When it comes to Agile planning, our team's user story estimates aren't very accurate. How can we improve them?
It's the average of all work over many iterations that helps teams plan for future work.
In my experience with Agile planning, teams waste a lot of time trying to improve the accuracy of their story estimates. But it's more useful to focus on the average of all work across many iterations, than to get hung up on getting individual story estimates rights.
The primary value in estimating comes from the team's discussions with stakeholders about the purpose of each user story, how the business will measure the success of the feature once it's in production and how it can be tested.
A team's average velocity over time is a useful planning tool. It gives the development team and the business managers a realistic, if general, idea about how long it may take to complete certain themes or epics. So it's useful to give stories point or "T-shirt size" estimates (small, medium, large). Teams tend to work on large stories first, because they're usually higher risk. Small stories should be quick to complete, but since they're lower risk, they may get less attention and, as a result, be completed in about the same amount of time as the larger stories. It's the average of all work over many iterations that helps teams plan for future work.
This was first published in January 2013