The design-science perspective draws upon Simon’s (1969) notion of
a ‘science of the artificial’ and his idea that science develops knowledge about what already is, whereas design uses knowledge to create
what should be, things that do not yet exist (Romme, 2003: 562). An
orientation towards design distinguished the professions, such as engineering, architecture, education, law and medicine, from the sciences
(Romme, 2003: 558). As such, it seems to fit the fields of management
and organization studies (Argyris, 1996; Avenier, 2010; Bevan, Robert,
Bate, Maher, & Wells, 2007; Denyer, Tranfield & van Aken, 2008; Jelinek, Romme & Boland, 2008; Trullen & Bartunek, 2007; van Aken,
2005;), since, like engineering or architecture, the focus of management research ‘is not a natural phenomenon but something humanmade’ (Hodgkinson & Rousseau, 2009: 536).
In the context of design sciences, management research aims at producing instrumental and prescriptive knowledge which is put into action
in collaboration with practitioners in order to solve one of their problems
(Denyer, et al., 2008: 395). Although the design-science perspective in
management comprises several distinct approaches, a core element in
these approaches is the conviction that, in management, ‘more room
for the development of solution-oriented or prescriptive knowledge
would increase its relevance’ (Denyer, et al., 2008: 393). Design rules
could be the way to bridge the relevance gap in management research
(Romme, 2003: 567).
Instrumental and prescriptive knowledge: design propositions or rules
Drawing on the idea that management and organization science has
reached a certain level of maturity and should now venture ‘out from its
adolescence’ (Romme & Endenburg, 2006: 287), the design-science
perspective promotes a pragmatic view of knowledge which consists
in designing systems ‘that do not yet exist – that is, change existing organizational systems and situations into desired ones’ (Romme, 2003:
559). Since managing means ‘creating intended consequences’ (Argyris, 1996: 402), management research should produce designs, understood as ‘specifications of actions to be taken to achieve the intended
consequences’ (Argyris, 1996: 396).
Proponents of the design-science approach to management have used
a variety of terms for referring to the type of instrumental and prescriptive knowledge it produces: designs, design propositions, design rules,
design solutions, technological rules, and so on. Design rules refer to
‘any coherent set of detailed guidelines for designing and developing
organizations’ (Romme & Damen, 2007: 110). Likewise, ‘technological
rules’ refer to ‘knowledge that can be used by professionals in the field
[…] to design solutions to their field problems’ (van Aken, 2005: 22).
A key point about these technological rules or design propositions is
that they never are the complete solution for a particular problem; rather, they are ‘an input to the designing of the specific solution’ (Denyer,
et al., 2008: 396). The specific solution itself demands ‘professional
knowledge and expertise […] along with the evidence from fieldtesting
and intimate knowledge of the local situation and business domain in
question’ (Denyer, et al., 2008).
While agreeing on the end result – design propositions – proponents
of the design-science approach significantly differ as to the ways to
produce these design propositions. More specifically, what varies from
one approach to the other is the degree to which design propositions,
and management research in general, are coproduced by scholars and
Design propositions extracted from research synthesis
A second conception of design science stresses that research results
produced by descriptive research can often be translated into design
propositions and rules (van Aken, 2005: 29). Rather than embarking
on new empirical studies with the explicit aim of producing design rules
about a particular field problem, a lot of progress in management and
organization science can be made by revisiting existing research results
and evidence and extracting the design propositions that are implicitly
contained in them (Hodgkinson & Healey, 2008: 451). This perspective
is akin to the evidence-based approach in management which seeks
to ‘distil actionable principles from systematic reviews of prior studies’
(Hodgkinson & Healey, 2008: 437).
Denyer, Tranfield & van Aken, for example, argue that ‘the development of design propositions can result from synthesizing previously published research’ (2008: 393) and illustrate this position with research
about high-reliability organizations. Design propositions follow what
they call the ‘CIMO-logic’: a problematic context (C), calls for a design
proposition that suggests an intervention type (I) based on generative
mechanisms (M), in order to reach an intended outcome (O) (Denyer,
et al., 2008: 393).
They give the following example of a design proposition following the
CIMO logic: ‘If you have a project assignment for a geographically
distributed team (class of contexts), use a face-to-face kick-off meeting (intervention type) to create an effective team (intended outcome)
through the creation of collective task insight and commitment (generative mechanisms)’ (Denyer, et al., 2008: 396). A design proposition
they derive from their synthesis of the literature about high-reliability
organizations runs as follows: ‘In contexts characterized by social and
political pressure, interactive complexity and high hazard, in order to
avoid high-impact failure and reduce error rates, continuously communicate rich, real-time information about the health of the system and
any anomalies or incidents (Denyer, et al., 2008: 406).
