Event Storming: Thoughts from a Facilitator
There was a point on Monday morning where I was looking at a 25 metre long board full of randomly arranged orange stickies and thinking… we are never going to be able to sort out this mess. I had to fight a sensation of panic and I know the other facilitators were thinking the same thing, despite the mantra of “trust the process”.
Yet somehow over the next day a meaningful narrative emerged. It still feels slightly miraculous, but groups of people have this super-power of finding the order inside chaos.
Event storming is a method which at heart consists of getting lots of people to write down events on stickies, and then order them into a simple narrative. In a way, event storming pedals a lie. Time-lines are clearly too simple for almost any business, but time is the best ordering mechanism we have. Trying to stick with time as long as possible lets us create journeys. Eventually though creative additional interventions can be forced upon us. During these two days of event storming at Gousto, for instance, I saw two such interventions.
On Monday, we split the stories into chapters, coherent stories where timelines actually worked. Chapters gave a high-level index that allowed the events to cluster more rapidly. On Tuesday, the participants added swim-lanes, noting that there were coherent stories — but there was more than one happening at any moment in time. Below is part of a high-level picture with swim-lanes (the rows), chapters (in blue) and events that was used to guide the much more detailed story on the main wall:
By the end of the two days I now see Gousto more like waves crashing on a beach. As the waves progress toward the shore, we move from the realm of forecasting into planning until the moment when we have to commit to customer orders. Our team aren’t just tracking one wave at a time. They are on top of the whole cycling journey of the waves.
Anyway, now I’ve wrung out the towel of the beach analogy, here are a few thoughts on what works really well.
Event Storming Crowd-Sources Knowledge
There were about 25 attendees on Monday, and 50 on Tuesday. The increased numbers were because we brought in software developers. If you put that together and add time pitching the idea, preparatory work and facilitator time, we are probably at 100 days of effort. That may sound like a lot of work but ask yourself: how long do you think it should take to get a meaningful consensus between 50 domain experts? 2 days seems to me to be astonishingly efficient.
Event Storming Builds Consensus on Issues
During the day, we also asked people to highlight issues as we walked through the story. At the end, we gave people two votes and clear points of pain emerged. The idea is of course that most people will give one vote to their pet peeve, and then the other to something which might not specifically be a problem for them, but where they understand the pain.
Some concerns were raised about this way of voting. For instance, the voting is likely to be weighted towards how many representatives there are from particular domains. However, such an interactive process encourages empathy, and when there are genuine stand-out issues, this simple consensus feels sufficient to see what really needs sorting out.
Event Storming Is Team-Building
I’ve already mentioned the process of empathy we saw at work. People are sociable and whilst we like to be able to work from home, most of us like getting to know people too. The steps of listening, questioning, speaking up, and occasionally just stepping away all encourage and reinforce our sense of togetherness. Challenging the board, rather than people, is very effective at avoiding the occasional tensions you can see in ordinary meetings.
Event Storming Finds Natural Domains
Event storming is obviously good at building a model, and it’s good at highlighting issues. Another objective of event storming is to spot domains and see how well they map onto the organisational structures that we have. To some extent I’m not even sure if this is different from spotting problems. If an area has few issues, who cares what the structure is? However, if we look at the cut-off flow stage where there are many issues, we see that there are many squads involved. Of course we have to ask ourselves: is that part of the problem?
Event Storming Is a Ball
Event storming has a boot-strapping problem. It’s impossible to work out who should come in advance, especially given that we don’t know what we need to discover. If event storming is a ball, then many people did not come to the ball and we did not benefit from what those people could offer. For this I apologise, but consider how democratic this process is compared to a purely top-down approach.
During the event storming, participants were asked to share their hopes and fears, but as a facilitator I didn’t get to share mine. My fear I’ve already told you: could we go from chaos to order? The most common fear is probably: will we act on what we find? Personally, I’m not worried about that. Holding the event storm demonstrates the massive commitment to pursuing the information that comes out of it. My hope is simply that event storming will become another tool we use at Gousto for building understanding, not necessarily at the org level in future, but at the right level for whatever problem it is that we find ourselves wanting to solve.