Core Principle
DORA (DevOps Research and Assessment) provides the most empirically grounded framework for measuring and improving software delivery performance. It consists of five delivery metrics and 35+ capabilities that predict performance. The central finding: speed and stability are not a tradeoff; they reinforce each other.
Why This Matters
DORA gives concrete, measurable language for arguments about engineering practices. Instead of “we should use trunk-based development because it’s better,” you can say “DORA’s research across thousands of organizations shows trunk-based development predicts superior delivery speed, stability, and availability.” The framework also situates itself within a broader measurement landscape alongside SPACE and DevEx.
Evidence/Examples
The Five Metrics
| Metric | Measures | Category |
|---|---|---|
| Change Lead Time | Commit to production duration | Throughput |
| Deployment Frequency | Deploys per time period | Throughput |
| Failed Deployment Recovery Time | Time to restore after failure | Stability |
| Change Fail Rate | % deployments needing remediation | Stability |
| Deployment Rework Rate | % unplanned deploys from incidents | Stability |
Key Capabilities (35+)
Technical: trunk-based development, continuous integration, continuous delivery, test automation, deployment automation, infrastructure-as-code, monitoring and observability, database change management, code maintainability, version control
Process: working in small batches, streamlining change approval, team experimentation, customer feedback, WIP limits
Cultural: generative organizational culture, learning culture, transformational leadership, job satisfaction, well-being
Relationship to SPACE and DevEx
Nicole Forsgren: “DORA is an instance of SPACE, as a measure of the performance of software.”
- DORA (2014-): Team/org delivery outcomes via five metrics
- SPACE (Forsgren et al., 2021): Five dimensions: Satisfaction, Performance, Activity, Communication, Efficiency
- DevEx (Greiler, Storey & Noda, 2022): Developer-centric perception data alongside system metrics
Implications
- Use DORA metrics as leading indicators for engineering practice decisions
- No single metric captures productivity; use SPACE dimensions to avoid Goodhart’s Law
- DORA’s 2025 finding: “AI acts as an amplifier, but the greatest returns come from focusing on the underlying sociotechnical systems”
Related Ideas
- The Speed-Stability False Tradeoff
- Trunk-Based Development
- Small Batch Sizes and Feedback Loops
- Lean Flow Theory Applied to Software
- AI-Native Infrastructure The Nix-LLM Virtuous Cycle
Questions
- How do DORA metrics apply to solo developers or very small teams where “deployment frequency” may not be meaningful?
- The 2025 AI amplifier finding: does this mean infrastructure maturity is prerequisite to AI productivity gains?
Sources
- DORA Team. dora.dev
- Forsgren, N., et al. “The SPACE of Developer Productivity” (2021). ACM Queue.
- Greiler, M., et al. “An Actionable Framework for Understanding and Improving Developer Experience” (2022). arXiv:2205.06352.
- Forsgren, N., Humble, J., & Kim, G. Accelerate (2018). IT Revolution Press.