16 November 2009

EA Quick Start Guide (Part 1): How to Set Up an EA Practice

This article is the first in a sequence providing an introduction to Enterprise Architecture suitable for those starting out in the profession, or for those outside the architecture profession interested in what the fuss is all about.

In this first article I would like to address an issue that has particularly bothered me recently; the setting up of an EA practice. Much is made of the activities required to establish such a practice, but in my opinion, the advice given is at the heart of why others are frustrated with the lack of value delivered by Enterprise Architecture in the early stages of its adoption.

What Not to Do
Firstly, let me address what not to do. Common “wisdom” seems to believe that the initial steps should be as follows:
  1. Identify the fundamental principles that the architecture will support.
  2. Establish the governance framework that will ensure compliance with the architecture.
  3. Establish the architecture framework that the EA practice will adopt.
  4. Start to engage with projects.
“What's wrong with that?” you might ask. It sounds reasonable enough at face value, but this approach is unlikely to result in quick acceptance of EA as a valuable activity.

Firstly, of what use are principles to anyone? They are usually a re-articulation of the obvious, and stating them (or more importantly, wasting time word-smithing them) will simply establish right from the outset that your so-called EA practice has every intention of being a talking shop that makes a profession out of navel gazing. What is more, what exactly is anyone else supposed to do with these principles? They can try with all their might to conform to them, but at the end of the day, they are simply words with many possible interpretations.

Principles may be useful in guiding your architects in the development of the architecture, but they are no more use to the recipients of that architect than a set of coding standards (without any code) are to a tester. In reality, principles in isolation are simply ambiguous tools that self proclaimed architects can use to impose their opinions on others.

…And this brings me to the second step. What are you doing establishing a governance framework when (a) you have nothing against which to measure compliance, and (b) you have not yet established the credibility or respect to demand that others do as you say. Establishing a governance framework at this stage also sends out a very strong message that the primary intent of your practice is to block progress, rather than enable it.

The third step is not so contentious, but I would consider it to be too burdensome to undertake during the emergent phase of your practice. In my opinion, frameworks are useful sources of reference information that can be dipped into and cherry-picked as work progresses on your architecture. Use them as you would a dictionary or encyclopaedia. As for the fourth step; engaging with projects at this stage will simply destroy any credibility you may have had, as you will simply be exposing your lack of real deliverables to a sceptical audience. What is more, you will be dragged into the day-to-day point solution world of the project, and become just one more design resource. This will take your architects away from the bigger picture, and prevent your EA from being developed.

What to Do
So what should we do? The hint is in the previous paragraph. The one thing that the organisation needs you to do as an EA practice is develop your Enterprise Architecture. Until this starts to happen, you cannot provide any real guidance or value, and you certainly cannot justify what you are saying if it is not clearly recorded. You will, instead, always appear to be “making it up as you go along”.

For me, the ideal first steps in establishing an EA practice are:
  1. Ensure you have the right people.
  2. Establish basic working practices.
  3. Start building your architecture!
Those are the basic steps. The following sections give some guidance on how best to perform them.

What Makes a Good Architect?
Enterprise Architecture is a skilled, specialised activity, to be performed by a small number of people. One weak link could fundamentally undermine your effectiveness. (This should be obvious, but for some inexplicable reason EA practices more than any other area of business are set up from a random selection of people, rather than those selected for their previous experience, or potential to be architects. Some key attributes required by an architect are:
  1. Ability to communicate verbally, visually and textually.
    Architecture is (roughly) 80% communication. Bad communication will frustrate others and negate the effectiveness of your architecture.
  2. Ability to listen and quickly assimilate new information.
    Architects are always dealing with the future and thus the ability to learn fast and think fast is essential. What you know today will almost certainly be out of date with respect to a true “to-be” architecture.
  3. Ability to influence and enthuse.
    For your architecture to be successful it must be followed, and for this to happen others must be bought into, and enthusiastic about your architecture. This will only happen if your team have the presence and ability to make it happen.
  4. Credibility.
    Some people have it, some do not, and worse still, some have at some point in the past lost it. There is no point in starting with a disadvantage.
  5. Ability to see the big picture.
    An architect must be able to consider the organisation as an entire system. Those who prefer to focus on point solutions or drill down into the detail of one thing at the expense of the rest are not suited to architecture.
  6. A natural preference for simplicity.
    The best architecture is the simplest possible one that can fulfil the business vision. Those who gravitate towards, and enjoy complexity do not make good architects. Those who abhor complexity and take please from finding the simplest possible solution to a problem are the ones you want.
But what about specific business or technical knowledge? For me, it is not specific knowledge that matters, but a demonstrable ability to adapt to new situations, and understand new ways of working, and new enabling technologies at the drop of a hat. Those with specific in-depth knowledge generally measure their own value in terms of the relevance of that knowledge. As a result these people often fear change, as it may negate that value, and thus negate their importance. They are not best placed to view the future from an unbiased point of view. Unfortunately, businesses too often focus on finding people with the specific knowledge to support their current position, and then wonder why these people bring nothing new to the table.

