We often hear about people who do retrospectives but struggle to get the team to agree to actions that really make a difference. Sometimes people struggle to come up with actions if they did well in a sprint. We’ve had this problem ourselves. Something we did recently in our retrospective was to change the idea of a retrospective action, into a retrospective experiment. The benefit is that you automatically think about how to tell if the action is working or not. It also forces you to think about the learning rather than just the doing.
Here is an example of our last experiment. We do a combination of onsite work at clients, talking to people (in person and via phone), and work from our office. We noticed that visits out of our office are quite disruptive to some of the focussed work we do like writing our book.
So we constructed an experiment. For the next month, we would tag 2 days a week as out of office days. We would schedule any visits or onsite work on these days. That would mean 3 days of uninterrupted work each week.
We speculated that it would result in the following differences:
- We would be better at getting work done on the in office days
- Our exercise and lunch routines would be more predictable as we would only have potential disruptions twice a week.
- We would spend more time connecting with people via phone or email, which would mean we’d be more likely to connect with those outside Cape Town.
- We would time box visits with clients better because we would have somewhere else to be later.
- We would appear to have less ‘admin’ workload because we wouldn’t be squeezing it in to small gaps before meeting a client, but would rather tackle it in a focussed way.
When we next retrospect (next week) we will see which of these are true as a result of the experiment.
This not only gets you to take an action, but gets you thinking about the impact of that action, and evaluating if it had the desired effect. If it didn’t make any of these things true, we can decide not to keep the action, but rather to try something else.
So give it a try. Instead of wording the outcome of your retro as an action. Try set it up as an experiment. Let us know what you learn.
