Together with Professor Jan Pries-Heje I have just had an article accepted for the 42th Hawaii International Conference on System Sciences in January 2009.
To answer our research question on the purpose and drivers of EA, we used two research approaches: In the first part we gathered two focus groups, one with CIOs and one with chief enterprise architects in the Danish central government. In the second part we interviewed two experts – or “Gurus” – and analyzed official documents, newspapers and official websites.
We found the causal direction reversed from EA being transformative and prescriptive in its nature to EA being reshaped and adopted in step with the institutional forces in public organizations and their macro environment. To be more than just another fashion fad, future EA programs in government must provide a comprehensive and coherent view across business, information, and technology; not just to guide the design of IT systems – but to deliver business change supported and enabled by IT.
Abstract:
Enterprise Architecture (EA) has been promoted as a key tool for transformation and modernization of government. In this paper we study what has driven the use and adoption of the EA concept in the Danish central government. Based on analysis of focus group and ‘guru’ interviews with government CIO’s, enterprise architects, and consultants, as well as extensive document studies, we find that there are two streams in public sector EA programs: a stable element of it-architecture and a fashion driven business architecture element – used in parallel, but with different focus, approach and artifacts. We conclude that EA in government to a large extend is driven by fashion. Finally, we discuss the role of EA in the future and point out that EA can not transform government by itself. Fundamental transformation to the tasks performed in public organizations is only achieved if institutional forces promote transformation.
Let me know if you want further details.
Kristian,
It is my belief that we need to rethink Enterprise Architecture from being something that is focused on understanding the Enterprise to something that is focused on delivering value to a specific project. To some extent, this calls into question the name, “Enterprise Architecture”, but I still think the name has value because it implies that we are considering project level issues that can only be addressed at the enterprise level.
To go even further, the main value that “Enterprise Architecture” can deliver to projects is in reducing their overall complexity. One might argue that we can also focus on improving business/IT alignment, but I believe that this flows naturally once we have reduced project complexity.
Why do we need to focus on complexity at the Enterprise Level? Why not the IT level or the business level? The answer to this requires an overview of the Science of Simplicity. This is a bit much to cover here, but let me refer you to my last book, Simple Architectures for Complex Enterprises and numerous white papers at my web site. In brief, simplification requires a good understanding of both synergistic relationships (which are best understood by the business side of the enterprise) and strong partition boundaries (which are best understood by IT side of the enterprise). We can thus only address these issues at the juncture point between business and IT, and that juncture point is what we usually call “Enterprise Architecture”.
I think that government is perhaps the area that can most benefit from a focus on simplification. Government projects tend to be highly complex and often involve multiple vendors. When complexity is not managed methodically and early on, failure is too often the result. An excellent example of the problems inherent in such failure is the UK NHS system known as NPfIT (National Programme for Information Technology). This sytem will likely cost 50-100 billion dollars and most likely will end up in failure. Why? Essentially, catestrophic complexity.
Why is complexity such a problem? Several studies (and numerous personal observations) have shown that increasing the functionality of a system by 25% increases the complexity of the system by 100%, unless steps are taken to manage the complexity. This means that taking a system from 10 functions to 100 functions increases the complexity by over 1000 times! Clearly this is not acceptable. The only way we can dampen this increase is by focusing on the problem of complexity. And Enterprise Architecture is where we have the opportunity to do so.
So let me summarize my thoughts on where Enterprise Architecture needs to go in the near term, especially in the public sector. First, it needs to move from an enterprise focus to a project focus, however it needs to focus on these projects from an enterprise perspective. Second, it needs to specialize in the one problem that can only be solved from an enterprise perspective: complexity.
Best wishes,
Roger Sessions
Pingback: Architecting for Better Government - Kristian Hjort-Madsen