Skip to content
Calcrivo

Commit Velocity Calculator

Calculate weekly commit velocity for a team or repository and compare it against a team or organization average.

Inputs

commits

Total commits made in the analyzed period.

weeks

Length of the analyzed window in weeks.

commits/week

Baseline average velocity to compare against (e.g. org-wide average).

Commit Velocity

40.00/week

vs. Team Average

Above average

Difference from Average

14.3%

Step by step

  1. Velocity: commits / weeks

    320 / 8

    = 40.00/week

  2. Difference from team average

    40.00 - 35

    = 5.00/week (14.3%)

How it works

Commit velocity tracks throughput over time. Formula: velocity = commits / weeks, benchmarked against a team or organization average to see whether a repository or contributor is pacing above, at, or below the norm. As with commit frequency, this is a volume metric, not a value or quality metric — use it alongside review and defect data.

Formula

velocity = commits / weeks

commits
Total commits in the period
weeks
Length of the period in weeks

Frequently Asked Questions

How is this different from commit frequency (per day)?

It's the same underlying idea at a different granularity — velocity uses weekly buckets, which smooths out day-to-day noise like weekends and is more common for sprint-based team comparisons.

Should I use commit velocity to evaluate individuals?

Be cautious — commit count varies hugely with style (squash vs. incremental commits), task type (a one-line config fix vs. a complex feature), and can be gamed if used punitively. It's better suited to repository/team trend analysis.

What causes velocity to differ across teams fairly?

Team size, codebase maturity (greenfield vs. legacy), and task complexity all legitimately shift velocity — always contextualize a comparison against team average with these factors.

How do I compute a team average baseline?

Aggregate commits per week across all repositories/teams in your org over a representative period (e.g. a quarter) and divide by team-weeks to get a fair per-team average.

You might also need