Agile Estimation Woes – Is your Team Imprecise or Inaccurate?

arrow-and-target

Uncertainty Isn’t Necessarily a Problem

When I work with Agile teams that are struggling to meet their sprint commitments, they often blame the fact that there’s too much uncertainty with their stories for them to be able to meet their commitments.  There’s a lot of ways that the team can address this problem – the most direct is to perform further analysis to reduce uncertainty in the story.

However, even if we take story uncertainty as a given (after all, the Agile Manifesto includes that teams should welcome changing requirements, even late in the development), the team should still be able to make their commitments if they are pointing accurately.  The uncertainty will reduce precision, but if the team is accurate, they should still have a good chance of completing their committed stories.

Accuracy vs. Precision

Accuracy refers to how close a measured value is to the known value [1].  For story estimation, we’ll refer to accuracy as how close a story estimate is to the true size of the story, as determined after the work has been completed.

Precision, on the other hand, simply refers to how close multiple measurements are to each other [1].  A team could be precise if they’re consistently sizing stories as 5 points that end up truly being 8 point stories.

If a team is struggling to meet their commitments, the team may have been precise, but not accurate. Likely, the team is consistently underestimating the size of the stories, so more of the stories end up being larger than expected.

precise-but-not-accurate

When uncertainty is high, we want to strive for the team to be accurate, even if they aren’t precise.  If accuracy is high, even if the estimates are wrong, the estimates should be centered around the actual story size.  Thus, some stories will be wildly overestimated, and some stories will wildly underestimated, but these differences will come out in the wash over the course of the sprint, and the team will be able to meet its commitments.

accurate-but-not-precise

Eventually, our goal is for the team be accurate and precise:

accurate-and-precise

Recalibrating an Inaccurate Team

We’ll focus on recalibrating a team that’s regularly underestimating their story size.  I very rarely encounter a team that’s regularly overestimating – if they are, they typically like to keep that quiet!

Don’t Assume Happy Path

Some teams make the mistake of assuming everything will go perfectly, or nearly perfect.  Most people know about Murphy’s Law, which states where anything that can go wrong will go wrong.  Despite knowing this, teams I’ve worked with tend to be overly optimistic when planning for upcoming sprints, and plan for the happy path, or close to it.  Make sure your team is accounting for the average size of the story, not the best case scenario.

Include Risk

Part of your story sizing should include the risk associated with the story. It’s important for teams and stakeholders to understand that this is different than padding a story.  Padding would lead to a consistent overestimation – including risk just helps to make sure the team is accurate in the face of uncertainty.

Don’t Worry About Being Right – Be Accurate Instead

Teams typically focus on getting each story estimate correct.  After all, if all of the individual stories are estimated correctly, the whole sprint will go smoothly!

This line of thinking works well for mature teams that are successfully getting their estimates correct.  For teams that are struggling to complete a sprint, it can help to remove the focus from each individual story, and focus on being accurate over the course of the sprint.  Even if some stories end up being overestimated and some are underestimated, that’s ok if the entire sprint is successful.

Journey toward Precision

Precision is something that the team will improve on over the course of multiple iterations, but the team needs to be accurate first. Separating out accuracy from precision can help the team focus on a smaller chunk of improvement, which can help the team get over the mental hurdle of the difficulty of making all of their commitments.  Once the team is keeping their Product Owner happy by keeping their commitments, there will be a greater opportunity to focus on precision.

References

[1] https://www.ncsu.edu/labwrite/Experimental%20Design/accuracyprecision.htm

Advertisements

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s