In this perspective, the design mode serves to ‘translate empirical findings into design propositions’ (Romme, 2003: 569) and thus occurs
after some empirical findings have been produced by researchers
through a traditional – not necessarily collaborative - research process.
Hodgkinson and Healey (2008) suggest that this process of translation can occur with research in other fields in order to produce design
rules about management or organization problems. This strategy can
be effective when only limited evidence or research results about a
particular problem are available in the management field. For example,
arguing that research results and evidence about scenario planning
are lacking for design purposes, Hodgkinson and Healey (2008) draw
upon three theories from social psychology to extract design propositions that could inform the design scenario planning about team composition and the facilitation process.
An example of such design propositions extracted from basic research
is as follows: ‘To increase the likelihood of attaining requisite forms
of group information processing with informationally diverse scenario
teams, wherever possible select participants with greater intrapersonal
functional diversity’ (Hodgkinson & Healey, 2008: 442).
Of course, in this approach to design science, extracting design propositions from existing research is a first step that needs to be followed by
one consisting in subjecting them to field-testing, ‘in order to ascertain
what works and what does not work’ (Hodgkinson & Healey, 2008:
Coproduction of design rules by scholars and practitioners
Finally, a third approach in design science consists in the coproduction
of design rules by scholars and practitioners. This process of coproduction implies intense forms of collaboration between researchers
and practitioners. Principles of design science involve a particular way
of conducting research and producing knowledge that goes beyond
‘old’ research paradigms – such as action research. Collaboration with
practitioners is based on the idea that the use of design rules is a natural activity which is routinely performed by management practitioners,
albeit implicitly (Plsek, Bibby, & Whitby, 2007: 154).
This perspective in design science has been developed chiefly in the
sub-field of organization development (OD) (Mohrman, 2007) with the
aim of facilitating ‘empowerment and participation in decision making
at all levels’ (Romme & Damen, 2007: 115). With respect to this ‘core
OD value’ (van Aken, 2007: 81), the design-science approach enables
us to give ‘the direct stakeholders an active role in the change process’
(van Aken, 2007: 81).
In OD, Endenburg’s ‘circular design approach’ is one of the first concrete
applications of the design perspective. In order to explore new ways of
facilitating employee participation, Endenburg ‘started to develop the
circular OD approach in which feedback rather than power was to become the basic organizing principle’ (Romme & Damen, 2007: 111).
This approach to organization design is ‘very similar to the research
and development cycle connecting the natural sciences, engineering,
and technology’ (Romme & Endenburg, 2006: 295) and ‘draws on a
research cycle involving organization science, construction principles,
design rules, organization design, and implementation and experimentation (Romme & Damen, 2007: 287).
Construction principles and design rules are seen as ‘boundary objects’
which ‘can serve as a conceptual framework for productive interaction
and collaboration between practitioners, consultants, and academics’
(Romme & Endenburg, 2006: 295). Examples of construction principles
(the antecedents of design rules) for any organization that wishes to
‘build capacity for the circular flow of power and information’ include
such principles as ‘make mistakes’, ‘continually explore and set boundaries but recognize that deviating too much from your course is risky’, or ‘set and agree on acceptable limits in the case of collaboration’
(Romme & Damen, 2007: 111-112; Romme & Endenburg, 2006). Corresponding design rules anchored in such principles are, for example,
‘each circle makes decisions on policy issues by informed consent’
and ‘every member of the organization belongs to at least one circle’
(Romme & Endenburg, 2006). Construction principles are used as’ tools to create a specific set of design rules, acknowledging that there is
an infinite number of possibilities and combinations’ (Romme & Damen,
Romme & Damen (2007: 110) provide another illustration of the relationship between construction principles and design rules: an example
of the former would be ‘to increase innovative capabilities, the firm
needs to develop absorptive capacity—the ability to recognize the
value of new, external information, assimilate it, and apply it to commercial ends’, while the corresponding design rules grounded in such
principles would provide ‘guidelines regarding when and how to invest
in R&D, engage in cooperative R&D ventures, and so forth’ (Romme &
Damen, 2007: 110).
