AI Adoption Score — max 100
Same formula for all roles — AI adoption is expected equally from everyone.
| AI commit ratio × 0.40 | max 40 pts |
| PRs with AI commit × 0.30 | max 30 pts |
|
| Tool bonus — Claude | +20 pts |
| Tool bonus — Copilot | +10 pts |
| Tool bonus — Cursor / other | +0 pts |
|
| Trend bonus (+2 to −2 scale) | max 10 pts |
trend_bonus = (trend + 2) / 4 × 10
Velocity Score — max 100
ICs are rewarded more for shipping code — Managers more for reviewing and unblocking others.
| Sub-weights vary by role |
| Component | IC | Manager |
|---|
| PRs merged / mo ÷ 40 | 35 pts | 20 pts |
| PRs reviewed / mo ÷ 60 | 30 pts | 45 pts |
| Review speed (P75 h, target 4h) | 20 pts | 20 pts |
| Approval rate % | 15 pts | 15 pts |
speed = (1 − (p75h − 4) / 44) × w_speed, clamped [0, w_speed]
approval = (rate / 100) × w_approval
Engineering Excellence Score
Weighted composite of all available dimensions. Components are only included when data is present — the denominator adjusts automatically.
| Component | IC | Manager |
|---|
| AI Adoption | 55 | 25 |
| Velocity | 35 | 60 |
| Collaboration (when available) | 15 | 15 |
| Code Output (when available) | 10 | 5 |
excellence = Σ(score × weight) / Σ(weights of present components)
Collaboration Score — max 100
From Slack analytics. Only shown for engineers matched by GitHub login.
| Messages posted ÷ 500 | max 70 pts |
| Days active ÷ 20 | max 30 pts |
Code Output Score — max 100
From GitHub stats. Only shown when lines and PR data is available.
| Lines added ÷ 5000 | max 40 pts |
| PRs merged ÷ 20 | max 30 pts |
| PR size discipline | max 30 pts |
|
| PR size weights (higher = better) |
| XS (≤10 lines) S (≤100) | 1.0 |
| M (≤500 lines) | 0.7 |
| L (≤1000 lines) | 0.3 |
| XL (>1000 lines) | 0.2 |
discipline = Σ(weight × count) / total_prs
Claude Code Score — max 100
Based on Anthropic Admin API data (last 2 months). Independent from Excellence score.
| Edit acceptance rate × 35 | max 35 pts |
| Sessions ÷ 40 × 40 | max 40 pts |
| Lines added ÷ 2000 × 25 | max 25 pts |
Cursor Score — max 100
Based on Cursor Admin API data (last 30 days). Independent from Excellence score.
| Tab acceptance rate × 35 | max 35 pts |
| Total requests ÷ 200 × 40 | max 40 pts |
| Accepted lines ÷ 2000 × 25 | max 25 pts |
Grade Scale
| A+ — elite | ≥ 90 |
| A — strong | ≥ 80 |
| B — solid | ≥ 70 |
| C — developing | ≥ 55 |
| D — early | ≥ 40 |
| F — needs focus | < 40 |