How to avoid a dog’s Brexit – an EA approach to massive change

How can we get out of a slime hole?

There can’t be many people that think Brexit has been a great success. But if an Enterprise Architect was involved – what might they do to help??

Spoiler: if you want the quick summary it's (1) recognise the current approach is a mistake; (2) start again by separating concerns; (3) find the best ways to change the current architecture to address each concern; (4) find ways to combine these changes into a coherent future architecture that addresses all concerns for all stakeholders. But please read on for the details...

A dog’s breakfast – British slang for a complete mess


Enterprise Architects work with large-scale, complex change – and sometimes things get out of control. A key reason to use EA is to have managed change, rather than random or chaotic change. Brexit was initiated following a referendum – in effect a yes/no vote about a specific stakeholder concern: should Britain leave the EU?

What is the best strategy when things get totally out of control? One option is to try to find an alternative approach. Brexit has tried this, but no-one seems to be able to come up with a viable alternative to the current Chequers Plan. But what do you do when none of the options actually addresses any of the stakeholder concerns? This is more or less the situation now with Brexit – no-one feels that the current architecture vision will resolve the stated concerns!

Maxim: Don’t be afraid to start afresh

When you use the SatNav and you make a wrong turn, the next instruction is “turn around when possible”. Large-scale change is usually very complex. Sometimes you only gain a full understanding by attempting the change. And sometimes it is the problems and issues that you face when you first attempt the change that give you the insights to learn, start again, and do it properly.

At the start of the Brexit process, at the time of the referendum, no-one was able to fully articulate the concerns and no-one really understood the feelings of the stakeholders (the British public, business owners, institutions, other countries, different regions within the UK…. heck, this list goes on and on). But the discussions that have ensued since the referendum have given light to a hugely complex problem – arguably the largest enterprise architectural change ever attempted on this planet!

Brexit is the largest enterprise architectural change ever attempted on this planet!

Roger Evernden

As an enterprise architect you should never be afraid to go back to your sponsors, decision makers and stakeholders to say: we didn’t fully understand the situation; we need to start afresh. You should never carry on regardless when it is clear that you are not addressing the real, underlying concerns.

When you start again, it is an opportunity to learn from the initial mistakes and avoid them the second time around.

Technique: Separate Concerns

Separation of Concerns is one of the most important techniques available to an enterprise architect. Most architectural change is too complex without breaking it down into more manageable chunks.

It’s nigh on impossible to start with a holistic solution. This is why there is a mantra in EA about making the distinction between “architecture” and “solution”. You start by thinking about, and understanding, the architecture before you leapt into proposing a solution. With Brexit the “architects” started with a solution: “leave the EU”, rather than canvasing concerns and then working out an architectural response that addressed those concerns, and only then moving on to look at solutions that matched the proposed architectural framework.

EA starts by separating concerns, then creating patterns that describe the concerns by explicitly explaining how they are caused by the components, configuration and behaviour of the architecture. By separating concerns, architects it is possible to explain each one in a clear and straightforward way. Each concern is given detailed and full attention, without clouding the issues by muddying the debate with other concerns. It is much easier to focus on one thing at a time… to try to understand the architecture behind one concern at a time.

Custom CSS

It is only when the individual concerns are well understood that it is possible to bring them together and gradually combine the single-concern patterns into bigger, multi-concern patterns.

There are two key EA principles involved in Separation of Concerns:

  1. Necessary deconstruction: this principles says that concerns must be deconstructed (not decomposed) to fully understand the architectural constraints that are causing them. Concerns must be deconstructed until each concern has been fully separated. It is necessary to separate concerns to be able to address those concerns.
  2. Stable composition: this principle says that, following necessary deconstruction, it is possible to combine single architectural elements into compound constructs. (This is the distinction John Zachman makes between primitive and compound.) This is the principle that makes it possible to combine single-concern patterns into multi-concern patterns, and ultimately to group these into a coherent architectural vision.

Brexit made the big mistake of assuming that there was only one concern: do you want the UK to stay in or to leave the EU? The debate at the time of the referendum made it clear that there were actually many different, complexly-interrelated concerns. Following the referendum would have been a good time to separate each concern, and help stakeholders understand the architectural issues behind each concern….

