Dataset: Issues & Pull Requests
Entity: Pull Requests, Issues
Field ID: first_peer_comment_at
Type: Datetime
Description: The datetime of the first comment made by someone other than the author, excluding bots (e.g. CodeRabbit, Dependabot). Use it to measure the time to first human feedback.
Use this field, not First comment at, for time to first review. First comment at counts every comment, including the author's own and those of AI reviewers and scanners, which often comment within minutes of a pull request being opened. On a team using such a bot, it measures the bot rather than the team.
Source: Calculated
Transformation logic: It is the creation datetime of the first peer comment on the issue or pull request. A peer comment is a comment:
whose author is known;
whose author is not the author of the issue or pull request. Accounts are compared, not displayed names, so two people sharing a name are not confused;
whose author is not a bot, as Author type defines it: from the application's own account type and from a list of known bots maintained by Keypup.
Comments left as part of a pull request review count. Comments copied from another tool by a sync app (e.g. Exalate or Unito on Jira) do not: they are posted under the sync app's account, at the time of the sync, and credit no one on your team. The field is empty when the item has no peer comment.
App Mapping
GitHub (PRs, Issues) | Calculated |
GitLab (MRs, Issues) | Calculated |
Bitbucket (PRs) | Calculated |
Azure DevOps (PRs, Issues) | Calculated |
Jira (Issues) | Calculated |
ClickUp (Issues) | Calculated |
Trello (Issues) | Calculated |
Reporting Use Cases
The first_peer_comment_at field marks when someone else on the team first engaged with a piece of work. It sits on the same row as the item's other dates, so it can be combined with them in one insight.
Time to first human feedback: the custom formula
(first_peer_comment_at - created_at) / HOUR()gives the hours between a pull request being opened and its first comment by a teammate. Filter ontype == "pull_request"andIS_NOT_NULL(first_peer_comment_at), so that pull requests nobody commented on are not counted as zero.Trend over time: a line chart with
YEAR_MONTH(merged_at)as the dimension andMEDIAN((first_peer_comment_at - created_at) / HOUR())as the metric, over merged pull requests, shows whether reviewers pick up new work faster or slower. The median is less sensitive than the average to the few pull requests that wait for weeks.First review action, comment or approval: a pull request can be approved without a single comment.
IF_NULL(first_peer_comment_at, first_review_created_at)takes the first peer comment, or the first review when nobody commented, and(IF_NULL(first_peer_comment_at, first_review_created_at) - created_at) / HOUR()gives the time to first review action.Work waiting for a reviewer: open pull requests with
IS_NULL(first_peer_comment_at)have received no human feedback yet, even when a bot has already commented on them.
