Tuesday, December 15, 2009

EA maturity models

Following the popularity of SOA maturity models, a number of EA maturity models have started to appear.

However, if the purpose of this kind of exercise is to assess and improve the effectiveness of an Enterprise Architecture program, then I think it's more useful to think of these models as Excellence models rather than Readiness or Maturity models, and to take inspiration not just from the SEI CMMI but also from excellence models such as Baldrige and EFQM.

There are two important points about these excellence models. Firstly, they don't only look at the capabilities but also at the outcomes produced by these capabilities. And secondly, they provide a systematic framework for developing a customized capability model for a particular enterprise. We shouldn't expect people to invest in specific EA capabilities just because some EA guru thinks it's a good idea, or just because lots of other organizations have adopted this as a "best practice", but ultimately because this capability can be demonstrated to produce the desired effects in this particular enterprise. Whereas an immature organization may have to take some of this kind of thing on trust, a mature organization should be striving for EA excellence, which means that every EA activity is tied to results.

Incidentally, when I have talked to people about Adoption and Excellence models in the context of SOA (in my work with the CBDI Forum), I have perceived a little resistance to the word "excellence". Some people seem to interpret it as meaning quality for its own sake. But in Baldrige and EFQM, excellence means value-for-money, doing exactly those things that contribute to the short-term and longer-term goals of the enterprise and its customers. So perhaps we could call it a Cost-Effectiveness model instead.

So I think any EA assessment should include three vital questions. What does EA cost - not just the architects' own time but also the time of developers, users and other stakeholders in participating in EA activities and complying with EA edicts? Secondly, what added-value is EA producing for the enterprise? And thirdly, what is EA doing to monitor and control the answers to these questions?

To get good answers to these questions, we clearly shouldn't just ask the architects; interviews should involve developers, users and other stakeholders: thus we end up with a 360-degree assessment.


Jon H Ayre said...

Interesting post with a different slant on maturity models. I would be concerned, however about the question "What does it cost" on its own, as it implies that EA will cost more, and so must deliver value in some other way.

IMO truly excellent EA takes less time and resource in the long run as the work is done once and only once, and so the overall cost of doing excellent EA is less than the cost of not doing it.

E.g. If successful EA can reduce the start up and analysis time for each project by 10% (say) then the total time saved across all projects in a large organisation is enormous and may far outway the cost of doing the EA in the first place. This effectively makes excellent EA "Cheaper than free".

Jon H Ayre
The Enterprising Architect

Tom G said...

I agree with your comments about the limitations of the more simplistic 'excellence'-based 'maturity'-models: as you say, the problem is that whilst they provide some kind of a metric, they don't really say anything about capability. The maturity-model that I use in my book "Doing Enterprise Architecture" focusses almost exclusively on capabilities and scope: specifically:

- 'what business are we in?' (context)
- 'clean up the mess' (horizontal integration)
- 'strategy to structure' (top-down)
- 'real-world impact' (bottom-up)
- 'dealing with the pain' (spiral-out, using mix-and-match of other capabilities)

Where I disagree with you is using money as the only measure of value, because frequently that will trap us into the common error of mistaking efficiency alone for 'effectiveness'. For real effectiveness, we need to link the architecture to all of the enterprise's values - not solely to money.

Richard Veryard said...

Many EAs share your opinion that "truly excellent EA takes less time and resource in the long run", but where are the metrics to support this opinion?

Anyway, the business case for EA may well be based on the claim that it will more than pay for itself "in the long run", but it is still necessary to have some form of cost control in the short term. Unless of course the EAs are willing to work for nothing, on a payment-by-results basis? (No, I thought not.)

In most organizations, costs and benefits don't always belong to the same cost centre. So although the total benefits may eventually outweigh the total costs, we still have to know what the costs are and how they are distributed.

So an immature EA organization says "one day you'll get back your investment in us, just wait and see". A mature EA organization is able to say "here's the evidence that you are now getting back your investment in us". And an excellent EA organization says "here's the evidence we have focused all EA activity to maximize cost-effectiveness".

My worry about the "cheaper than free" argument (whether for EA or for other kinds of quality management) is that it sends the wrong message in a world where the long run is never as important as the current crisis. What manager is going to devote resources to something that is supposed to be cheap? Surely we should be able to get it on the "never-never"?

Richard Veryard said...

If I ever imply that costs and benefits can only be measured in terms of money, then Tom would be right to slap my wrists. Costs include time and energy - people in organizations have finite attention and capacity for change, and even the most worthy initiative has to compete for management attention with all sorts of other initiatives. As for benefits, these can be measured in all sorts of ways.

But my point is this. If you can't find an objective way to reason about benefits, then you can't demonstrate that your EA is being effective at delivering these benefits.

Accountants and management consultants have developed ways of expressing all sorts of value in terms of money. If we don't like this, then we have to find some other way of expressing value.

Char said...

I am wondering if maturity models are the antithesis of Lean thinking. Increasingly feeling they are a dead end approach.