Enterprise Architecture as Applied Design Science
Every EA framework needs to be adapted and tailored to the local enterprise context
Enterprise Architecture is applied Design Science – a Science of the Artificial applied in the domain of management, organisation and business.
In 1969, Herbert Simon published the first edition of “The Sciences of the Artificial”. In this book, he made the distinction between :-
- the ‘Natural Sciences’ – sciences which study natural systems – i.e. systems of Nature not created by humans or and other intelligence; and
- the ‘Sciences of the Artificial’ – i.e. a scientific approach to the study of systems created by people.
In 2014, some 45 years after Herbert Simon published his book, looking back at the closing years of the 20th Century and the early years of the 21st, Raghuraman Krishnamurthy wrote :
“With increasing usage possibilities, and the attendant complexity, the need for a scientific understanding of how IT systems behave in an enterprise rose significantly.”
He then went on to observe that, “… EA [Enterprise Architecture] should have the focus at [the] enterprise level much beyond the contours of IT”.
Why should EA not limit itself to IT or even focus on IT? Because Information is not an IT issue.
Krishnamurthy is right, EA should not limit itself to the small subset of technology called “Information Technologies” – and nor should it limit itself to just the generalised technological, but should concern itself with all artificial systems in an enterprise – which is to say all systems in an enterprise since an enterprise is itself an artificial structure or construct created by people.
Why should EA not limit itself to IT or even focus on IT? Because Information is not an IT issue – and, more importantly, informed, coordinated activity for a purpose or towards a common set of objectives – which is the very definition of an enterprise – is not an IT issue.
IT is only a set of technologies that sometimes puts some information in the metaphorical hands of people who need to make decisions about what to do. But making decisions is not the only, and often, not the most important of things that people do (with or without the help of IT or other technologies).
And all too often, all that IT does, is hurl data at people in the hope they might metaphorically catch the important, clear, crystalline stones hidden in the information hailstorm.
So, EA should focus on the complex, composite structure and system of artificial systems that is the enterprise – it should be what the noun-phrase says it is: about the architecture of enterprises. In any case, to focus on the small fraction of artificial systems in an enterprise that are IT-intensive systems, is contrary to the holistic approach of EA and a disproportionate allocation of analytical, scientific effort.
Having distinguished the systems of Nature – natural systems – from the systems of human construction – artificial systems – Simon goes on to observe that the methods of natural science, collectively summed up as “analysis”, need expansion with some of the methods of construction – “synthesis” – if they are to apply to artificial systems.
Simon summarises that “science” is concerned with “analysis” (of natural systems) while “engineering” is concerned with “synthesis” (of artificial systems). In Simon’s idiom, “Engineering” is the usage of knowledge to change the world through the conscious, deliberate construction of artificial systems.
In the UK, we have a cultural, intellectual hobble or shackle that (mistakenly) conceives “engineering” as more- or less- sophisticated forms of metal-bashing or assembly. This is nothing more than powerful but unfounded cultural/ intellectual snobbery – and a profound societal mistake; engineering at Loughborough is a much more challenging and culturally or intellectually respectable pursuit than PPE at Oxford (for example) – Herbert Simon has it right!.
Simon goes on to develop his thesis into a plea for a new form of systematic analysis/synthesis – different from traditional Science, and different from heuristic, personal, subjective Craft and broader, more holistic than the systematic narrow-scope methods of construction that are the engineering subjects – mechanical, electrical, aeronautical, chemical, naval, electronic, genetic, etc. – a “Science of the Artificial.
IT is only a set of technologies that sometimes puts some information in the metaphorical hands of people who need to make decisions about what to do.
Simon develops his thesis and identifies a systematic approach to “Design” as a core component of the Science of the Artificial. Chapter 5 of his book – entitled “The Science of Design: Creating the Artificial” – outlines what a science of design looks like, including :
- a theory of design;
- a methodology of how designs are arrived at;
- a discussion of how designs (and artificial systems) are represented;
- Means-Ends analysis; and
- the logic of (artificial systems-of-systems) design. A lot of this looks very familiar to the Enterprise Architect.
However, Simon was doing his thinking and writing, 45 years ago, at a time when the positivist-inductivist theory of Science was very dominant. This notion, that Science infers and explains regularities in systems’ behaviours, from the patterns identified in ‘natural’ but careful observations with unaided human faculties – is both naïve and outdated, an 18th Century view of Science.
In the last half century, the philosophy of science has made huge revising steps forward – and a modern view of Science is that it starts with the putative theory of the way the Universe is and builds technological systems to create (and then observe) the phenomena that theory entails. The observation also often requires complex technological systems to be constructed just in order to ‘see’ the phenomena.
