Engineering Delivery Metrics
AI writes the code.
Review is your new bottleneck.
AI tools have doubled PR volume — but review time has grown just as fast. FlowMetrics shows where work is piling up, how you compare with teams your size, and which pull requests are stuck right now. Straight from GitHub, no spreadsheets required.
Reads pull request metadata only. We cannot read your source code.
acme/api · Overview
last 14 daysBottleneck Awaiting review holds 41% of lead time, and time to first review is slower than 81% of teams your size.
Lead time (p50)
1d 6h
DORA high
First review (p50)
14h 24m
slower than 81% of peers
Throughput
68
+9 vs last period
Stalling now
- #482 Add rate limiter Approved · 2d 4h
- #489 Warm the cache on deploy Awaiting review · 1d 9h
- #477 Fix retry loop In review · 1d 6h
Illustrative data
"Your app just replaced a spreadsheet I spent 40+ hours building."
- CTO
Compared with 38 teams with 6–15 authors
last 28 daysIllustrative data
New · Peer Benchmarks
Know where you stand.
Not just where you've been.
Your trends tell you whether you're improving. Benchmarks tell you whether you're good. FlowMetrics compares your lead time, review wait, approval time, PR size, change requests and throughput with teams your size, so "your review wait is slower than 80% of peers" is a fact, not a hunch.
- Teams like yours. Peers are grouped by how many people actively ship, from 1–5 authors to 51+. Peer groups start with active open-source projects and grow with every team that joins.
- Built into your summary. When a metric lands in the slowest quarter of your peer group, it shows up in your summary and next to the metric itself.
- Anonymous by design. We only publish totals that combine at least 10 repositories from 5 organizations. Never names, people or code. How it works
The one comparison you can't build in-house.
A homegrown dashboard only ever sees your own repositories. It can tell you your review wait, but never whether it's normal. FlowMetrics peer groups combine teams across organizations, and every team that joins makes everyone's comparison sharper.
Feature
Find the friction,
not the slow developer.
Teams using AI coding tools are shipping nearly twice as many PRs — but review time has grown at the same rate. FlowMetrics tracks every PR from open to merge — time to first review, time to first approval, and lead time — and the cumulative flow diagram shows exactly which stage work is piling up in. Fix the system, not the people.
Cumulative Flow
Illustrative data
Value stream
Where merged PRs spent their time · 128 PRs · rework rate 14% · flow efficiency 32%
| State | Type | Median | Share |
|---|---|---|---|
| Open | Work | 6h | 18% |
| Awaiting review Likely bottleneck | Wait | 21h | 41% |
| In review | Work | 5h | 14% |
| Approved | Wait | 9h | 27% |
Illustrative data
New · Value Stream Mapping
See where every hour
of lead time goes.
Map your team's real workflow using the labels, drafts, reviews, and approvals you already use in GitHub. FlowMetrics walks every merged PR through it and shows how long work sits in each state — and which of that time is active work versus waiting in a queue. The state eating the biggest share of lead time is flagged as your likely bottleneck.
- Your workflow, not ours. Define states like "Awaiting review" or "QA" and the GitHub events that move a PR into them.
- Work vs. wait. Median and P75 time in every state, plus flow efficiency: how much of lead time is spent actually moving.
- Rework, made visible. See how often PRs bounce back to an earlier state, and where they land.
Up to
90
days of baseline data
Ready in minutes after install
Feature
Up to 90 days of baseline data.
Ready in minutes.
No manual tagging or waiting for data to populate. The moment you install the GitHub App, we backfill up to 90 days of history (60 on Personal) and begin calculating P50 and P75 lead and review times for each repository over 7, 14, 28, and 84-day windows. P75 is where slow PRs hide. Averages aren't.
Feature
Your metrics and
your stuck PRs, in Slack.
Get a digest of your team's key delivery metrics posted directly to the Slack channel of your choosing — so the people who need the numbers don't have to go looking for them.
Turn on stalled-PR nudges and every weekday morning FlowMetrics lists the pull requests that have sat in their current stage longer than 3 in 4 of your recent PRs, with each author named. The threshold comes from your own value stream, so there's nothing to configure, and nothing is posted when nothing is stuck.
📊 FlowMetrics Weekly Digest
Lead time (p50): 1.4 days · Throughput: 34 PRs · Review time (p75): 3.1 days
⏳ 2 PRs are stalled in acme/api
- #482 Add rate limiter · alice · Approved for 2d 4h (usually ≤ 9h)
- #477 Fix retry loop · carol · In Review for 1d 6h (usually ≤ 20h)
More code isn't the problem.
Invisible bottlenecks are.
FlowMetrics measures the system — not the individual.
How it works
Connect your GitHub account
Sign in with GitHub and install the FlowMetrics GitHub App on your organization or specific repositories. Takes about 60 seconds.
Always up to date
GitHub webhooks record each pull request as it opens and merges, and a nightly sync adds review history. No manual exports, and nothing to change in your CI/CD pipeline.
Act on it
Your summary flags where you trail teams your size, and stalled-PR nudges name the stuck pull requests in Slack every weekday morning. Lead time, throughput, WIP and value stream stay current for every repository.
Simple, volume-based pricing
Never pay per seat. Free for personal accounts.
Pro
$49/mo
500 PRs/mo · 90-day history
Peer benchmarks · Slack digest and stalled-PR nudges
Start free trialNeed more? See full pricing including Enterprise →
Stop guessing. Start measuring.
Connect your GitHub account and have answers in minutes.
Generate Your Dashboard →Reads pull request metadata only. We cannot read your source code.