GitHub Copilot Productivity Impact

Measuring the impact of GitHub Copilot helps you understand its effect on your team's productivity, quality, and overall development velocity. By comparing key performance metrics between Copilot users and a baseline cohort, you can identify meaningful differences in outcomes. The results can inform adoption strategies, help optimize usage, and pinpoint where Copilot delivers the most value.

Copilot Developers

The Copilot developers group includes any active developer with a seat in your organization.

• Users join this group when the seat is created
• Users are removed from this group if their seat is removed or if they do not have GitHub Copilot activity for more than 90 days
• When a user is removed from this group, their history in the Copilot group is retained
• Any user that had a license before GitHub was connected to Software.com is a member of the group from first adoption to when GitHub connected

Baseline Groups

To measure the productivity delta, we compare the performance of Copilot developers against a baseline group.

You can choose from three baseline options for your comparison:

• Not Copilot developers (default): This baseline compares Copilot developers to their peers who have not used Copilot in the last 90 days. This helps answer the question, “Are our Copilot developers more productive than developers not using Copilot?”
• Copilot developers, before adoption: This baseline compares Copilot developers against their own performance from before Copilot was first adopted by your organization (when the first seat was created). This helps answer the question, “Are our developers more productive with Copilot than they were before?”
• All developers, before adoption: This baseline compares Copilot developers against the performance of your entire organization from before Copilot was first adopted (when the first seat was created). This helps answer the question, “Are Copilot developers more productive than our organization was before we started using Copilot?”

When calculating the average performance before adoption, we do not include the period in which Copilot was adopted. For instance, if Copilot was first adopted by your organization on March 2, 2025, the 6-month period before adoption is September 2024 to February 2025.

When calculating the current average performance, we include the period in which Copilot was adopted. If Copilot was first adopted within the last 6 months, the current period is adjusted to start from when Copilot was first adopted. The current period also includes the current week, month, or quarter. Data is prorated for the in-progress period for New Deliveries per Developer and Cost per New Delivery.

Example Comparison Windows

Adoption Date: March 2, 2025, Today: August 12, 2025

Key Impact Metrics

We compare the productivity of your Copilot developers group to the baseline across the following key metrics:

• Productivity Impact: Measures the percent change in New Deliveries per Developer between the Copilot Developers cohort and your selected baseline.
• Adoption Rate: Calculated by dividing the number of Copilot Developers by the total number of developers in your organization over a 90-day rolling window.
• Other Impact Metrics: We also measure the performance delta for Lead Time, Rework, and Cost per New Delivery, comparing your Copilot cohort against the selected baseline.

Drill Down Into Usage

You can drill down into who is using GitHub Copilot during any given time period by clicking on a column in the Adoption Rate graph.

• Copilot status: whether the user is part of the Copilot developers group
• Last Copilot activity: the last Copilot activity recorded for that user (chat, autocompletion, etc.)
• Coplot license created: when the user’s Copilot license was created

Next, read more about other usage metrics we collect from the GitHub API and show on our Copilot Usage dashboard.