect to comparing the potential value of need
already contains a significant element of direct search for solution pair search with solution space search based
viable need-solution pairs. If need-solution pair iden upon formulated problems, we conjecture that, under
tification without problem formulation is indeed more conditions where need landscapes are complex and many
efficient or effective than problem formulation followed and diverse need-solution pairs are potentially viable, a
by problem solving, it is very probable that people both problem-solving process based upon a search for viable
individually and within organizations are already fre need-solution pairs is likely to be effective and effi
quently doing this form of solving, although perhaps not cient. In contrast, under conditions where only a small
consciously. If the practice is not found, our theoretical portion of the need landscape can give rise to a viable

This content downloaded from

[Link] on Wed, 28 Jan 2026 17:03:46 UTC
All use subject to [Link]
Hippel and Krogh: Identifying Viable "Need-Solution Pairs"
Organization Science 27(1), pp. 207-221, ©2016 INFORMS 215

need-solution pair because of organizational, technical, presented to experimental subjects. One could then study
temporal, or other constraints, formulation of problems related variations in those subjects' abilities to discover
before solving may well be the most efficient way to viable need-solution pairs.
focus solver efforts on the relevant portion of the need It is also the case that individuals are likely to
landscape. have greater "power" to identify a particular need
It may be possible to compare the two processes by solution pair as a function of their individual person
running experiments based upon the variations in the alities and other characteristics (e.g., cognitive skills).
content of need and solution landscapes. For example, Thus, research has shown that the basic personality trait
recall that we have argued that problem formulation "openness to experience" characterizes someone who

can hide or prevent access to potentially viable need is intellectually curious and tends to seek new expe
solution pairs. To understand the extent of that problem, riences and explore novel ideas (Barrick and Mount
it may be possible to perform experiments to progres 1991, Zhao and Seibert 2006, p. 261). Such individu
sively shift the need landscape/solution landscape parti als are also described as creative, imaginative, reflec
tion by systematic variation in problem formulation. For tive, curious, and untraditional (McCrae and Costa 1985,
each formulation, more or less of the need landscape is George and Zhou 2001). They tend to produce creative
transformed into the solution landscape. Search for suc ideas (Martinsen 2011). Other individual characteristics,
cessful solutions can then be conducted and compared such as cognitive styles, divergent and holistic think
with the outcomes of unconstrained searches for viable ing, the use of metaphors, and analogical reasoning are
need-solution pairs. also linked to the production of creative ideas (Koestler
1964). Emotional states likely matter, too. Positive affect
7.2. Exploring Conditions for Effective can lead individuals to explore and seize opportune solu
Need-Solution Pair Search tions to satisfy their needs, rather than shy away from
Next, it will be useful to better understand the nature what they perceive as "difficult problems" (Adler and
and effectiveness of need-solution pair search. At this Obstfeld 2007).
point, it is not clear whether that process is intuitively Further, individuals' ability to conceive of novel need
done well (i.e., "as natural as breathing"). If so, we will solution pairs is likely to be affected by their prior
still want to better understand how it works. Alterna experience with objects and concepts related to those
tively, if it is difficult to do well, there will be benefit pairs. It has been shown that, due to an effect called
from research to explore how effectiveness and effi "functional fixedness," subjects who use an object or
ciency could be improved. Below, we suggest what we see it used in a familiar way are strongly blocked from
think are some useful starting points for further research using that object in a novel way (Duncker 1945, Birch
for both of these purposes. and Rabinowitz 1951, Adamson 1952). It has also been
First, it is clear that human agency considerations will shown that experimental subjects familiar with a com
have an impact on individuals' and groups' ability to plicated problem-solving strategy are unlikely to devise
discover viable need-solution pairs. (In sociology and a simpler one when this is appropriate (Luchins 1942).
philosophy, "agency" frequently refers to the capacity These effects have also been demonstrated in real-world
of an agent (a person, other living being) to act in a research settings (Allen and Marquis 1964). Relatedly,
specific context, using its abilities and resources. For individuals display the highest creativity when they
example, human beings act with intention and pursue apply problem-solving approaches that they do not rou
goals in specific contexts (Emirbayer and Mische 1998).) tinely use (Dane et al. 2011). Still further, organization
In that regard, recall our §1 example of Bernard and team-level factors (e.g., climate, openness) also have
Sadow, the inventor of the rolling suitcase. This individ been shown to shape creativity (Amabile et al. 2005,
ual was employed by a suitcase producer. For this reason Amabile and Pillmer 2012).
he is perhaps more likely than an ordinary suitcase Research will be required to determine whether these
transporting tourist facing identical stimuli to bring to individual-, group-, and organization-level factors matter
mind viable need-solution pairs with respect to suit a lot or a little with respect to the effective and efficient
cases. Similarly, in our §1 example of the individual identification of viable need-solution pairs. Once that is
happening to notice a safe and sturdy bike baby carrier understood, further work can determine how to apply
in the store window, recall that that person is described these insights to make the identification of viable need
as a parent who has the task of transporting a child to solution pairs more effective and efficient. For exam
preschool. This contextualization of the activity imbues ple, if functional fixedness is a major impediment, guid
the person with mindfulness, intent, and selective atten ance toward envisioning novel need-solution pairs can
tion relevant to the need-solution pair identified. An also be more systematically supplied. Thus, toolkits that
example of an experiment to explore the importance of can be helpful offer lists of possibly useful associations
human agency and context is to systematically vary sub for use by need formulators and problem solvers. TRIZ
ject agency and/or need-solution landscape information (Altshuller 1998) is an example. It is a toolkit that is

