For a relatively standard description and history of EA one need look no further than A Common Perspective on Enterprise Architecture by The Federation of Enterprise Architecture Professional Organisations1A Common Perspective on Enterprise Architecture, by The Federation of Enterprise Architecture Professional Organisations, retrieved 03/11/2015 pdf here, where it says,
Enterprise Architecture is a well-defined practice for conducting enterprise analysis, design, planning, and implementation, using a holistic approach at all times, for the successful development and execution of strategy.
In other words it’s something that the enterprise should be able to use in order to help it achieve its goals.
If formal EA were any good, it should be everywhere – a discipline of incomparable power and undisputed success whose necessity was universally acknowledged2Relationships between EA and more fashionable endeavours such as Agile and DevOps will also be discussed in due course..
But it isn’t (though it is used very effectively by some), and from time to time people study the issues and try to explain why.
Only Connect. Fail.
Sven Roeleven and Jonathan Broer in their paper Why two thirds of Enterprise Architecture Projects Fail3Why two thirds of Enterprise Architecture Projects Fail, Roeleven & Broer, 2008 pdf here say,
The most frequent explanation given for this failure is that connecting EA to business elements such asand BPM is found to be difficult in practice.
The headline message from their survey is that two thirds of EA projects failed because the very thing that EA should do – connect strategy to execution – is “difficult”.
Over the coming months I intend to look at why formal EA as currently approached is so difficult in practice and why it needn’t be, but for now here’s a taster.
I don’t blame senior executives for not “getting” EA – it’s not a listening problem. They are not “buying-in” and offering “commitment” because they don’t understand the idea (so it is also not architects’ fault for being unable to communicate effectively), they are holding off because they don’t believe EA can deliver4I’ll talk about the psychology of motivation too at some point. – they don’t see much evidence in favour of EA or hear sufficiently persuasive arguments; and as Roeleven & Broer and others demonstrate, what may be called intuition is actually supported out by evidence against EA5Though it is also, to a certain extent, a self-fulfilling prophecy/vicious circle..
Nor is it a speaking problem: architects can’t communicate the reliable effectiveness of EA because it isn’t reliably effective. EA has an execution problem.
Yet it Should Work
All this negativity however entirely misses the point that EA should work. If you understand your enterprise well enough and can use EA to navigate the strengths and weaknesses of various approaches to implementing strategy, you should be much better off than simply “doing the vision thing” and hoping for the best.
So, what is it that makes formal EA difficult – and how do we make it easier?
I am going to put the fundamental blame for EA’s difficulty squarely on the ridiculously complex and incoherent frameworks and metamodels used to describe the enterprises, and in close second place, the disproportionate effort required to develop and maintain EA descriptions. In third, fourth and fifth places (order to be determined) will also come: the effort required to extract answers from architecture “products”, managing compliance with reference architectures, the division of powers, responsibilities and accountability between EA functions, operational and strategic management. And other things.
Then I’m going to propose a much simpler and more flexible approach to the conceptual aspects of EA modelling6and explain why existing structures fail and talk about how we can devolve a lot of the maintenance work so that the descriptions maintain currency without a permanent army of business analysts peering over the shoulder of everyone in the organisation and modellers trying to wrangle their descriptions to fit some baroque metamodel.
How I’ve Already Lied to You
I will also talk about why EA isn’t actually the absolute disaster I have just implied it is because, in fact, everybody already does EA – they just don’t do it formally and don’t recognise what they are doing as fulfilling the objectives of EA (and approximating its formalised approach) and don’t call it EA.
In other words, I will be saying that the fundamental nature of EA has been misunderstood; that will involve some more discussion of psychology and intelligence, references to Maslow and so on. I will let EA stand for other things – such as Enterprise Advantage, Executive Awareness, Evolutionary Action, etc. – to highlight a variety of perspectives on what this thing called EA really is.
The order in which I deal with these things – and other matters, such as the roles and responsibilities of architects, what we need from our tool-sets, how EA relates to business analysis and especially requirements development, testing, and so on – isn’t decided yet, but I hope that over the course of the next year good progress should be made in spelling out precisely how EA does work and how it can work even better.
I will be relying on a variety of concepts that I have developed in tedious detail (in order to confirm their consistency) that are documented in the kumu (you are welcome to start looking at the articles now), including my current view of EA as traditionally conceived.
Physician, Heal Thyself
The bottom line is this: neither EA nor its practitioners are foolish, but EA’s methods have evolved without being properly subject to its own criteria and consciously redesigned, so traditional practice hinders rather than helps. Now we can see where the problems are we can use EA’s own principles to fix it.
Then then C*O won’t be declining requests to discuss EA, she will be making them.
Notes [ + ]
|1.||⬆||A Common Perspective on Enterprise Architecture, by The Federation of Enterprise Architecture Professional Organisations, retrieved 03/11/2015 pdf here|
|2.||⬆||Relationships between EA and more fashionable endeavours such as Agile and DevOps will also be discussed in due course.|
|3.||⬆||Why two thirds of Enterprise Architecture Projects Fail, Roeleven & Broer, 2008 pdf here|
|4.||⬆||I’ll talk about the psychology of motivation too at some point.|
|5.||⬆||Though it is also, to a certain extent, a self-fulfilling prophecy/vicious circle.|
|6.||⬆||and explain why existing structures fail|