Velocity, capacity, and unplanned work
Velocity usually includes unplanned work, which limits its usefulness for capacity planning and forecasting.I’ve not a fan of velocity as a metric, but many people use it. What’s worse, many people misuse it in a subtle, but important way that I want to discuss.
Velocity, as commonly understood, is a measure of all work done during an interation. Most often, it’s an expression of story points per sprint.
Ostensibly, velocity is typically used for one of two things: Capacity planning (how much work can we get done in the next sprint), or completion forecasting (how long will it take to complete this epic?).
The problem is that velocity doesn’t give us much insight into these things for the simple reason that it typically includes unplanned work. This makes velocity less useful for capacity planning, and essentially worthless for forecasting.
Say you pull in some work half way through a sprint because of some urgent bug, and at the end of the sprint you calculate that you completed 25 story points worth of work. What does this tell you about your working capacity? Well, maybe that you can complete 25 points worth of work in a sprint. But how much work should you plan for the next sprint? That depends on how much unplanned work you may get.
And how long will a 100-point epic take to complete? 100/25 = 4 sprints, right? Oh, but again: What about any unplanned work that may arise?
To avoid this mistake, simply don’t count unplanned work in your velocity tracking. Often this measure is called capacity rather than velocity, although semantically the two words have very different usages.
In our 25-point velocity example, if 5 points were unplanned work, then we should consider only 20 points/sprint as our working velocity, as that’s what we can plan for in the future, and it’s what we can more reasonably count on for our epic forecasting (100/20 = 5 sprints).
Now to be clear, velocity tracking, and indeed story points, are at best imprecise (by design), so using them for planning and forecasting does come with caveats (which I may get into in the future). If you’re not using velocity now, I don’t suggest you start. But if you are using it, be sure to use it “correctly”, and only count your planned work.
