DORA (DevOps Research and Assessment, now part of Google Cloud) has studied software delivery for over a decade. Its four classic delivery metrics are the most widely used language for engineering performance:
- Change lead time: from code committed to code running in production.
- Deployment frequency: how often you deploy to production.
- Change failure rate: the share of deployments that cause a failure needing remediation.
- Failed deployment recovery time: how long it takes to recover from a failed deployment.
The first two measure throughput, the last two stability. DORA's research has found that the two go together: good teams are fast and stable.
The 2024 performance clusters
Each year DORA groups survey responses into clusters. In 2024:
| Level | Lead time | Deploy frequency | Change failure rate | Recovery | Share |
|---|---|---|---|---|---|
| Elite | < 1 day | On demand (multiple per day) | 5% | < 1 hour | 19% |
| High | 1 day – 1 week | Daily – weekly | 20% | < 1 day | 22% |
| Medium | 1 week – 1 month | Weekly – monthly | 10% | < 1 day | 35% |
| Low | 1 – 6 months | Monthly – every 6 months | 40% | 1 week – 1 month | 25% |
Notice the medium cluster's failure rate is lower than the high cluster's: DORA itself says improving matters more than reaching a label, and "the best teams are those that achieve elite improvement, not necessarily elite performance."
Why DORA matters more with AI
AI changes the input (more code, faster). DORA metrics measure the output of the system. If AI adoption rises while lead time stays flat and change-failure rate rises, AI is not helping delivery yet — DORA's own 2024 data found exactly this risk at the aggregate level.
Practical tips
- Use the definitions consistently; document how your tools compute them.
- Compare a team with itself over time, not teams with each other.
- Add AI-specific context (adoption, AI-authored share) next to DORA, not instead of it.