Escaping the edge of chaos

I’ll admit it — I spent the first 3 months of this year in a bit of a haze. A couple of reasons — firstly, the arrival of my twin boys (amazing) and secondly I was in a pretty challenging place at work.

We had a vision of what we wanted to achieve, and boy was it ambitious. We were pushing the boundaries of what is possible within our organisation, surfing the edge of what’s technically possible (federated clusters across multiple cloud providers — come on Ubernetes).

Six months into the year and we were out of the haze. Partly because my wife is amazing and the twins are thriving. Also partly because we got a grip on what we were doing at work.

I’ve spent a bit of time reflecting on what we’ve done and how to tackle similar challenges better next time.

When trying to understand the dynamics of the situation, I’ve found the Cynefin model and Ralph Stacey’s work on understanding the level of agreement and certainty really useful. Jurgen Appelo describes these models nicely in Management 3.0 and has some handy definitions, which I’ll ruin by paraphrasing:

  • Something complicated is difficult to understand but with a predictable outcome
  • Complexity is a measure of predictability of outcome

The management team craved (and continues to crave) more and more pace and predictability of outcome. Back at the start of the year we were nowhere near delivering any outcome — we were at the edge of chaos.

A quick word on chaos… chaos would be something like a global systems failure (ahem, BA) or a massive security breach (ahem, NHS). In that situation you’ve surely got to just act, and hope that those business continuity and disaster recovery processes you’ve never really tested pull through. We weren’t in that situation. Just on the edge of it.

In a corporate environment, pretty much the number one thing that management look for when making decisions on what to do is predictably — what am I going to get for my money?

Try to do something new and difficult in an environment that is going through a transition and you’ve got a recipe for unpredictable outcomes.

And it’s a rare business that will tolerate too much unpredictability for too long.

We were in a situation where we couldn’t really predict an outcome. And a management team that was going to run out of patience pretty quickly.

So what did we do?

Went back to basics and focused on the most important problem to solve next. We formed small teams to tackle the problem, and approached things in short cycles. We helped the teams adopt SCRUM to give some structure and tools to organise their work.

We followed the core principles.


Inspect and Adapt.

Gradually the logical order to solve the most important problems emerged.

Gradually teams start to get a handle on the problems and a rhythm to tackling them.

Gradually we were able to start providing some degree of predictability — at first the horizon was only a couple of weeks, but gradually we extended that horizon.

So this being Medium, we need a list. What steps would I recommend? What will I strive to do next time I find myself in this situation?

  1. Communicate
  2. Get comfortable with the ambiguity
  3. Communicate
  4. Take action on the most important problem
  5. Communicate
  6. Stick with the vision
  7. Communicate
  8. Reflect and adapt
  9. Communicate

If you are in some sort of leadership position (rather than actually doing the work), more than 50% of what you (individually) do should be communicate. Tell people what’s happening. Explain the challenge. Explain what you’re doing. Show them what you’ve done. Ask for their help. Keep talking to them.

The rest is about being confident and methodical:

  • You are in a complex, maybe chaotic, state where there is a complete lack of clarity — accept it and then embrace it.
  • There is only one way out of your current state — take action on your most important problem.
  • You should have a good idea of where you want to be — hold the vision front and centre and don’t compromise.
  • Whilst taking action is mandatory, take regular stock of that action, what should you do next and how can you improve on your process.

There are certainly other things for that list — let me know what you would prioritise.