The Good, the Bad and the Ugly of one week sprints
I recently saw Bob Martins Keynote on Øredev. Around time index 14:30 he is talking about length of iterations. This reminded me on some experiences we made recently when trying sprints of one week in some of our teams. So here are some aspects we found.
- Scrum zeremonies are done very often. The people on the teams were used to scrum, but the teams were new. So the training effect of doing planning, review and retrospectives so often was good. By the way, the zeremonies are proportionally shorter. There is no waste of time just because there done more often.
- The short iterations gave a good feedback on where the project was going and how long this will take.
- The focus is very high on completing the sprint-tasks, but this is not only a good side. See also the Ugly.
- Teambuilding happens faster than with longer iterations. I think this is mainly, because teambuilding needs all phases of an iteration to be completed and shorter iterations mean the phases happen more often.
- Tasks need to be prepared very well. A one week iteration is not only challenging for the team. It is even more challenging for the product owner. Tasks need to be well prepared and small enough to fit into one week.
- Some people on the team will not like such a short iteration. You need to get the commitment of everybody that they want to try this for a while.
- Things that go wrong, hurt faster and harder. This is good as long as the real reason will be removed. But it is easy to think of the short iteration to be the problem instead of the real reason.
- As already said, it is easy to get in a situation where the one week iteration is seen as the problem even if there are other reasons. This can remove this option for the future. So make sure the situation is ready, before you try.
- People can feel like being in a hamster’s wheel. This prevents from taking a step back and seeing the whole picture. Short sprints can show a lot of problems in the process, but they can hide larger issues or prevent alternative approaches. Everybody just feels they have no time for it.
So what is your experience with one week sprints? Is it just like Uncle Bob says: “It takes a real man to do one week iterations”?