Since its inception in the 70s, circular design ‘has been applied in about
65 OD projects in the Netherlands, the Unites States, Canada, and Brazil’ (Romme & Damen, 2007: 115). Romme and Endenburg report several attempts from a diversity of companies at using the circular design
approach. In one of these cases, the attempt failed ‘as a result of the
“hit and run” strategy adopted by the CEO as well as his strong need
to be in control’ (Romme & Edenburg, 2006: 292). In another case, the
approach resulted in substantial improvements for the company and
showed that ‘participative decision-making processes generated and
bounded by the circular structure can be effective in a crisis situation’
(Romme & Edenburg, 2006: 292).
A different yet complementary view about scholar-practitioner collaboration in design science is the one developed by Plsek, et al. (2007). In
their perspective a crucial stage of design-oriented research consists
in trying to extract explicit design rules that are enmeshed in the experience of practitioners. Any form of organizational action and any effort
at organizational change have ‘embedded design rules’ (Plsek, et al.,
2007: 154). Management scholars, in collaboration with practitioners,
can convert practitioners’ tacit knowledge into explicit design rules that
are statements in the form ‘If you want to achieve outcome Y in situation S, something like X might help’ (Plsek, et al., 2007: 153).
Bate and Robert (2007) have pushed this approach further by focusing
on a different type of ‘practitioners’. Denouncing the strong management orientation of OD, they plead for a more ‘user-centric’ OD, ‘one
that seeks to mobilize and privilege change on behalf of the consumers or users of an organization’s product or service, involving them at
every stage of the design process, from problem diagnosis to solution
generation and implementation’ (Bate & Robert, 2007: 41). They have
applied the approach of experience-based design in a cancer clinic
with the aim of ‘improving the care and treatment experience of head
and neck cancer patients and their carers’ (Bate & Robert, 2007: 42).
In this process of codesign, the patients worked with staff, senior managers and physicians. They used one of the methods described by Plsek, et al., (2007) for extracting design rules, namely stories and narratives. For example, one of the design rules that emerged from patients
and staff’ stories about the fact that patients do not always know what
the next stage in their treatment is was ‘Never do anything that might
take away from the resilience of the patient’, a rule that is markedly different from the more immediate response of ‘Let’s tell them everything’
(Bate & Robert, 2007: 58).
This last example is particularly interesting since it demonstrates both
the potential and the limits of design rules for dealing with human
processes. In the authors’ own words: ‘The two possible alternative
rules in this case of “tell them everything” (based on the logic that we
should never deceive patients, surprise them, or keep them in the dark
about their own illness) and “preserve maximum patient resilience” are
tramlines, boundaries, or polarities that need to be managed and within
which difficult judgments and decisions will always need to be made’
(Bate & Robert, 2007: 58). In their view, it shows the fundamental difference between designing physical objects and designing human processes (2007: 58).
van Burg, Romme, Reymen, and Gilsing (2008) have developed
a ‘science-based design approach’ which combines the second approach explored above – extracting design propositions from research
synthesis – with the idea of extracting design rules from practitioners’
knowledge-in-action. They aim more specifically to connect pragmatic
knowledge about how to create university spin-offs to scholarly work
explaining why certain practices in this field work and others do not (van
Burg, et al., 2008: 116). They produced a synthesis between practice-based principles and research-based principles about the performance
of a particular spin-off.
First, following Plsek, et al.’s path, practice-based principles were developed by ‘converting the largely tacit knowledge of key agents in university spin-off creation into explicit principles’ (van Burg, et al., 2008:
118). One of these practice-based principles was, for example: ‘Make
potential entrepreneurs (e.g., students, Ph. D. students, staff members)
aware of opportunities to start a venture based on a research finding
(van Burg, et al., 2008: 121).
Second, research-based principles were derived from the scholarly literature about university spin-offs. One example of a research-based
principle was: ‘Screen technologies and ideas for new ventures, and
subsequently provide start-ups with advice and coaching from skilled
people’ (van Burg, et al., 2008: 121).
Finally, the authors set out to synthesize the two sets of principles in
a new set of ‘design principles’, defined as ‘principles that are tested
in practice as well as grounded in the existing body of research’ (van
Burg, et al., 2008: 121). One of these design principles runs as follows:
‘Create university-wide awareness of entrepreneurship opportunities,
stimulate the development of entrepreneurial ideas, and subsequently
screen entrepreneurs and ideas by programs targeted at students and
academic staff’ (van Burg, et al., 2008: 123).
Can design science bridge the relevance gap?
