It has been known for quite some time in the agile community that estimation is waste. And yet, I still see many, many teams spend huge amounts of time in story pointing sessions. I’ve experienced people passionately debate whether a user story is 3 points or 5 points. If that isn’t a waste of time, I don’t know what is.
I’ve stumbled across a pattern over the last 5 years or so that has probably always been there, but, hey, sometimes I’m a little slow (and, I’m not the only who’s noticed this by the way). When I first noticed it, I was a little surprised. Once I noticed it, I saw it everywhere. But it made complete sense.
When estimating with story points, the purpose is to determine a velocity (total number of story points completed in a sprint) so that you could estimate how much a team can deliver within some period of time. I was a big believer in this…it really worked. It helped us be able to say “no” to more work during a sprint. Drawing that line in the sand was awesome as we had real data to know where to draw that line.
While teams discover their velocity in points, I thought I would see how velocity correlated to number of user stories completed. And then I found the pattern. Teams consistently not only delivered approximately the same number of story points during a sprint, they also delivered approximately the same quantity of stories. When I first discovered this, the team I was working with consistently delivered between 9-11 user stories every sprint. I went back at least 5 sprints, and it was always the same. I thought maybe that was a fluke. So, I looked at other team’s (on the same initiative) product backlogs. The same thing. Consistent number of stories every sprint.
Afterwards, when I worked with other teams, I noticed the same thing. It’s extremely consistent. And it made complete sense. When you decompose user stories to the point where they fit into a sprint, you will find that some stories you thought were tiny are a little bigger, and vice versa. Ultimately, it “comes out in the wash”.
I do have one caveat. If, for some reason, your team(s) have a tough time decomposing stories and the size of them vary wildly, then this may not work so well. However, instead of trying to estimate better, you should put your efforts into learning how to decompose your work into smaller batches. There will be a much bigger payoff.
Having experienced the widespread and consistent pattern, maybe it’s time to deprecate story points. I know what you’re thinking, “what about long-term planning?”. Long term planning smells like waterfall. However, it is completely reasonable when a customer asks, “when do you think this will be done?” or “how much can you complete by October?”. Ultimately the answer is always “we’re not sure”, but that doesn’t always fly. There are other ways to provide a target (or a guesstimate). But, alas, I’ll leave that for another post!
What do you think? Is it time to deprecate story points?