What are the Right Working Practices?
When setting up an EA team, the best working practices are those that will allow you to both engage with projects and continue to build your architecture, without the demands of one impeding progress on the other. You need to quickly establish who will support projects, and who will build architecture, and then you need to ensure that these two groups retain a close working relationship as each will feed the other.

Your working practices need to support a proactive approach where the build activity seeks information regarding the future strategy, (and in turn helps to guide future strategy) and quickly articulates this information as future architecture.

You also need to support a reactive approach that allows the architecture to engage with projects at an early stage and either feed them with information already recorded in the architecture, or re-prioritise build activities to focus on the areas of the architecture that will fulfil the needs of the project.

How Should We Build It?
Build it? That’s the scary bit, isn't it? This brings us to the heart of the problem. So many EA practices spend time avoiding the really important bit because they are frightened of it. I strongly believe this is why so many processes for establishing an EA practice start with a sequence of what I feel are little more that prevaricating steps. In a consultancy led world, an outside “specialist” can use this approach to provide a significant amount of content without actually having to get involved in building anything.

In the early stages you need to recognise that initial demand will come fast, and will come from two sources; above and below. From above, you will need to engage with the business to articulate the future vision in the form of a Business Architecture. From below, in-flight projects will start to demand guidance relating to low level implementation decisions. You will need to have the courage to make some decisions relating to future technologies with very little guiding information, and articulate these decisions in your Technology Architecture. Some of these decisions may turn out to be wrong, and you must be willing to accept this, adapt your Technology Architecture accordingly, and help projects to absorb these changes in a pragmatic manner. Remember, that sometimes the simple act of making a decision and sticking to it is what makes that decision the right one.

(I will cover what goes into an Enterprise Architecture in terms of levelling and content in later posts in this series).

It is this build focused activity that guides your engagement with projects, and more importantly with those who set the vision and strategy of the organisation as a whole (and yes, that does include you as an architect).

And the Rest?
Let’s now return to the original steps that I declared to be the wrong ones. They were:
  1. Identify the fundamental principles that the architecture will support.
  2. Establish the governance framework that will ensure compliance with the architecture.
  3. Establish the architecture framework that the EA practice will adopt.
  4. Start to engage with projects.
Am I suggesting that these are not important, or can be ignored? Of course not, but what I am stating is that these are not the very first things you need to do. Once you are up and running and developing the architecture (and thus hopefully providing some real value) you can address these other areas on an as-needs basis.

Principles can be recorded if you like, but remember that your architects, if properly selected, will already naturally apply sensible principles to everything they do. Do not expect others to act on these principles however. They need you to act on them and provide an architecture that fulfils these principles.

A governance framework will be important as your architecture develops, but you must start in a spirit of cooperation and assistance to win support and credibility. Once this has been established and you have some content in your architecture you can start to introduce more formal governance mechanisms based on this content and the trust you have already established.

An EA framework is also important, but it should be your own, and it should be adapted to and derived from the unique circumstances that exist within your own organisation. Once again, attack this on a just-in-time basis, cherry picking from the various reference frameworks those features and details that you find you have a need for, and always consider carefully whether it really adds value to what you are doing. (Roger sessions wrote a good article on EA framework comparison available for download at http://www.objectwatch.com/whitepapers/4EAComparison.pdf).

Is This the Only Way?
Of course this is not the only way. You are an architect, and you are in charge of your own approach. If you cannot set your own direction of travel, then you need to question whether you can set the direction for others to follow. The intent, however, should always remain the same:
To start producing an Enterprise Architecture and to become a useful addition to the organisation in as short a time as possible.
The longer you take to do this, the more likely it is that someone will decide (correctly) that you are surplus to requirements, before you even get started.

The Enterprising Architect

1 comment:

  1. In the past months I worked my way up and just recently I was asked to "get into the EA". I'm reading so much and I can't just get my head around things at the moment, but I'm sure they will in time.
    I like your pragmatic approach and it feels so right for me at this time. The only thing I wonder is how I will manage all of this, starting up an EA with only myself seems just too big atm....