This content downloaded from

[Link] on Wed, 28 Jan 2026 17:03:46 UTC
All use subject to [Link]
 Hippel and Krogh: Identifying Viable "Need-Solution Pairs"
216 Organization Science 27(1), pp. 207-221, ©2016 INFORMS

based on the study of patterns of invention in the global many successful discovery processes often show impor
patent literature and on the premise that basic types of tant serendipitous events (Campbell 1960, Fioratou and
problems and solutions are repeated across industries Cowley 2009). A recent study of scientists involved in
and sciences. It offers a knowledge base and model interdisciplinary research shows that when these individ
based technology to enable a problem solver to sys uals become sufficiently and positively "surprised" about
tematically "try out" success-associated possibilities that the value of unexpected content, they will often redi
might otherwise not have been envisioned. rect their ex-post solution- search on new paths or even
Success might also be more likely if solver individuals restructure their problem specification (Foster and Ford
or teams are chosen to already have both need and solu 2003, p. 334).
tion landscape information, encoded and nonencoded, in Serendipitous events can also trigger new need
their minds. Individuals so equipped could more cheaply solution pairs, viable solutions to problems that had
do searches involving both landscapes relatively inex not previously been formulated in a solver's mind. The
pensively via thought experiments. This is why individ discovery of the first synthetic dye by William Henry
uals who, for example, themselves have a need (reside Perkin came about in 1856 quite serendipitously dur
in a real need landscape) and themselves also have solu ing the search for synthetic quinine to treat malaria
tion expertise are such powerful developers of novel (Banerjee 2014). In more recent times, the cosmetic
need-solution pairs. An example is software users who effects of Botox were serendipitously noticed by a
directly experience a need for improved software and are patient who was being treated with that drug for an
also expert programmers (Raymond 1999, von Krogh entirely different medical problem. The general value of
et al.2012). A more general example is user innovators this serendipitous discovery was then recognized by and
of all types who directly experience needs and discover acted upon by the patient's doctor (Carruthers and Car
that they also have the expertise required to solve them, ruthers 1992, Coondoo and Sengupta 2015).
often at low cost (von Hippel 1994, 2005, Chap. 5). Research on serendipity is also directed at improve
ments to practice. Some researchers are exploring
7.3. Integrating Existing Literature Streams how problem solvers can more frequently and effec
We suggest that research related to identification of tively make and recognize serendipitous discoveries
need-solution pairs can usefully draw upon and also (Van Andel 1994, Denrell et al. 2003). With respect to
reciprocally enrich other research streams. Here, we note increasing the sagacity that enables those encountering
how research on serendipity, gestalt solution recognition, happy accidents to recognize their potential utility, some
and emergent design development may be useful to the conceive of serendipity as a structure, a capability, an
study of need-solution pairs. opportunity, or a resource, rather than an event, involv
The term "serendipity" refers to "accidental sagac ing conscious efforts to turn chance into luck (Denrell
ity." Horace Walpole first devised this "very expres et al. 2003, de Rond and Thietart 2007, de Rond 2014).
sive word" in 1754 by referring to a Persian fairy tale The work we previously mentioned on human agency,
The Three Princes of Serendip in which the three were functional fixedness, richness of contextual information,
"always making discoveries, by accidents and sagacity, and the like clearly relates to enhancing or reducing that
of things which they were not in quest of..." (Walpole capability for specific solvers with respect to specific
1960, pp. 407-408). As he further noted, "You must events.

observe that no discovery of a thing that you are look Research on serendipity also explores how to artifi
ing for comes under this description" (Walpole 1960, cially increase the level of possibly valuable accidents
p. 408). Serendipity, as it is used today, is often con that problem-solvers can then assess and exploit. Thus,
sidered a very valuable mode of scientific discovery. computer scientists have experimented with systems that
Merton, having experienced his own happenstance dis increase the "chance" element in information retrieval by
covery of the word, transformed it into a concept in using data-mining systems in ways that offer unexpected
the domain of the sociology of science and cogently links to new and interesting information (Beale 2007,
argued for the inclusion of serendipity as a scientific Liang 2012). Some approaches found that it is even more
method alongside purposeful discovery by experimen useful to "augment user skills" in their information seek
tation (Merton 1968, Merton and Barber 2004). In the ing by visualization and artificial intelligence techniques
social sciences, grounded theory development empha (Beale 2007, p. 421) or by supporting the actual strate
sizes the value of serendipitous discovery of findings for gies that users said that they used in "making my own
constructing novel theory (see Glaser and Strauss 1967, luck" (Makri et al. 2014, p. 2181). Rich browsing venues
Bourgeois 1979). (i.e., internal and external knowledge repositories, pro
Sometimes, as researchers have found, a serendipitous fessional journals, scientific conferences, business con
event triggers a new solution approach to a formulated ferences, trade shows, and so forth) can also serve these
problem already being worked upon, and the records of functions (Bathelt et al. 2004).

