Sprint Length

When we started with Scrum a year ago, we decided that our sprints should work on the calendar month.  Specifically, the sprint ended on the last Thursday of the month and started the following Monday.  That means some sprints were 4 weeks and some were 5 weeks, but we felt the consistency of ending the sprint at the end of the month was more valuable than having a consistent length.  We also wanted all teams on the same sprint schedule to enable broader team sprint reviews as I described in an earlier post.

The calendar month sprints have served us well overall, but we have had at least a couple of exceptions worth noting. 

The first exception had to do with the quality of the backlog going into a sprint.  In our experience, there are cases where you just can’t define a good goal or set of goals for a month long sprint.  This is often the case when you are early in a milestone and there are a lot of unknowns.  In this situation, we found that the best approach was to shorten the sprint.  We only attempted to plan what you reasonably could do, with the output of the shorter sprint feeding the next sprint.

The other exception is with a team that I am actively working with now.  The team is building a Sample Application for an MBF milestone that will be delivered to some early adopter partners.  The Sample Application is delivered right with the framework.  It has the unhappy requirement of starting late in the cycle due to dependencies, but still shipping at the same time as the rest of the framework.

The milestone turned out to be 9 weeks of development time for the Sample Application.  As the ScrumMaster, I didn’t want to wait 4 weeks to review and re-plan if needed.  So we decided to use three 2 week sprints and one 3 week sprint.  Since many of the features are small, we felt that we could deliver value in each of these sprints.

We just completed our third 2 week sprint.  At the Sprint Retrospective yesterday, I asked the team how they liked the 2 week sprints.  The team was unanimous in favor of the short cycles.  It has helped keep us very focused and helped manage time more effectively – exactly what we wanted it to do!

In summary, monthly sprints have been a good thing.  But exceptions which have caused us to shorten the sprint have also worked quite well.  Let common sense be your guide…