Although design science and the various design-oriented approaches
outlined above have opened a promising path towards bridging the relevance gap, a few mitigating points should be raised. First, contrary
to the assumption that technology-products can easily be deemed relevant and have a direct impact on practice (Hodgkinson & Rousseau,
2009: 541), a glance at other fields that are considered as exemplars of
design science shows that issues of ‘relevance’ are very much at stake
there too. In medicine, for example, ‘the transfer of research findings
into practice is often a slow and haphazard process. For example, ‘patients are denied treatment of proven benefit because the time it takes
for research to become incorporated into practice is unacceptably long’
(Graham, et al., 2006: 13).
With every illustration that the impact of hard evidence and design rules
on practice is direct and pervasive – for example, evidence about the
effects of feedback on performance has been translated into ‘contemporary guidance regarding how to give employees performance feedback’ (Hodgkinson & Rousseau, 2009: 540) – there are dozens that
probably show the opposite.
Another issue with the design-science perspective has to do with the
stage of extracting the implicit design rules embedded in practice. Acknowledging that practitioners routinely develop local theories (Bartunek & Louis, 1996: 5) and design rules which are embedded in their
actions is essential. Attempting to capture this practical knowledge and
to extract the design rules through scholar-practitioner collaboration is,
however, an arduous and uncertain process.
Research findings about ‘good practice’ such as Schön’s (1983) notion of ‘reflective practice’ indicate neither that good practice implies
making the tacit explicit nor that all tacit knowledge can be made explicit. On the contrary, Schön quite convincingly showed that ‘when
the professional practitioner tries, on rare occasions, to say what he
knows – when he tries to put his knowing into the form of knowledge
– his formulations of principles, theories, maxims, and rules of thumb
are often incongruent with the understanding and know-how implicit in
his pattern of practice’ (Flyvbjerg, 2001: 20). The process of codifying
tacit knowledge needs to be understood as a knowledge creation process rather than a mere process of ‘conversion’ of tacit knowledge into
explicit knowledge claims (Cohendet & Meyer-Krahmer, 2001: 1564).
Having tested four methods ‘for extracting tacit knowledge design
rules’ from experienced practitioners, Plsek, et al. conclude that there
are many complexities associated with doing this and that it entails an
intense process of interactive collaboration on the part of scholars and
practitioners (Plsek, et al., 2007: 168); what Bartunek and Louis had
already shown regarding outsider/insider research (1996: 18). We thus
end up with the same kind of limit that we outlined earlier, namely that
given the high ‘costs’ of such practitioner/scholar collaboration, it is
rather unlikely that it should proliferate in the near future, which casts
doubt on the ability of the design-science approach to ‘bridge’ the relevance gap.
The most important issue raised by the popularity of the design-science approach in management is the risk of taking it too literally
rather than as a metaphor or analogy (van Aken, 2004: 239). This very
literal conception of management ‘engineers’ is patent, for example,
in the following extract from Hodgkinson and Healey’s work presented above. They reach the following conclusion about the usefulness
of their design propositions on team composition and the facilitation
process: ‘Facilitators will be better equipped to engineer the requisite
forms of group cognitive processes that yield changes in decision makers’ mental models and enhance the flexibility of their thinking about
the future, by introducing various techniques as and when appropriate’
(Hodgkinson & Healey, 2008: 449).
As many of its proponents have stressed, the comparison between
designing objects and designing human processes can only go so far
(van Aken, 2007: 72). In management, which is more ‘context-bound’
than disciplines such as medicine or engineering (van Aken, 2004:
239), ‘using’ or ‘applying’ a design rule always involves a ‘comprehensive learning process rather than the straightforward execution of a
single rule’ (Denyer, et al., 2008: 396).
Proponents of the design-science approach are generally careful to
avoid a mechanistic view and thus end up producing very general
design rules such as ‘look for external initiatives that might provide
a stimulus for change’ and ‘spot and deal with resistance’ (Plsek, et
al., 2007: 163) that have little efficacy if practitioners do not have an
intimate knowledge of the context, since ‘each situation is unique’
(Romme, 2003: 559), and a certain level of expertise, since design
rules about human processes can never be simply ‘executed’ (Denyer,
et al., 2008: 396).
Finally, the limit of the design-science metaphor also lies in the types of
knowledge that management scholars produce. Although focusing on
instrumental and prescriptive knowledge is certainly useful, we should
not lose sight of the fact that other types of knowledge do have a role to
play in scholarship and in the management field especially.