This content downloaded from

[Link] on Wed, 28 Jan 2026 17:03:46 UTC
All use subject to [Link]
Hippel and Krogh: Identifying Viable "Need-Solution Pairs"
Organization Science 27(1), pp. 207-221, ©2016 INFORMS 217

Of course, it is also likely that research on need closer examination, will be found to deal with funda
solution pairs can reciprocally inform research on mentally different matters. Thus, in the seminal garbage
serendipity. By studying serendipity and serendipitous can model of organizational choice, already formulated
events as a trigger for combining and evaluating points problems and solutions are put into a "garbage can" and

on need and solution landscapes, scholars can systemat linked and acted upon via managerial decisions (Cohen
ically explore, for example, the characteristics of land et al. 1972, Padgett 1980, Fioretti and Lomi 2010). The
scape points and combinations that appear to be most utility of this research to studies of need-solution pair
susceptible to such events. solving is currently limited precisely because inputs to
Another research field that contains findings and the garbage can are preformulated problems and solu
thinking useful for those exploring need-solution pairs tions rather than precursor information such as that con
is the gestalt school of psychology. This school argues tained in need and solution landscapes.
that the basis of problem solving is restructuring, a type
of process in which the problem solver gains "insights"
about a problem and thus comes to see the require 8. Next Steps in Practice
ments of a problem situation in a new way (Ohlsson The eventual outcome of research such as that described
1984, p. 65; for an interesting historical perspective, see in the previous section will be an understanding of
ter Hark 2010). Gestalt psychologists distinguish "solu the value of need-solution pair problem-solving, and
tions that pop into mind" from solutions identified by improved methods to do it. What can practitioners do
purposeful search. A central assumption in gestalt theory today, while researchers are engaged in such activities?
is that a problem has underlying structural characteristics We suggest that it is possible to see present aspects
that the problem solver can manipulate in order to dis of problem-solving practice that can have the effect of
cover a solution (Wertheimer 1959). Recent work sug more broadly exploring both need and solution land

gests that pop-up solutions tend to result from parallel scapes simultaneously, and to utilize these with a clearer
processing of experience and the constraints people per understanding of that fact. In this section we will sug
ceive in a problem specification (Novick and Sherman gest some practical steps to utilize present-day problem
2003). Again, this research stream can offer valuable solving methods in a way that is sensitive to the value
insights, via both examples and theory, regarding the
of the discovery of need-solution pairs.
mental processes that may underlie need-solution pair
identification. 8.1. Need-Solution Pair Search Triggered by
A final example of a research stream that may use Broadly Formulated Problems
fully contribute to the study of need-solution pair solv A hybrid of solving a problem as formulated and open
ing is found in the design literature. Here most of the ing the way to the discovery of novel need-solution pairs
formal methods tend to both specify the design prob exists when problems are purposely formulated very
lem a priori and fix the solution set before problem broadly, leaving room for richer variation than a more
solving begins (Simon 1979). Such specification ide narrowly specified problem would allow. For example,
ally allows designers to partition the solution set into a firm manager might implicitly or explicitly formulate
subsets, wherein they can conduct more constrained a very broad problem statement such as, "Our firm is
search for the "best" designs. Researchers and prac willing to produce anything we can sell at a profit with
ticing designers both argue that such tight a priori our existing production machinery and distribution chan
problem specifications and solution-set approaches are nels." They then look to similarly situated firms to see
rarely applicable to design situations addressed by real what those firms are producing at a profit as a form of
world designers (Schon 1983). As a more appropri discovery focused on the identification of potentially rel
ate approach, Hatchuel and Weil (2002) developed the evant and viable need-solution pairs. Entrepreneurs not
concept-knowledge (C-K) theory of design. This down encumbered or empowered with preexisting capabilities
plays problem formulation in the design process and can look even more broadly: "What can the basis for
seeks to explain the systematic expansion of the solu a profitable business be for me? Let me search widely
tion space by generating new knowledge about hith for what others are doing successfully." As a second
erto unknown solutions (Le Masson et al. 2010). This step in each case, the firms test each of these solutions
kind of work can be insight generating with respect to (a successful need-solution pair for someone, but not
both how viable need-solution pairs emerge and how necessarily for them) against their own need and solution
the efficiency and effectiveness of that process might be landscapes, to see if those solutions form a potentially
increased. successful pairing for them as well.
Many additional research streams will doubtless be Formal problem-solving methods exist that centrally
found that can contribute useful inputs to the study of involve a search for successful need-solution pairs
need-solution pair solving. However, it is also impor guided by broad and flexible problem formulation. These
tant to note that some apparently related streams, upon include positive deviance studies and lead user studies.

