Activity metrics measure motion: tickets closed, test cases executed, hours logged, coverage percentages. Outcome metrics measure results: defect escape rate, cycle time, and the DORA four from Accelerate —
| Metric | What it tells you |
|---|---|
| Deployment frequency | How often you successfully ship |
| Lead time for changes | Commit to running in production |
| Change failure rate | Share of changes needing remediation |
| Time to restore | How fast you recover |
The research finding that makes these four special: they predict organizational performance better than team size, methodology, or process-maturity theater.
The trap with activity metrics isn’t just that they’re uninformative — it’s that they’re steerable. Any team under pressure can raise test-case counts without touching product risk. Outcome metrics resist gaming because they’re anchored to what actually happened in production.
Two qualifiers I want to keep honest about:
- Outcome metrics are lagging. You still need leading signals day-to-day — just don’t let the leading signals become the goal (Goodhart’s Law).
- A metric without psychological safety becomes a weapon, and weaponized metrics get gamed fastest of all.
Related:
- Mechanisms, not memos — outcome data is how you audit whether a mechanism works
- Small batches, fast feedback — the DORA four largely measure batch size and feedback speed
- Theory of Constraints — activity metrics reward polishing non-constraints