Amazed to find this post that suggests that Enterprise Architecture is the next big thing. I thought that EA was a late 90s, early 00s silver bullet and that we were now all moving on to a concentration on fractal business architectures supported by federated IT. Now I realise that judicious EA can encompass this view – indeed that it should – but realistically EA as a term was hijacked by geeks who used it to enforce technical standards and who gibbered at their business colleagues in a threatening way about governance and compliance whilst creating impenetrable mountains of useless documentation that was out of date as soon as it was created (and made no sense to business people or developers).
Worse than that most EA practitioners genuinely believe that they can document the whole organisation top down in a complex chain of dependencies – a horrible fallacy in a world of federation – and then micro-govern people by putting cumbersome processes in the way of getting anything done (and given the difficulty (nay – impossibility) of doing this they end up obsessing on forcing shared and inappropriate IT on people instead of supporting business improvement).
Taking it all a bit further it’s also unfortunate that most EA efforts top out at business processes (rather than capabilities) and so are a pure expression of how things get done rather than a sensible framework for enabling decisions about what should be done (I enjoyed Steve Jones’s post about this very issue).
Before I get flamed to the extent that I am consumed by a terrible conflagration I have to say a couple of other things; are the aims of EA crazy (i.e. to understand and govern the organisation)? Umm, no. Should we still be seeking to do this? Umm, yes. Does the kind of top down, all encompassing approach taken by most EA efforts actually work? Umm, no. By saying that EA is no longer useful I’m basically saying that the concept is absolutely right, it is more necessary than ever in a world of services – indeed services could be the missing abstraction that finally fill the gap between intent and execution – but that initial efforts in this direction didn’t take sufficient notice (indeed why would they have) of unbundling and federation. So I believe that EA needs to evolve to help govern the portfolio of capabilities required by an organisation to function and to include some light touch policies that specify how they should work together. Below this level we start again, essentially with a smaller EA for each capability, recognising federation and unbundling will remove our ability to control every aspect of the way that an organisation works centrally and from top to bottom – EA essentially needs to become explicitly fractal.
Either way I’m not sure that EA is the next big thing, but then again maybe I’ve just dreamt all of this and the future is bright etc.