This content downloaded from

[Link] on Wed, 28 Jan 2026 17:03:46 UTC
All use subject to [Link]
 Hippel and Krogh: Identifying Viable "Need-Solution Pairs"
218 Organization Science 27( t), pp. 207-221, © 2016 INFORMS

If, for example, one has the broad goal of helping a The users then apply the prototype in their own envi
community in the developing world to improve fam ronment using their own data. Based upon what they
ily health, the positive deviance study approach (e.g., learn in the trial, users then modify their need specifica
Krumholz et al. 2011, Bradley et al. 2012) would be to tion (Gronbaek 1989, pp. 114-116). A revised prototype
search for families in the community who stand out as incorporating the modified problem formulation is then
positive deviants with respect to good family health. The quickly developed and sent to the user. The trial-and
error learning and iterative problem-and-solution refor
assumption is that such individuals have somehow found
and are using a need-solution pairing that is successful mulations by developer and user are repeated until a
for them and that fits conditions found in the community. successful need-solution pairing is found.
Researchers then seek to understand what these deviants In the software field, it has been found that rapid
are doing that is linked to better health. Then, if that prototyping methods are not only less costly than tradi
need-solution pairing looks viable for others, they may tional, noniterative methods such as the waterfall devel
opment approach based upon a fixed problem, but that
attempt to diffuse it as a best practice in the community
they are able to "better satisfy true user requirements and
(see Szulanski 1996).
Firms doing lead user studies execute a similar ap produce information and functionality that is more com
plete, more accurate, and more meaningful" (Connell
proach (von Hippel 1986). For example, a lead user and Shafer 1989, p. 15). Various implementations of
search for collision avoidance methods to be applied to
these basic operating principles are today collectively
automobiles will be purposely broadened to encompass
termed "agile software development," where the empha
how anyone or anything successfully causes or avoids sis is on frequent and intense interaction between devel
collisions with other individuals or things. This can lead
opers and users. For example, one variation involves
to, for example, an exploration of how blind individuals a software user and a software developer jointly and
successfully navigate crowded sidewalks, and how mil frequently inspecting the emerging product (e.g., Ferre
itary munitions successfully seek and execute collisions et al. 2001). In these reviews, both need statements and
with targets—and how targets successfully deploy coun solutions developed are topics for review and change
termeasures. The focus in these studies is always discov (for a review see Conboy 2009).
ery of successful need-solution pairs that may also fit the
need-solution landscape of the firm engaged in search. 9. Conclusion
The next step is to see if any of these pairs actually do Years ago, Herbert Simon (1973) stated that the only
have a viable fit to points on both the need and solution well-structured problems were ill-structured problems
landscapes of that solution seeker or if they can be mod formulated for problem solving. We now think it use
ified to fit (Lilien et al. 2002, Churchill et al. 2009). ful to add that the only ill-structured problems extant
are those identified as problems to be solved. In this
8.2. Iteratively Reformulate Problems to Discover essay we have explored the additional option of discov
Need-Solution Pairs ery of viable need-solution pairs without ex ante identi

As was noted in the literature review, it is also possible fication or formulation of a problem. This option may be
to start problem solving with a formulated problem, and frequently used in practice today and may be valuable

then iteratively adjust it as problem solving proceeds in under conditions yet to be assessed.

order to increase the chance of discovering viable need

Problem solving through search for viable need
solution pairs (e.g., Kurup et al. 2011). In each cycle, a solution pairings is a generic process that can be con
ducted by individuals, teams, and organizations. It can
point on the solution landscape (i.e., a solution) is tested
be in-house, and/or can be outsourced via
against the intended point on the need landscape for via
pursued

crowdsourcing calls. We propose that it will be impor

bility. If the fit is not good, the problem formulation is
tant to learn how it works and can work in these differ
changed and solving recommences. The trial-and-error ent contexts and governance regimes (Felin and Zenger
cycle continues until an acceptable need-solution pair
2012, Lakhani et al. 2013, von Krogh et al. 2012). It is
ing is found or created (Marples 1961, Simon and Simon well known that a broad of contextual factors such
range
1962, Allen 1966, von Hippel and Tyre 1996b, Thomke
as organization culture and structure shape the effec
1998, 2003, Hsieh et al. 2007, Nelson 2008). tiveness and efficiency of organizational problem solv
Another example of such a method is the rapid pro
ing (Lyles and Mitroff 1980, MacDuffie 1997, Marengo
totyping method of product development, as practiced et al. 2000, Cohendet and Simon 2007, Le Masson et al.
by software development firms and in many other fields 2010). Clearly, if need-solution pair identification is
as well. In rapid prototyping, problem solvers respond
proven to be valuable, there is a great deal of work to do
to initial user need specifications by quickly developing to understand both how it works and how it can be most
and delivering to users (usually within weeks) an inex effectively implemented in the light of existing and new
pensive, easy-to-modify, working model that provides understandings. We hope that others will find it interest
much of the functionality contained in the initial request. ing to join us in considering and exploring these matters.

