DeployMeter

Live demo · example data

Northwind

Northwind is a fictional organization: 12 developers across 3 products. Your dashboard shows your own pull requests.

AI Multiplier
3.4x
across 3 products · P25–P75 2.1–5.6x

estimate · 67% measured

5%
AI Time Saved
2.1person-months
range 1.5–2.8 person-months

estimate · 67% measured

0%

What changed

Sep 18 – Oct 18, 2026 · compared with Aug 19 – Sep 18
Moves of more than 5% (or 5 pt) against the previous period

Improved

  • Time Saved

    1.7 person-months → 2.1 person-months vs previous period

    +25% better

  • Review Turnaround

    10.6h → 8.0h vs previous period

    −24.5% better

  • AI Multiplier

    2.9x → 3.4x vs previous period

    +17.2% better

Needs attention

  • Unreviewed PRs

    5% → 11% vs previous period

    +6 pt worse

DORA metrics

Deployment Frequency
6.10/day
+13.0% better
Elite

183 deploys · 30d

Lead Time
3.1h
−11.4% better
EliteP50 3.1hP75 14hP90 28h

to deploy

Change Failure Rate
5.1%
−10.5% better
High
Failed Deployment Recovery
6.5h
−11.0% better
MediumP50 6.5hP90 22h
Rework Rate
8.0%
−11.1% better
High

derived from: GitHub / GitLab Deployments, GitHub issues (label: incident)

Trend by product

Direction over the period — per-product daily medians. Tap a row to expand.

Last 30 days. Deploys per day and the rates are period totals (failures ÷ deploys); lead time and recovery average the daily medians. For “All products”, lead time and recovery are the whole scope's median for the period.

DORA performance bands

Where each product sits today against industry Elite / High / Medium / Low thresholds.

Focus
Better ←
EliteHighMediumLow
All products
Deployment Frequency (DF)
6.1/dElite
Lead Time (LT)
3.1hElite
Change Failure Rate (CFR)
5.1%High
Failed Deployment Recovery (FDRT)
6.5hMedium
Rework Rate (Rework)
8.0%High
Incidents2none open
Uptime99.9%no open SEV1 incident
View incidents →

derived from: GitHub / GitLab Deployments, GitHub issues (label: incident)

Human vs AI contribution
Agent-authoredHuman, AI-assistedHuman only
PRs merged75% with AI · 99 / 132
Commits32% with AI · 251 / 779
Lines changed69% with AI · 17,688 / 25,608

Agent-authored is a floor: only accounts classified as agents count. A commit is AI-assisted when it carries an AI co-author trailer; a PR, with all its lines, when at least one of its commits does.

Product comparison

Every product side by side across DORA, impact, volume, and quality.

  • All productsHigh
    Deploy Freq6.1/day
    Lead time3.1h
    CFR5.1%
    FDRT6.5h
    Rework8.0%
  • CheckoutElite
    Deploy Freq4.2/day
    Lead time3.1h
    CFR3.1%
    FDRT0.9h
    Rework6.0%
  • Mobile AppHigh
    Deploy Freq1.4/day
    Lead time18.0h
    CFR7.8%
    FDRT6.5h
    Rework11.0%
  • Data PlatformMedium
    Deploy Freq0.5/day
    Lead time62.0h
    CFR14.0%
    FDRT22.0h
    Rework17.0%
AI

Insights

Based on last 30 days of team activity

Updated 2d ago

Checkout leads on AI multiplier

aiMultiplier

Checkout's median multiplier is 4.1x, the highest of the three products, driven by short, well-scoped PRs with measured coding time across most of the team.

Action:

Share Checkout's PR sizing habits with Data Platform, whose PRs average 3x more lines.

Data Platform lead time is 20x Checkout's

leadTime

Median lead time is 62h against 3.1h. The gap is review wait, not coding: first review lands after 21.7h on average.

Action:

Set a review turnaround goal for Data Platform and route reviews to a second reviewer after 8h.

Fewer deployments failing

changeFailureRate

5.1% of deployments failed or were attributed an incident, down from 5.7% in the previous period. The 2 SEV2 incidents were on Mobile App and Data Platform.

Action:

Keep the current release cadence; watch Mobile App's post-release checks.

Measured time on two thirds of developers

timeSaved

8 of 12 developers send measured coding time, so the time-saved band is narrow. The remaining 4 are on Data Platform.

Action:

Run npx deploymeter init on Data Platform's repositories.

Data Platform carries the most rework

reworkRate

17% of Data Platform's deployments are rework, against 6% on Checkout — concentrated in two files touched by every release.

Action:

Schedule a refactor of the two hotspot files before the next release train.