Review quality trend: watching rigor drift before it becomes a problem

Review Quality Trend plots two lines, week by week or month by month: the % of merged pull and merge requests with a genuine peer approval, and the % with a real written comment. Same genuine-peer-review logic as the Review Quality Index scorecard — an approver or commenter who isn't the author — just spread across time instead of collapsed into one number. This is the widget that catches review quality sliding before anyone notices it in a single week.

Coverage and comment rate, both steady over eight weeks.

What it measures, and where the data comes from

Every merged pull or merge request in the window, cohorted by the period it was opened in, bucketed weekly or monthly. Two lines come out of the same cohort:

• Peer-approval coverage % — share with at least one approval from a reviewer other than the author.
• Comment % — share with at least one written comment from someone other than the author, distinct from a silent approval click.

Sourced from GitHub, Azure DevOps, and GitLab — merged work only, so a period with a lot of abandoned or still-open PRs won't dilute the reading.

How to read it

Watch the two lines relative to each other, not their absolute height. Two shapes matter most: both lines drifting down together, and coverage staying flat while comment % falls. They mean different things, and they call for different responses.

Both down together is the broader signal — review rigor eroding across the board, not just one dimension of it. Coverage flat, comment falling is narrower and sneakier — approvals are still technically happening on schedule, they're just getting shallower. The click survives; the substance underneath it quietly leaves.

What it tells you over time

A single bad week on either line is noise — a holiday week, one release crunch, a reviewer out sick. Three or four consecutive periods moving the same direction is a trend worth naming out loud in a retro. Also watch for a step change that lines up with something you already know happened — a headcount change, a new deadline, a reorg — that's usually the actual cause, not "the team got lazy."

Example situations

1. Both lines sliding down together

Coverage and comment % both trending down over eight weeks.

What you're seeing: both lines heading the same direction, steadily, for two months straight. This isn't one team's bad sprint — it's a broad erosion of review rigor, and it's the kind of pattern that tends to show up right when a whole org is under the same kind of pressure at once.

How to react: don't respond with a review-process memo before you understand the cause. A synchronized decline across coverage and comments almost always traces back to something structural — a compressed release calendar, a hiring freeze stretching everyone thinner, a big deadline looming over multiple teams.

2. Coverage holding, comments quietly disappearing

Coverage % flat; comment % sliding underneath it.

What you're seeing: the coverage line is nearly flat — approvals are happening exactly as reliably as they were two months ago. The comment line is not; it's falling steadily underneath. Every metric that just counts "was there an approval" would look perfectly healthy here. That's the whole problem with reading coverage alone.

How to react: this is silent approval creep — reviewers still clicking approve on schedule, but reading less carefully, or trusting the author more than the diff warrants. It rarely gets flagged by anyone because the process, on the surface, looks like it's working.

Related widgets

• Review Quality Index — the full five-row scorecard for any single window this trend flags.
• Reviewer Participation — checks whether a falling comment rate traces back to review load concentrating on fewer people.
• DORA Metrics — review rigor eroding often shows up a few weeks later as a softer change-failure-rate proxy.
• Back to the widget reference.

Last updated 17 August 2026