This content downloaded from

[Link] on Wed, 28 Jan 2026 17:03:46 UTC
All use subject to [Link]
Hippel and Krogh: Identifying Viable "Need-Solution Pairs"
Organization Science 27(1), pp. 207-221, ©2016 INFORMS 219

Acknowledgments Bourgeois LJ (1979) Toward a method of middle-range theorizing.

The authors thank professors Zur Shapira and Paul Adler, and Acad. Management Rev. 4(3):443^47.
three reviewers at Organization Science for excellent com Bradley EH, Byam P, Alpern R, Thompson JW, Zerihun A, Abeb Y,
ments on earlier versions of this manuscript. The authors Curry LA (2012) A systems approach to improving rural care in
are also very grateful to Burcu Kucukkeles, Oliver Alexy, Ethiopia. PloS One 7(4):e3402.
Nikolaus Franke, Jochen Runde, and Neil Thompson, as well Brooks PF (1979) The Mythical Man-Month: Essays on Software
as participants in research seminars at Warwick Business Engineering (Addison-Wesley, Reading, MA).
School and Universidad Carlos III Madrid, for very valuable Campbell DT (1960) Blind variation and selective retentions in cre
comments. The authors also thank Jessie von Hippel for her ative thought as in other knowledge processes. Psych. Rev.

very valuable editorial work. 67(6):380-400.

Carruthers JD, Carruthers JA (1992) Treatment of glabellar frown
lines with C. botulinum-A exotoxin. J. Dermatologic Surgery
Endnotes
1 Oncology 18(11):17-21.
See [Link]
Churchill J, von Hippel E, Sonnack M (2009) Lead user project
_english/problem.
handbook: A practical guide for lead user project teams.
2See [Link]
(October 1), [Link]
%20Project%20Handbook%20%28Full%20Version%[Link].
Cohen MD, March JG, Olsen JP (1972) A garbage can model of
References
organizational choice. Admin. Sci. Quart. 17( 1): 1—25.
Cohendet P, Simon L (2007) Playing across the playground: Para
Abbott J (2005) Understanding and managing the unknown: The
doxes of knowledge creation in the video game firm. J. Organ.
nature of uncertainty in planning. J. Planning Ed. Res. 24(3):
Behav. 28(5):587-605.
237-251.
Adamson RE (1952) Functional fixedness as related to problem solv Conboy K (2009) Agility from first principles: Reconstructing the
concept of agility in information systems development. Inform.
ing: A repetition of three experiments. J. Experiment. Psych.
Systems Res. 20(3):329-354.
44(4):288-291.
Connell JL, Shafer LB (1989) Structured Rapid Prototyping (Prentice
Adler PS. Obstfeld D (2007) The role of affect in creative projects
and exploratory search. Indust. Corporate Change 16(1): 19-50. Hall, Englewood Cliffs, NJ).
Coondoo A, Sengupta S (2015) Serendipity and its role in dermatol
Allen TJ (1966) Studies of the problem-solving process in engineering
ogy. Indian J. Dermatology 60(2): 130—135.
design. IIEE Trans. Engrg. Management 13(2):72—83.
Allen TJ, Marquis DG (1964) Positive and negative biasing sets: The
Corneil DG (2004) Lexicographic breadth first search—A survey.
Hromkovic J, Nagl M, eds. Graph-Theoretical Concepts in Com
effects of prior experience on research performance. IEEE Trans.
puter Science (Springer, Berlin), 1-19.
Engrg. Management 11(4): 158-161.
Cutrell E, Guan Z (2007) What are you looking for? An eye-tracking
Altshuller G (1998) 40 Principles: TRIZ Keys to Technical Innovation
study of information usage in web search. Proc. SIGCHI Human
(Technical Innovation Center, Worcester, MA).
Factors Comput. Systems (ACM, New York), 407-416.
Amabile T, Pillmer J (2012) Perspectives on the social psychology of
Dane E, Baer M, Pratt MG, Oldham GR (2011) Rational versus intu
creativity. J. Creative Behav. 46( 1):3—15.
itive problem solving: How thinking "off the beaten path" can
Amabile T, Barsade SG, Mueller JS, Staw B (2005) Affect and cre stimulate creativity. Psych. Aesthetics Creativity Arts 5(1):3-12.
ativity at work. Admin. Sci. Quart. 50(3):367—403.
de Rond M (2014) The structure of serendipity. Culture Organ.
Baer M, Dirks KT, Nickerson Microfoundations of strategic
JA (2013)
20(5): 342-358.
problem formulation. Strategic Management J. 34(2): 197—214.
de Rond M, Thietart RA (2007) Choice, chance, and inevitability in
Baldwin CY, Clark KB (2000) Design Rules: The Power of Modular
strategy. Strategic Management J. 28(5):535-551.
ity (MIT Press, Cambridge, MA).
Denrell Fang C, Winter S (2003) The economics of strategic oppor
J,
Banerjee J Sir William Henry Perkin and the Coal-Tar Colours.
(2014)
Management J. 24(10):977-990.
tunity. Strategic
The Victorian Web (August 28), [Link]
Duncker K (1945) On problem-solving, trans. Lees LS. Psych. Mono
science/[Link].
graphs 58(5):i—113.
Barrick MR, Mount MK (1991) The big five personality dimen
sions and job performance: A meta-analysis. Personnel Psych. Emirbayer M. Mische A (1998) What is agency? Amer. J. Sociol.
103(4):962-1023.
44(1): 1-26.
Felin T, Zenger TR (2012) Open or closed innovation? Problem solv
Bathelt H, Malmberg A, Maskell P (2004) Clusters and knowledge:
Local buzz, global pipelines, and the process of knowledge cre ing and governance choice. Res. Policy 43(5):914—925.

