In 30 seconds
- Deployment frequency and lead time for changes measure speed.
- Change failure rate and time to restore measure stability.
- Research shows speed and stability go together.
The four metrics
| Metric | Measures |
|---|---|
| Deployment frequency | How often you deploy |
| Lead time for changes | Commit to production time |
| Change failure rate | Share of changes causing failures |
| Time to restore service | How fast you recover |
Where they come from
The DevOps Research and Assessment (DORA) programme, led by Nicole Forsgren with Jez Humble and Gene Kim, published the research in Accelerate (2018). Google acquired DORA in 2018 and continues the annual State of DevOps reports, which have since added further measures such as reliability.
Using them well
- Track trends for your own team, not league tables
- Look at all four together, so speed doesn't trade off stability
- Use them to start improvement conversations
Myth vs reality
Myth · Speed and stability are a trade-off.
Reality · DORA research found high performers do well on both.
Myth · DORA metrics measure developer productivity.
Reality · They measure the delivery system, not individuals.
Myth · Targets improve the metrics.
Reality · Targets invite gaming; use them for learning.
The AI angle
DORA metrics are a useful check on AI adoption: if deployment frequency rises while change failure rate and recovery time worsen, speed is coming at the cost of stability.
FAQ
How do we collect them?
From version control, CI/CD and incident tools; many tools automate it.
How often should we review them?
Monthly or quarterly trends.
Do they apply to non-web products?
Yes, adapted to how the product is delivered.