Technique: Concerns, viewpoints and views

A good enterprise architect must separate concerns and explain each concern to stakeholders holding different viewpoints. And then this is the clever bit…. they must be able to magically come up with a target architecture that takes every one of these concerns and viewpoints into account. It is a very special skill, but vital if you want to help everyone fully understand the paradigms and architectural context that created the concerns in the first place! It is the only way to get everyone on board – to get everyone behind you – to realise a colossal architectural transformation.

Guess what – the politicians involved in Brexit are more concerned (I choose that word carefully) with forcing their own opinion and view onto everyone else… which is an approach that leads to divergent thinking, rather than the convergent thinking needed to agree on a coherent future architecture. This comes back to to the need for necessary deconstruction of the architecture first (to fully understand the situation), before attempting to create a stable composition; divergent thinking comes first, followed by convergent thinking.

Technique: Patterns

Enterprise Patterns – patterns at the highest level of abstraction that encapsulate and summarise all of the lower level more detailed patterns – are an essential tool for explaining complex architectures. They are great for presenting alternatives and options, and they provide the necessary criteria and information to choose one option over another. They are generally presented in summary form as a diagram, supported by information about the risks, benefits, consequences, value, costs, savings, etc. associated with that specific option.

I’ve followed the Brexit process closely, and I can only recall one attempt to summarise some of the complexity inherent in the current architecture through a diagram that equates to an enterprise pattern. It was a graphic by Ian Bott in the Financial TImes that provided a Visual Guide to the UK in Europe. [If you click Skip This you’ll get to see the diagram in full, and then if you scroll down you’ll see the type of supporting information that is needed to support an enterprise pattern.] You can also see this as a video. It’s well worth watching, and I can’t believe that only 7,000 people have done so!

The key point here is that most people like to, and try to, make sense of the situations they face. And a lot of us tend to think in patterns – we extrapolate abstractions from huge volumes of data so that we can understand what’s going on. These patterns may start life as hypotheses because they are based on intuition and assumption, but gradually they become clearer and more fact-based. Gradually it becomes much easier to see if you want this or that.

The referendum that started Brexit wasn’t a choice between clear expositions of the alternative patterns for a future EU architecture. It was a choice between YES or NO; to REMAIN or to LEAVE. And without the necessary information or patterns to help the public decide, it became a purely emotional decision… which is not a good way to plan architectural change!

Manage expectations

Managing stakeholders, and in particular managing their expectations, is a vital EA skill. I advise clients to think about the three TOP areas for managing expectations. This is, of course, an acronym:

  1. Timings – how long will it take to achieve?
  2. Outcome – what are the likely consequences and results?
  3. Process – what process do we need to follow to achieve the outcome?

Some people thought that the result on referendum day meant that the UK was, from that point onward, out of the EU! Some people think that it will all be over on 29th March 2019 – that the UK has managed to extricate itself from the enormous complexities of the EU.

Teresa may satisfy a few stakeholders with the Chequers plan, but tht doesn’t sound like a good commendation for an enterprise architect!

Roger Evernden

Any EA project takes time. The more complex, the bigger the change, the longer the time, the greater the cost. No-one explained to the British public that Brexit is a long-term project. Administrators and civil servants will be working on this for decades. And that is when the realisation of the target architecture begins. But we haven’t even got the detail of what this future architecture state will be! No one has yet fully developed the target architecture; there isn’t even a clear architectural vision.

In conclusion…

Brexit is a mess – and that is clearly an understatement. The current “plans” are unlikely to satisfy any of the stakeholders. “Leaving the EU” is too big a task to handle without breaking it down into its constituent concerns.

Image a family of four living in a house. No one is fully happy – it has some good points and some bad. The mother and son want to stay in the current house. The father and daughter want to leave and find a different house. They argue for months about what to do. The father finds a house that he likes and forces everyone else to agree with his decision. The family move home, but the house they move to isn’t any better than the one they left behind. And it has cost them all of the removal costs and legal fees…

Their neighbours also want to move. The father, mother, son and daughter talk openly about what they don’t like with the current house, and what they would like in a new home. The list criteria that help them choose a new home. They are able to describe their ideal new home to an architect, who offers them some options.

One is to redesign their old home so that it now fits their needs. Another option is to find a new home that is an exact match to their requirements. There is no option for a new home that is no better, and may be a lot worse, than their existing home.

  • One is to redesign their old home so that it now fits their needs.
  • Another option is to find a new home that is an exact match to their requirements.
  • [There is no option for a new home that is no better, and may be a lot worse, than their existing home.]

The only conclusion that an enterprise architect would make is: learn from the first attempt and start afresh!