So, the science of the natural and the science of the artificial, are not so very far apart – not as far as Simon imagined – and modern natural science is much closer to the science of artificial systems than it is to positivist-inductivist dogma and preconceptions.
“Design” is important and central to both – design of experimental systems to allow scientists to develop knowledge on the one hand, and design of other artificial systems to allow engineers to deliver other benefits to other groups of people on the other – in the mostly-artificial world humanity has constructed for itself.
The Science of the Artificial continued to develop, slowly, through the 1970s, 1980s and 1990s but took a significant step forward about a decade ago with the publication of papers around “Design Science” by Hevner, March, Park, Gregor, Ram and others (all of which naturally point back to Herbert Simon’s contribution).
The seminal paper by Hevner, March, Park and Ram identifies “Design Science”, conjoined and contrasted with “Behavioural Science”, as the twin foundational paradigms of the “Information Systems” Discipline.
Hevner et al., cite the Henderson and Venkatraman strategic alignment model (an adapted and generalised form reproduced here) and observe that the derivation from business strategy to organisational infrastructure takes place within the Behavioural Science paradigm, while the derivation from technology strategy (or in their case IT Strategy) to technological infrastructure (or again, in their case, IT Infrastructure) takes place in the Design Science paradigm – and the two need to be conjoined if an aligned (coherent) technology-organisation enterprise is going to be achieved. It is not too far of a ‘stretch’ to think that behavioural science, applied in the context of enterprises, is synonymous with sociological science.
Hevner, et al., develop their thinking into a 3-cycle view of “systems research” (in their case, applied to “Information Systems” in enterprises), depicted in a slightly adapted form below.
The “Relevance Cycle” is about applying relevant knowledge to engineer the right systems in enterprises according to “business needs”.
The “Rigour Cycle” is about developing rigorous (in the academic sense), empirically-grounded theory and knowledge (of how social and technological systems in enterprise change and evolve.
Linking these two cycles is the third “Design Cycle”. This cycle is based on model artefacts and serves two different-directional purposes:
1) the adapting of general theories and reference models of technological and sociological systems to the specific context of a particular enterprise according to its business needs; and
2) abstracting general models and theories of sociological and technological change from observations of particular instances in a specific enterprise’s context.
The first result and principle that Hevner et al. extract from their theorising is that of “Design as an Artefact” – “a purposeful [IT] artefact created to address an important organisational problem. It must be described effectively, enabling its implementation and application in an appropriate domain.”
If the “organisational problem” is considered to be that of adapting to a changing business and technological environment, at the whole enterprise level, then the “purposeful artefact” will be complex – comprising many separate parts and interacting factors – and be “described effectively” by a collection of models of similar complexity – ie a set of Enterprise Architecture products / descriptions.
Hevner et al., position their three-cycle as a “Design Science Research” framework (or paradigm). I would speculate that this is because they are academic-scientists (of sorts) whose preoccupation is research – presumably in the context of some university somewhere – to extract general theory from a systematic series of ‘observations’ of particular enterprises.
This is a left-to-right direction of travel in their framework. From a practitioner-engineer perspective however, the preoccupation is delivering the correct (coupled) evolution of organisation and technology, and the direction of travel is right-to-left.
However, whereas most EA frameworks have little in the way of theoretical justification or empirical grounding, on which to base their claims of effectiveness, the Hevner-Mark-Park-Ram three cycle framework is both empirically grounded, theoretically sound and academically rigorous.
Thus I arrive at the conclusion: Enterprise Architecture is applied Design Science – a Science of the Artificial applied in the domain of management, organisation and business.
Every EA framework, as the basis of actual practice, has to be adapted, tailored to the local enterprise context – and so, is unique to the enterprise. Is the one used in your enterprise as theoretically sound and empirically justified as Applied Design Science?
References Simon, H.A., (1996), “The Sciences of the Artificial”, (3rd Ed.), The MIT Press.  Krishnamurthy, R., (2014), “Architecture Leadership and Systems Thinking”, (chapter 5) in Saha, P., (2014), “A Systemic Perspective to Managing Complexity with Enterprise Architecture, IGI Global.  Hevner, A.R., March, S.T., Park, J., and Ram, S., (2004), “Design Science in Information Systems Research”, MIS Quarterly, Vol. 28, No. 1, pp. 75-105. Available as a pdf.  Hevner, A.R., (2007), “A Three Cycle View of Design Science Research”, Scandinavian journal of Information Systems, Vol. 19, No. 2, pp. 87-92. Available as a pdf.  Henderson, J. and Venkatraman, N., (1993), “Strategic Alignment: Leveraging Information Technology for Transforming Organizations”, IBM Systems Journal, Vol. 32, No. 1.
If you think this post is interesting, please help spread the word – share this!
Headline image courtesy Sergio Rola@unsplash.com.