<- all tokdocs

How Long Do Pull Requests Wait for Review? The Two-Day Median, Checked

By , founder of Mavric and Apso

View on TikTok ->

A pull request from a colleague that draws review comments waits a median of 2,881 minutes, or 48 hours, according to a study of 3.3 million GitHub pull requests. That matches the "sits for two days" in a TikTok from Danielle (@danielleintech), who argues code review should come first each morning and that review depth should match risk. Her fixes line up with the research, though the strongest cause of the delay is slow first response, not PR size.

How long do pull requests take to get reviewed?

The largest dataset is "Pull Request Latency Explained: An Empirical Overview" by Xunhui Zhang, Yue Yu, Tao Wang and Huaimin Wang (2021), covering 3,347,937 pull requests from 11,230 GitHub projects. Its median latencies:

  • Commented PR from someone other than the merger: 2,881 minutes (48.0 hours)
  • Commented PR merged by its own author: 1,348 minutes (22.5 hours)
  • Uncommented PR from someone else: 311 minutes (5.2 hours)
  • Uncommented PR merged by its own author: 30 minutes

The video's case, a substantial PR from a colleague that needs discussion, sits in the first row. Two days is the population median.

Healthier projects are faster. "Wait For It" by Yu, Wang, Filkov, Devanbu and Vasilescu (MSR 2015) studied 103,284 pull requests from 40 projects using Travis-CI and found a "median first comment time is 16min" and a "median 11.2h" closing time. That sample leans toward large, well-staffed repositories.

Is a 30-second approval vs a two-day wait a fair comparison?

Not quite. The video says "One line change approved in 30 seconds, but the 400 line PR that actually needs eyes sits for two days." Thirty seconds is reviewer attention. Two days is calendar time, most of it spent waiting in a queue.

Compared on the same clock, the gap is smaller. Google's study "Modern Code Review: A Case Study at Google" by Sadowski, Söderberg, Church, Sipko and Bacchelli (ICSE-SEIP 2018) found developers "wait for initial feedback on their change a median time of under an hour for small changes and about 5 hours for very large changes." The overall median for the whole review process was under 4 hours. Developers spent a median of 2.6 hours a week reviewing, across a median of 4 reviewed changes, about 40 minutes each.

What causes slow code reviews?

Size matters, but discussion and first response matter more. "Wait For It" found "the number of comments is the best single predictor of the latency," with PR size "a positive, strong predictor." Size, churn and discussion length together explained 67% of the explained variance. Its conclusion: "the shorter it is the faster it will be reviewed."

Zhang and colleagues go further: "the first response time of the reviewer is most important when there exist comments." For commented pull requests, first response time ranked first among all factors.

The video blames the 400-line size, but its own first fix targets first response time, the stronger lever.

How fast should you review code? Google's rule

The video's first fix is "review before you open your own branch each morning." Google's Speed of Code Reviews guide shares the reasoning: "we optimize for the speed at which a team of developers can produce a product together, as opposed to optimizing for the speed at which an individual developer can write code."

Google's rule is a ceiling: "One business day is the maximum time it should take to respond to a code review request (i.e., first thing the next morning)." It also lets reviewers finish focused work first. If you are mid-task, "wait for a break point in your work before you respond." The promise covers response time, "as opposed to how long it takes a CL to get through the whole review and be submitted." The video's morning rule is a stricter version of the same idea.

Should review depth match the risk of the change?

The video's second fix is "match scrutiny to risk, rubber stamp the config bump, but go deep on the auth change." (The TikTok caption misprints "match" as "max.")

Google's study supports the first part. Of eight surveyed developers who called review comments on their change unhelpful, 3 explained that the changes under review were "small configuration changes" where review mattered less. Only 2 said the comments found a bug. Google's What to Look For page also requires qualified reviewers for "complex issues such as privacy, security, concurrency."

"Rubber stamp" goes further than Google's guidance. Its code review standard says reviewers "should favor approving a CL once it is in a state where it definitely improves the overall code health of the system being worked on, even if the CL isn't perfect," and also "Don't accept CLs that degrade the code health of the system." The guidance is to stop demanding polish, not to stop reading. Config changes can reach production without touching tested code paths, so they still deserve a look.

How big should a pull request be?

The video picked 400 lines as the PR that "actually needs eyes." SmartBear's write-up of a Cisco Systems study recommends reviewing "no more than 200 to 400 lines of code (LOC) at a time," says "a review of 200-400 LOC over 60 to 90 minutes should yield 70-90% defect discovery," and reports a "significant drop in defect density at rates faster than 500 LOC per hour."

Google aims smaller. Its Small CLs page says "100 lines is usually a reasonable size for a CL, and 1000 lines is usually too large," and that small changes are "Reviewed more quickly" and "Reviewed more thoroughly." Google's study found the median change modifies 24 lines, and over 10% change a single line. A 400-line PR is about seventeen times that median.

Bottom line

Two days is the real median wait for a commented pull request from a colleague. The fastest fix is a quicker first response, which is what reviewing first thing each morning delivers. Splitting large PRs into smaller ones and scaling review depth to risk both have research support, as long as low-risk changes still get read.

Sources

Published October 22, 2026. Based on a TikTok by @danielleintech. Full video notes (markdown)