Recently, I was invited to give a talk at Troux Directions on the evolution of enterprise architecture (gulp). I think it went pretty well, judging from comments and some of the tweets. The centerpiece slide was this two-dimensional view:

(click on the graphic for a better view)

EA is not evolving in just one dimension. First, it’s evolving as an IT practice, as part of a broader maturation of IT management. Neither IT management nor EA has  full maturity even on strictly IT terms yet. EA is also evolving up what can be called the “business food chain.” As a practice, EA is moving from simpler system and data models into understanding broader structural aspects of the business.

While I’m not usually a fan of sequential maturity models, I stuck my neck out and proposed a progression along each axis. (I already see a few things I would change; this is a sketch.) A key point I wanted to make is regarding the constraints, the heavy dotted lines. In terms of IT management maturity (horizontal axis), EA is currently crossing the boundary from understanding IT services and portfolios at a structural level, to understanding IT performance. I’ve talked a bit about this here, as part of my interest in the IT data warehouse.

Along the vertical axis, EA has long had a concern for structural understanding of the business, i.e. business capabilities and value streams, but has drawn a line at dynamic understanding.

Some attendees were clearly focused on maturing EA within the bounds of IT – the horizontal axis. I think that is an entirely reasonable approach, as there is still plenty of work to do improving IT. Others are more toward the vertical axis, moving towards a Corporate Strategy kind of function. (And what if there already is a Corporate Strategy group? That’s one of the very unsettled questions here.)

But both dimensions present a challenge to EA to become LESS abstract. Capability mapping can be very valuable, but often what the business really cares about is specific performance – not about the abstractions “Marketing” or “Product,” but the issues and potentials of a SPECIFIC product in a SPECIFIC market. That’s where money is made. Similarly, you may understand your IT portfolio, but that does not mean you understand your IT performance.

And it’s in thinking about this problem that I get a little cautious. It’s a no-win situation. If EA sticks to its abstractions, it can be dismissed as ivory tower, just painting pretty pictures. But if EA starts investigating matters that Marketing or Sales think they OWN, EA may really start to ruffle feathers, and does EA have the tools to help? You start to get into core MBA-type work, and how many EAs have MBAs? Even in the question of IT performance, fully understanding it requires deep investigations into IT execution areas like Availability Management and IT Finance that many EA groups might prefer to not dirty their hands with. Or even be able to get data from.

If you’re starting from IT origins, I think it’s  inadvisable to undertake an EA maturation path that is too steep. Business colleagues may see you as the barefoot cobbler’s child – “they can’t even manage IT well, and they want to move into business capability mapping!” I think the best course is a middle one that balances a concern for IT maturation with increasing coverage of business architecture. This may be different for EA groups that actually are moved out of traditional IT, over to more business-aligned reporting (a trend I’m increasing increasing reports of). Even then, I think that a big part of EA’s value will always be in drawing that line of sight from the business to IT.

But I don’t think there’s any stopping progress here. To cross that ceiling, I proposed in my presentation that EA explore methods like systems dynamics and Monte Carlo analysis, and I know I’m not the first to do so. What was surprising however was hearing from at least one EA group at Troux Directions who had actually established a dynamic modeling capability, to address the same concerns. (It’s always gratifying to propose something in a presentation and find support. And, in case you were wondering, the tool they use is AnyLogic. No, I hadn’t heard of it either. It also does agent-based simulation.)

Now, I saw this as a five-ten year out discussion. There is still much work to be done in understanding the more stable aspects of the business. Capability mapping is an essential part of the journey, a necessary precondition to any kind of dynamic understanding. Products like Troux and alfabet are essential to building out these kinds of maps, understanding where business capabilities are strong and weak, core and peripheral, what IT application services support them, and what infrastructure services and base technical products in turn support the applications. You can’t evolve EA in any direction without that foundation.

But nothing stays still. Sooner or later, you’ll have that structural view. And so will your competitors. At that point, dynamic understanding may be a critical next step.

Enhanced by Zemanta