ation. Progress Human Geography 28( 1):31—56. Ferre X, Juristo N, Windl H, Constantine L (2001) Usability basics
for software developers. IEEE Software 18(1 ):22—29.
Beale R (2007) Supporting serendipity: Using ambient intelligence
to augment user exploration for data mining and web browsing. Fioratou E, Cowley SJ (2009) Insightful thinking: Cognitive dynamics
Internat. J. Human-Comput. Stud. 65(5):421—433. and material artifacts. Pragmatics Cognition 17(3):549-572.

Ben-Menahem S, von Krogh G, Erden Z, Schneider A (2015) Coordi Fioretti G, Lomi A (2010) Passing the buck in the garbage can model
of organizational choice. Comput. Math. Organ. Theory 16(2):
nating knowledge creation in multidisciplinary teams: Evidence
from early-stage drug discovery. Acad. Management J., ePub 113-143.
ahead of print June 16, [Link] Fleming L, Sorenson O (2004) Science as a map in technological
Birch HG, Rabinowitz HS (1951) The negative effect of previ search. Strategic Management J. 25(8-9):909-928.
ous experience on productive thinking. J. Experiment. Psych. Foster A, Ford N (2003) Serendipity and information seeking: An
41 (2): 121—125. empirical study. J. Documentation 59(3):321-340.

This content downloaded from

[Link] on Wed, 28 Jan 2026 17:03:46 UTC
All use subject to [Link]
 Hippel and Krogh: Identifying Viable "Need-Solution Pairs"
220 Organization Science 27(1), pp. 207-221, ©2016 INFORMS

Garriga H, von Krogh G, Spaeth S (2013) How constraints and knowl Lilien GL, Morrison PD, Searls K, Sonnack M, von Hippel E (2002)
edge impact open innovation. Strategic Management J. 34(9): Performance assessment of the lead user idea generation process.
1134-1144. Management Sci. 48(8): 1042-1059.
George JM, Zhou J (2001) When openness to experience and con Luchins AS (1942) Mechanization in problem solving: The effect of
scientiousness are related to creative behavior: An interactional Einstellung. Psych. Monographs 54(6):i—95.
approach. J. Appl. Psych. 86(3):513-524. Lyles MA, Mitroff II (1980) Organizational problem formulation: An
Ghemawat Levinthal D (2008) Choice interactions and business
P, empirical study. Admin. Sci. Quart. 25(1): 102—119.
strategy. Management Sci. 54(9): 1638-1651. MacDuffie JP (1997) The road to "root cause": Shop-floor problem
Glaser BG, Strauss AL (1967) The Discovery of Grounded The solving at three auto assembly plants. Management Sci. 43(4):
ory: Strategies for Qualitative Research (Aldine Publishing,
479-502.
Chicago). Makri S, Blandford A, Woods M, Sharpies S, Maxwell D (2014)
Greiner R (1996) PALO: A probabilistic hill-climbing algorithm. Arti "Making my own luck": Serendipity strategies and how to sup
ficial Intelligence 84( 1—2): 177—208. port them in digital information environments. J. Assoc. Inform.
Sci. Tech. 65(11 ):2179-2194.
Greiner R, Hayward R, Jankowska M, Molloy M (2006) Finding opti
mal satisficing strategies for and-or trees. Artificial Intelligence March JG, Simon HA (1958) Organizations (Wiley, New York).
170( 1): 19—58. Marengo L, Dosi G, Legrenzi P, Pasquali C (2000) The structure of
Gronbaek K (1989) Rapid prototyping with fourth generation systems: problem-solving knowledge and the structure of organizations.
Indust. Corporate Change 9(4):757-788.
An empirical study. Office: Tech. People 5(2): 105-125.
Hamilton LA (2003) The Political Philosophy of Needs (Cambridge Marples DL (1961) The decisions of engineering design. IRE Trans.
Engrg. Management 8(2):55—71.
University Press, Cambridge, UK).
Martinsen O (2011) The creative personality: A synthesis and devel
Hatchuel A, Weil B (2002) C-K Theory: Notions and applications
of a unified design theory. Proc. Herbert Simon Internal. Conf. opment of the creative person profile. Creativity Res. J. 23(3):
185-202.
Design Sci., Lyon, 15-16 March.
McCrae RR, Costa PT (1985) Comparison of EPI and psychoticism
Hsieh C, Nickerson JA, Zenger TR (2007) Opportunity discovery, scales with measures of the five-factor model of personality. Per
problem solving and a theory of the entrepreneurial firm. J. Man
sonality Individual Differences 6(5):587~597.
agement Stud. 44(7): 1255-1277.
Merton RK (1968) Social Theory and Social Structure (Free Press,
Hui SK. Fader PS, Bradlow ET (2009) Path data in marketing: An
New York).
integrative framework and prospectus for model building. Mar
Merton RK, Barber E (2004) The Travels and Adventures of Serendip
keting Sci. 28(2):320-335.
ity: A Study in Sociological Semantics and the Sociology of Sci
International Electrotechnical Commission (2006) Fault Tree Analysis ence (Princeton University Press, Princeton, NJ).
(FTA), International Standard IEC 61025, 2nd ed. (IEC, Geneva).
Mintzberg H, Raisinghani D, Theoret A (1976) The structure of
Ishikawa K (1968) Guide to Quality Control (JUSE, Tokyo). "unstructured" decision processes. Admin. Sci. Quart. 21(2):
Ishikawa K (1976) Guide to Quality Control (Asian Productivity 246-275.
Organization. Tokyo). Nelson RR (2008) Bounded rationality, cognitive maps, and trial and
Jeppesen LB, Lakhani KR (2010) Marginality and problem-solving error learning. J. Econom. Behav. Organ. 67( 1):78—89.

