The median active reviewer submits 5 reviews per week
  • The median active reviewer submits 5 reviews and reviews 3 PRs each week.

    • We exclude popular end-of-year weeks when code freezes are common.

    • By taking the median of medians, we hope to reflect the most common developer experience.

  • Developers who review more are also likely to merge more code.

  • 24% of active reviewers are not active committers; this means that they merge, on median, 0 PRs a week. Dedicated active reviewers are more common in larger companies.

    • Following our second point, active reviewers who rarely ship code still review less frequently than developers who both review and ship code.
  • As we see below, there is a strong positive correlation between developers’ PR merge frequency and review submission frequency.

Reviewer merge relation

  • 23.8% of active reviewers (≥ 1 median review per week) are not active committers (0 median PRs merged per week).

    • Perhaps more surprisingly 7.8% of active committers rarely review and do not qualify as active reviewers.

Venn

  • Smaller companies tend to skew higher on “only merge” developers while larger orgs have more “only review” developers

Venn by org size

  • Active reviewers who are not actively merging PRs (0 median PRs merged per week) still submit fewer reviews compared to developers who are actively shipping code.

Review only vs both