Tracking raw stats is important, such as the actual velocity of each sprint. Trends on these stats are important as well, but can often show a lot of jitter that makes understanding the overall health of your teams more difficult.
Slow Growth in Velocity
Dev Team A had been struggling with delivery. They were working on many projects and would commit to doing a lot but rarely, if ever, completed a full sprint. The team was demoralized. They were all good developers but completing a sprint seemed beyond their grasp.
After the manager took over the team, they looked at how much the team was delivering vs committing to and reset things. Starting with Sprint 5, the team took on much less work. They focused on getting things done and followed good Dev Principles. They had a successful sprint. In the next sprint, they increased their commitment a bit. With each sprint, as they improved their velocity, they looked at the rolling 4-sprint average for their velocity and the team pushed a bit further. In Sprint 9, they got aggressive and fell a bit short of the commitment. However, they still delivered a solid sprint. In the following two weeks, they struggled a bit with finding the right commitment level but ultimately figured it out and kept pushing.
If the team and their manager were only looking at the past sprint and reacting accordingly, they'd have far more ups and downs in their velocity graph. But by looking at the Rolling 4-sprint average, singular drops are smoothed out and the team has a clearer idea of what they're capable of.
We can see how the peaks and valleys in sprints 5-11 are smoothed out in the second chart. In Sprint 9, the average was 31 points but the team pulled 38 points. While their manager loved their newfound confidence, the team wasn't quite ready for that much work. For the next few sprints, despite the average being in the low 30s, the team challenged themselves to get into the mid-30s for a velocity. By Sprint 13, they had done exactly that. By sprint 16, 3 months after their manager joined them, they'd increased their rolling average velocity by 30%.
Incremental Gains
Velocity is not increased in large amounts overnight. By pushing sprint commitments up slowly, based on the rolling 4-sprint average, the team gains confidence in their ability to deliver more and learns how to use the rolling average to plan their capacity for the next sprint.
This graph shows the difference between the rolling 4-sprint velocity average at the start of the sprint and the number of points planned into that sprint. This formula is (µ-1 - S) where S is the number of points planned at the start of the current sprint. You can see a direct correlation between where a team veered significantly from the rolling average and their sprint velocity.
You can see how the team was radically over-committing in their early sprints. Initially, they backed off dramatically to regroup and get a win. After a few sprints, you can see they, again, over-committed by 10 points in Sprint 9 before correcting. While the progress isn't smooth, the team is slowly moving towards a place where the sprint commitment is within a few points of the rolling average. This is desirable and shows a team that is both challenging themselves for improvement as well as one that isn't bowing to artificial pressures for velocity.
Comments