Skip to content
Switch branches/tags
Go to file
Cannot retrieve contributors at this time
title date slug image keywords
Story Points Aren't Units of Time
2019-07-16 23:45:29
digital transformation
software development

They just aren't.


Search for "story points agile" on Google. Try it. You don't even have to type it into Google; click the link!

You'll get, at this time of writing, approximately 12 million results. Accouting for the 8 million results are bots promoting something that requires your wallet, that leaves four million web pages, many of which will go on to describe story points to the letter and how they aren't about estimation.

Here are a few:

From To The New

A story point is an abstract measure of effort required to implement a user story. In simple terms, it is a number that tells the team about the difficulty level of the story. Difficulty could be related to complexities, risks, and efforts involved.

From Agile Alliance

Story points are relative measurement of the size and complexity of the user stories wherein a base story is assigned some story point/s to start with and rest of the stories are estimated in story points based on its size and complexity compared to the base story.

From Plataformtec, a result I pulled from the eighth page of search results:

Story points were developed because human beings lack abilities to forecast the future. Our bosses were saying “ok, ok… you wanna use this ‘agile’ thing, but I still wanna know at least when you are gonna finish this new feature!”, and we had to improve our forecasting beyond the “well it will be ready when it gets done” phrase.

(Sorry, Plataformtec. Google's a harsh mistress. Don't worry; this post will be on page 15.)

If I can go to the eighth page of search results and still find articles telling me how story points are about complexity and not time, then why are so many companies still using estimates as units of time?

"You're the consultant, Carlos! You should know," you're probably thinking.

And you're right. I should. But I really don't.

I've previously thought that this was an easy way for those with the purse-strings to better forecast their spending so they can know their budget burndown. But then I ran into groups that estimated based on time for no other reason than "because we always have.".

I've also previously thought that this was an artifact of scientific management that enforce and encourages managers to have tight control over their rank and file, or labor. But then I ran into engineers that were actually fine with time-based estimates, even if they had to work nights and weekends to uphold them.

So I don't know. My best guess is that when an authoritarian figure asks you, the engineer, "How long will this take?," responding with "about a Medium tee-shirt size" is a hard answer to give. But is that really better than saying "It'll be done tomorrow!" when you know that there's this bug that you've been wrestling with for the last sprint that Google has nothing on that might or might not be squashed before then?

Story Points Are Not Units Of Time

They really aren't.

The reason why story points exist is to acknowledge the fact that humans are really bad at estimating things.. We just are. However, we are much, much better at comparing things.

It's really easy to say that a feature is shippable in two days...until you hit a really nasty bug with datetimes that you didn't anticipate because you've only used that date/time library once or twice before. Or you get wrecked by your dependency manager (or, more commonly, a development environment), going down and hard-blocking you for a day.

On the other hand, it is much easier to say that a feature feels like a large effort because you've done projects that feel similar to it and, from that, can ascertain a better idea of how long it'll take to get done.

Why am I ranting about this? The motivation behind this is three-fold:

  1. I've seen so many projects slip or fail/get cancelled despite project managers having dedicated Marie Kondo-level efforts to updating Gantt charts, creating elaborate sub-tasks to stories, following up on variances and delicately managing "resource" time, amongst other efforts.

  2. I've also seen so many engineers (including myself) get beat up over deadlines and estimates that got blown due to factors outside of their control. In a world where every company is looking for the best engineers because they are so hard to find, why burn your engineers out this way?

  3. Trying to estimate on time is just so much work (and wasted labor costs) over using past experiences to judge the complexity of something.

If you've never done planning poker during your backlog grooming meetings, try it! It is fun, easy, makes engineers feel more highly valued and can actually improve the quality of your estimates over time (because we are better at comparing than precise estimating).

That is all.