On this page:
What is Estimate Accuracy?
It's not uncommon for work to take a lot longer than planned. It’s probably even more uncommon for work to be completed faster than planned. Even a slight variance to what was estimated can compound to derail an entire release, as well as disappoint stakeholders when actual results were different from original expectations.
We created Estimate Accuracy to help teams build greater predictability in relation to the amount of work they can complete, and more accurately forecast future productivity.
Estimate Accuracy is the comparison of the estimated time and the actual time taken to complete an issue. For story points, this refers to the average time take for each story point value.
Estimate Accuracy is the actual cycle time of an issue relative to the estimate (median cycle time of similar sized issues in the past 3 months). 'Estimate' is displayed for issues where time to complete is as planned. Issues taken faster than planned are displayed as being above 'Estimate' and issues taken longer than planned are displayed as being below 'Estimate'.
Why does Estimate Accuracy matter?
Estimate Accuracy provides teams an indication of how accurate their estimation practices are. This helps identify your team’s planning patterns. For example, whether they consistently overestimate or underestimate how long an issue might take or how complex an issue might be. Teams can also identify anomalies in their estimation practices. For example, the usual cycle time of an issue sized 3 usually takes their team 8 days to complete, whereas issues sized 8, usually take 3 days to complete. This helps teams to remediate estimation anomalies, plan more accurately, improve predictability and balance more consistent levels of throughput and rates of completion.
What does a 'good' Estimate Accuracy looks like?
Estimate Accuracy is presented as a percentage value relative to zero, where zero indicates there being no difference between your actual and estimated cycle time of issues. In other words, a zero indicates a perfectly accurate estimate.
The further away from zero you are, the less accurate you are. For example, the further above zero for your estimate, the faster your actual cycle time is relative to your estimate. The further below zero you are, the longer your actual cycle time is relative to your estimate.
How does TeamX measure Estimate Accuracy?
Estimate Accuracy is measured from the actual cycle time of an issue, relative to the median cycle time of an estimate of similar sized issues during the past three months.
A zero displayed in TeamX indicates time to complete an issue is as estimated.
Issues taking more time than estimated are displayed as being a negative percentage value. For example an issue that took two times slower than estimated is plotted on the chart as -200% below 0.
Issues taking less time than estimated are displayed as being a positive percentage value. For example, an issue that was completed two times faster than estimated is plotted on the chart as +200% above 0.
For each interval, TeamX surfaces the actual vs estimated cycle time for each completed issue, sorted by the order of magnitude in the variance.
TeamX also displays whether there are any patterns a team needs to be made aware of relating to how they estimate. For example, TeamX will surface if issues sized as being 3 story points are consistently taking longer than issues sized as 5 story points.
What drives Estimate Accuracy?
Activities that contribute to and influence Estimate Accuracy include the estimate value (being story points or time) for all issues completed within an interval. Also included is the median cycle time of similar sized issues completed in the past three months.
Activities that do not contribute to your Estimate Accuracy are all issues in ‘to-do’ or ‘in-progress’ status.
Estimate Accuracy is also influence by:
Total number of issues in the interval
Number of issues completed
Median story points of issues
Median time estimate of issues
Number of issues inherited
Median cycle time of issues
Number of issues added mid-interval
Tips for improving your Estimate Accuracy.
Identify any repeated blockers in your workflow that may be impacting time taken to complete an issue. TeamX's Time in Status and Stability will help to understand where issues taking longer than planned are getting stuck in your workflow
Refer to how long a story point value usually takes your team to complete, and apply this to how many issues you assign to the interval to be completed. You might also refer to TeamX's Efficiency Rate to understand the usual number of issues that can be completed per team member, and get a more accurate prediction of what can be completed.
Be sure to break down larger sized issues into smaller batches of work. This will help build a greater understanding within the team of what’s required, and remove as much variability as possible
If you’ve over-planned your interval and work is taking longer than planned, maintain momentum and focus by using your stand-up or check-in sessions to move lower priority issues back to the backlog
Be sure to keep you backlog maintained and prioritised for instances when you’ve completed issues faster than estimated and it’s clear what is the next issue to commence work on.