effectiveness in broadcast search. Organ. Sci. 21 (5): 1016—1033. Novick LR. Sherman SJ (2003) On the nature of insight solutions:
Koestler A (1964) The Act of Creation (Macmillan, New York). Evidence from skill differences in anagram solution. Quart. J.
Krumholz HM, Curry LA, Bradley EH (2011) Survival after acute Experiment. Psych. A 56(2):351-382.
myocardial infarction (SAMI) study: The design and imple Ohlsson S (1984) Restructuring revisited. Scandinavian J. Psych.
mentation of a positive deviance study. Amer. Heart J. 162(6): 25(2):65-78.
981-987. Padgett JF (1980) Managing garbage can hierarchies. Admin. Sci.
Kurup U, Bignoli PG, Scally JR, Cassimatis NL (2011) An architec Quart. 25(4):583-604.
tural framework for complex cognition. Cognitive Systems Res. Penin J, Burger-Helmchen T (2012) Crowdsourcing d'activites inven
12(3^):281-292. tives et frontieres des organisations. Management lnternat.
Lakhani KR, Lifshitz-Assaz H, Tushman ML (2013) Open innovation 16:101-112.
and organizational boundaries: The impact of task decomposition Pich M, Loch C, De Meyer A (2002) On uncertainty, ambigu
and knowledge distribution on the locus of innovation. Grandori ity, and complexity in project management. Management Sci.

A, ed. Handbook of Economic Organization: Integrating Eco 48(8): 1008-1023.

nomic and Organization Theory (Edward Elgar, Cheltenham, Raymond ES (1999) The Cathedral and the Bazaar: Musings
UK), 355-382. on Linux and Open Source by an Accidental Revolutionary
Laursen K, Salter A (2006) Open for innovation: The role of openness (O'Reilly Media Inc., Sebastopol, CA).
in explaining innovation performance among U.K. manufacturing Rittel HWJ. Webber MM (1973) Dilemmas in a general theory of
firms. Strategic Management J. 27(2): 131—150. planning. Policy Sci. 4(2): 155—169.
Le Masson P, Weil B, Hatchuel A (2010) Strategic Management of Sadow BD (1972) Rolling luggage. U.S. Patent 3,653,474, filed
Innovation and Design (Cambridge University Press, Cambridge, February 16, 1970, and issued April 4, 1972.
UK). Schon DA (1983) The Reflective Practitioner: How Professionals
Levinthal D, Posen HE (2007) Myopia of selection: Does organi Think in Action (Basic Books, New York).
zational adaptation limit the efficacy of population selection? Schwenk C, Thomas H (1983) Formulating the mess: The role of
Admin. Sci. Quart. 52(4):586-620. decision aids in problem formulation. Omega 11 (3):239—252.
Liang RH (2012) Designing for unexpected encounters with digital Sharkey J (2010) Reinventing the suitcase by adding the wheel.
products: Case studies of serendipity as felt experience. Internat. Bus. Day: New York Times (October 4), [Link]
J. Design 6(1):41—58. 2010/10/05/business/[Link].

