Centralized, Federated, Or Virtualized: What Works For Organizing The EA Practice?

If you’re trying to build an effective EA program, you’re in trouble from the get-go. I’d like to paint a rosier picture for anyone involved in this strategic, potentially very high-impact practice, but consider the fact that one of our more frequent client inquiries is about how to communicate EA’s value to non-EAers. How can I not say you’re in trouble if so many people doing EA look for outside help to explain to their own stakeholders that what they’re doing on a daily basis is worthwhile? There’s clearly something wrong with this picture.

So, OK, let’s say you want to build an EA practice anyway, despite the poorly understood value proposition — who should you staff it with? Misguided people with a desire to labor away in obscurity? Actually, no, you want your best and brightest. Those few very smart people who know your business very well, have both deep and broad knowledge and great analytical skills, and who display the potential for strategic-, system-, and design-thinking. That's a little challenging.

And then, when you find these people and attract them to your program, how best to organize them for effectiveness? Centralizing EA resources gives you the most control and makes it more likely that EA can deliver on its strategic value proposition. Decentralizing or federating EA resources puts the architects where the action is, making it more likely business and BT stakeholders will perceive value from the effort. But then those federated resources sometimes get so involved with their local — and usually tactical — issues that they go native and they’re not really working on the “E” in EA anymore.

In theory, it doesn’t matter how EA resources are organized. Centralized resources can work on common issues, distributed resources can work on local issues, and all architecture resources can matrix into a virtual organization that also works on the strategic enterprisewide mission of EA. But if theory were the same as practice, we would have learned everything we need to know about how the world works from high school physics.

At Forrester’s upcoming Enterpise Architecture Forum 2012 in Las Vegas in May and in Paris in June, we’re rolling up our sleeves to tackle this last issue. No, not high school physics, but how best to organize EA resources. Tim DeGennaro and I are focusing on this for one of our EA Forum “super-sessions” — longer sessions designed to enable drilling down into details with a combination of research delivery and highly interactive exchanges with attendees. We’ll look at the latest research on how organizations do it today, and discuss questions such as: What are the positive and negative attributes of the different organizational approaches? Are the rules changing? How can we best position EA for current and future needs? Will EA be moving into the business organization in the coming years? How best to have an effective EA practice — with all EA resources living in the same box in the org chart or by establishing formal linkage to a network of resources?

Join us in Las Vegas or Paris to explore this topic by learning from the latest research and your peers’ successes and mistakes — and perhaps share your experiences with us as well. Our goal for the session is for everyone to walk away with a clear idea of what has worked — and why it worked — for others, and what will work for you.

Meanwhile, those who feel strongly about one organizational approach can reply here in our EA Community and tell us: What approach to the EA organization works for you? Do you think that will that work for everyone, or has it been tailored for your enterprise?