Northwind Engineering grades C
6 of 8 organs measured · targets: expert baseline · as of 2026-09-25
What’s actually wrong
3 confirmed problems · see all of yours →What’s genuinely working
The eight organs
sub-scores 0-100 · expert baseline targets · A ≥ 85 … F < 40How it’s scored
Weighted blend of 4 of 4 inputs = 32 → grade F (0-39).
- Pickup latency (p50)35%
- Pickup latency (p90)20%
- Cycle time (p50)30%
- Stale work-in-progress15%
How it’s scored
Weighted blend of 5 of 5 inputs = 63 → grade C (55-69).
- PR size (p50 lines)25%
- Review depth (comments / 100 lines)20%
- Rubber-stamp reviews20%
- Hotspot files20%
- Regression Rework Share15%
How it’s scored
Weighted blend of 2 of 3 inputs = 85 → grade A (85+).
- Bus-factor-1 paths50%
- Blast radius (max)35%
- Onboarding to 10th PR (p50 days)15%
Only 85% of the designed weight is observed - the blend renormalizes over measured inputs, so their weights add to 100%.
How it’s scored
Weighted blend of 3 of 3 inputs = 73 → grade B (70-84).
- PR-ticket linkage50%
- Zombie tickets25%
- Orphan PRs25%
How it’s scored
Weighted blend of 3 of 3 inputs = 84 → grade B (70-84).
- Review-load Gini40%
- Top-person load share35%
- After-hours work25%
How it’s scored
Weighted blend of 1 of 2 inputs = 23 → grade F (0-39).
- Deploy frequency (per week)young60%
- CI duration (p50)40%
Only 60% of the designed weight is observed - the blend renormalizes over measured inputs, so their weights add to 100%.
⚠ deploy frequency per week is a young signal - its observable history is shorter than the window it claims; treat with care
What moved
Your org, read the same way.
Connect GitHub and the first verdict lands in minutes.