Fightin’ Words

Well, this is fairly provocative: Story Points Are Pointless, Measure Queues.

I’ve advocated for story points plenty of times, but I do see the article’s point. Estimation is hard. And I like the notion that in absence of other data, just having a count of tasks yet to be completed is a decent approximation of remaining work.

And perhaps it rightly centers the definition of the tasks themselves as the important part of estimation, versus the assignment of a numeric value to said tasks? A question well-asked is half-answered, or something like that?

It also made me think of a tension I’ve seen for much of my career. Product folks tend to think in terms of features and user outcomes. As they should, for outcomes is what ultimately matters to a customer. But estimating “features” is a fool’s errand. Engineers need to think in terms of tasks: add X, build Y, modify Z, etc. The latter are (generally) easier to estimate. And, as the article argues, if sufficiently broken down, don’t really need to be estimated at all, just counted.

Making the above work, though, requires that engineers own that translation of user stories to tasks (and consequent planning thereof). That takes work, no doubt, but probably less work in the long-run than arguing about numbers.