DORA Metrics

Full Definition

DORA metrics are four software delivery performance indicators derived from years of research by Dr. Nicole Forsgren, Jez Humble, and Gene Kim (published in "Accelerate: The Science of Lean Software and DevOps") that identify the highest-predictive metrics of software delivery capability and organizational performance. The four metrics—Deployment Frequency, Lead Time for Changes, Mean Time to Restore (MTTR), and Change Failure Rate—form a comprehensive measurement framework that balances throughput (how fast can you ship?) with stability (how reliably do you ship?). DORA research demonstrates that these four metrics predict both IT performance and organizational performance, including profitability, market share, and productivity. DORA research segments organizations into four performance tiers: Elite (multiple deployments per day, less than 1 hour lead time, less than 1 hour MTTR, 0-15% change failure rate), High (between once per day and once per week deployment, 1 day to 1 week lead time, less than 1 day MTTR, 16-30% change failure rate), Medium (between once per week and once per month), and Low (between once per month and once every 6 months, more than 1 month MTTR). Critically, DORA research disproves the intuitive trade-off between speed and stability: elite performers are simultaneously faster and more reliable than low performers. The explanation is that frequent small changes are easier to test, deploy, and roll back than infrequent large changes, creating a positive feedback loop where speed and reliability reinforce each other. Deploying DORA metrics as an organizational measurement framework requires accurate data collection for each metric, which is more complex than it sounds in organizations with multiple deployment pipelines, inconsistent tooling, or manual deployment processes. Deployment Frequency requires counting actual production deployments (not staging deployments or release tags). Lead Time requires measuring the time from first commit to production deployment for each change. MTTR requires measuring time from incident detection (not occurrence) to service restoration. Change Failure Rate requires tracking which deployments required rollback or hotfix. Investment in tooling to accurately capture these metrics is prerequisite to using them for performance management.

FAQs

Relevant Executive Roles