It’s no secret among agile practitioners that retrospectives, as commonly practiced, tend to suffer from a fatal flaw: inaction. Teams discuss all that irritates, perhaps cheer each other on a bit, and then leave the retrospective with faint hopes that things will improve, eventually losing faith in the activity due to its obvious ineffectuality.
The simple solution is to choose a single high-value issue each sprint, and allocate the time necessary to solve it. Some practical ways to make this happen are:
1. Set expectations properly: Ensure that your stakeholders expect a bit of effort toward process improvement and understand the value this brings, both generally and in each case.
2. Keep things small: Improvement actions should go in your backlog. As with other item, they work best when they’re able to be complete in about one to three days. If the action is too big or too broad, break it down into smaller, more manageable items.
3. Gather observations continuously: Much time in retrospectives can be spent just raising issues, which reduces the time available to craft and refine solutions. Keep an ongoing backlog of observations (and any suggested actions) that team members contribute to in real time, and this will help jumpstart meetings.
There are many great models for retrospectives available, and we’ve found that while there’s no one perfect structure for all situations, there are some consistently effective patterns. Perhaps we’ll explore these in a future post.