Sunday, October 11, 2009

Changing your Sprint Length

So we recently undertook an "experiment" to switch our sprint length from 2 weeks to 1 week. This was due to a high priority project that needed some extra visibility and ability to "course correct" if something came up.

Here are some things to keep your eye on if you decide to reduce your sprint length:
  • Meeting Overhead - You are now doing your sprint planning, review, and retrospective more often. While some may say that these meetings need should be reduced in length, the reality on the ground is that the shift in sprint length causes enough of a change where the meetings will likely take the same amount of time.
  • Product Owner Time frames - With a two week sprint you get 2 backlog grooming sessions before the next sprint starts. That allows you more than enough time to get the backlog estimated and in order to ensure a smooth sprint planning. With only 1 backlog grooming session, the churn is constant and its harder to stay ahead of the team. Remember if the team decides to quickly move in another direction, you also have to keep your management aware of the changes.
  • Work Size - The team is quite used to 2 week sprints. There is a comfort level with the size and estimate of the work. With 1 week sprints you can easily find it difficult to know the right amount of work to take on.
  • External Tasks - A 1 day vacation or an illness of a team member because material. All day meetings or long 2-4 hour meetings become material and difficult to manage. Don't forget to plan accordingly.
Be mindful and aware...if you know these "gotchas" ahead of time you can prepare and mitigate accordingly.

2 comments:

  1. As you say Peter, sprint length selection has a strong influence on the work dynamics of the team. Basic Scrum used to strongly influence 30 day (4 week) sprints. I found that these were too long from two key perspective. First, inexperienced teams really struggle with planning a 4 week sprint. And second, you have to wait too long for results feedback and your retrospective feedback. That's why I think so many teams have moved to 2 week sprints. In this case planning and feedback is quicker/better, but there's a lot of *waste* associated with ramping up and then down each sprint. I think that's the key problem I have with 1 week sprints - the fact that you probably only get ~ 3days of productive time every week. I think teams need to give careful consideration of sprint length give (1) the type of work and (2) the maturity level of the team.

    Good post!

    ReplyDelete
  2. Seriously though. I have worked in one and two week sprints and my favorite is the one week model. We found that there is less time for things to go astray, and the team has to demo at the end of each week. So you get more focus on the delivery.

    ReplyDelete