This content downloaded from

[Link] on Wed, 28 Jan 2026 17:03:46 UTC
All use subject to [Link]
Hippel and Krogh: Identifying Viable "Need-Solution Pairs"
Organization Science 27(1), pp. 207-221, ©2016 INFORMS 221

Sieg JH (2012) Problems on the Boundary: Involving External Service Volkema RJ (1983) Problem formulation in planning and design.
Providers in Organizational Problem Formulation. Ph.D. disser Management Sci. 29(6):639-652.
tation, ETH No. 20285, ETH, Zurich. von Hippel E (1986) Lead users: A source of novel product concepts.
Simon HA (1973) The structure of ill-structured problems. Artificial Management Sci. 32(7):791—805.
Intelligence 4(3—4): 181 —201. von Hippel E (1994) "Sticky information" and the locus of prob
Simon HA (1977) Models of Discovery and Other Topics in the Meth lem solving: Implications for innovation. Management Sci.
ods of Science, Boston Studies in the Philosophy of Science 40(4):429-439.
(D. Reidel, Dordrecht, Netherlands). von Hippel E (2005) Democratizing Innovation (MIT Press,
Simon HA (1978) On how to decide what to do. Bell J. Econom. Cambridge, MA).
9(2):494-507. von Hippel E, Tyre MJ (1996a) The mechanics of learning by doing:
Simon HA (1979) Models of Thought (Yale University Press, New Problem discovery during process machine use. Technology and
Culture 37(2):312-329.
Haven, CT).
von Hippel E, Tyre MJ (1996b) How learning by doing is done:
Simon HA, Simon PA (1962) Trial and error search in solving dif
Problem identification in novel process equipment. Res. Policy
ficult problems: Evidence from the game of chess. Behav. Sci.
24(1): 1-12.
7(4):425^t29.
von Krogh G. Haefliger S, Spaeth S, Wallin M (2012) Carrots and
Smith RP, Eppinger SD (1997) Identifying controlling features of rainbows: Motivations and social practice in open source soft
engineering design iteration. Management Sci. 43(3):276-293.
ware development. MIS Quart. 36(2):649-676.
Spradlin W (2012) Are you solving the right problem? Harvard Bus. Wallin MW, von Krogh G, Sieg JH (2013) A "problem in the mak
Rev.90(9):85-93.
ing": How firms formulate problems for innovation contests.
Szulanski G (1996) Exploring internal stickiness: Impediments to the
Working paper, ETH Zurich, Zurich.
transfer of best practice within the firm. Strategic Management
Walpole H (1960) To Mann, Monday 28 January 1754. Lewis WS, ed.
J. 17(S2):27—43. The Yale Editions of Horace Walpole's Correspondence, With Sir
ter Hark M (2010) Psychology of thinking before the cognitive revo Horace Mann, IV, Vol. 20 (Yale University Press, New Haven,
lution: Otto Seltz on problems, schemas, and creativity. History CT), 407-411.
Psychology 13( 1):2—24. Wertheimer M (1959) Productive Thinking (Harper, New York).
Terwiesch C, Loch CH, De Meyer A (2002) Exchanging preliminary Zhao H, Seibert SE (2006) The big five personality dimensions and
information in concurrent engineering: Alternative coordination
entrepreneurial status: A meta-analytical review. J. Appl. Psych.
strategies. Organ. Sci. 13(4):402-419. 91 (2):259—271.
Thomke SH (1998) Managing experimentation in the design of new
products. Management Sci. 44(6):743-762. Eric von Hippel is the T. Wilson Professor in Manage
Thomke SH (2003) Experimentation Matters: Unlocking the Potential ment and a professor of management of innovation and engi
of New Technologies for Innovation (Harvard Business School neering systems at the MIT Sloan School of Management.
Press, Boston). His research interests include explorations of problem solv
Thomke SH, Fujimoto T (2000) The effect of "front-loading" ing and studies related to the sources of and economics of
problem-solving on product development performance. J. Prod innovation.
uct Innovation Management 17(2): 128-142. Georg von Krogh is full professor of strategic manage
Tyre MJ, von Hippel E (1997) The situated nature of adaptive learning ment and innovation at ETH Zurich. He is also a member of
in organizations. Organ. Sci. 8( 1):71—83. the National Research Council at the Swiss National Science
Van Andel P (1994) Anatomy of the unsought finding. Serendipity: Foundation. His research interests include competitive strat
Origin, history, domains, traditions, appearances, patterns and egy, technological innovation, and organizational knowledge
programmability. British J. Philos. Sci. 45(2):631-648. creation.

This work is licensed under a Creative Commons Attribution-NoDerivatives 4.0

0 o o International License. You are free to download this work and share with others,
but cannot change in any way or use commercially without permission, and you
BY ND must attribute this work as "Organization Science. Copyright 2016 INFORMS.
[Link] used under a Creative Commons Attribution
License: [Link]

This content downloaded from

[Link] on Wed, 28 Jan 2026 17:03:46 UTC
All use subject to [Link]