Rethinking Delivery Performance: Lessons from DORA 2023 and Developer Experience

In 2023, the DevOps Research and Assessment (DORA) team released its latest State of DevOps report, highlighting the evolving landscape of software delivery. The findings were clear: high-performing teams aren’t just faster—they’re safer, more resilient, and more developer-friendly.

As engineering leaders, it's time we stop treating performance as just a speed game. Modern software delivery is a balance of velocity, reliability, and developer experience.

This post reframes key engineering metrics through the lens of DORA 2023—and introduces a few complementary metrics that expose the hidden friction in your pipeline.

The Core Five: DORA 2023 Metrics

These metrics form the foundation of performance analysis in elite engineering organizations.

1. Lead Time for Changes

Definition: Time from code committed to code successfully running in production.

This reflects the speed of innovation and feedback loops. DORA 2023 reaffirms it as one of the strongest indicators of delivery performance.

What to track: Time from PR creation to successful deployment
Why it matters: Long lead times discourage iteration, delay feedback, and frustrate teams

2023 Benchmark:

Elite: <1 hour
Low: >6 months

2. Deployment Frequency

Definition: How often your team deploys code to production.

This measures how continuously you can deliver value. Frequent deploys typically correlate with small batch sizes, less risk, and healthier teams.

What to track: Daily or weekly prod/staging deployments per team
Why it matters: High deployment frequency enables fast fixes, feature flags, and continuous validation

2023 Benchmark:

Elite: Multiple times per day
Low: Fewer than once every 6 months

3. Change Failure Rate

Definition: The percentage of deployments causing degraded service, bugs, or rollbacks.

This represents the cost of change. DORA teams emphasize failure rate as a critical counterbalance to deployment speed.

What to track: Failed deployments / total deployments
Why it matters: High rates indicate poor test quality, integration issues, or insufficient validation

2023 Benchmark:

Elite: 0–15%
Low: 46–60%

4. Mean Time to Recovery (MTTR)

Definition: Time from incident detection to full restoration of service.

Failures happen. What matters is how quickly you can detect, diagnose, and recover.

What to track: Incident start to incident resolution
Why it matters: Fast MTTR builds trust in your system and your team

2023 Benchmark:

Elite: <1 hour
Low: >6 months

5. Reliability

Definition: The degree to which your system meets operational expectations (uptime, latency, etc.).

Added in recent years, reliability reflects customer-facing system health, not just internal delivery metrics.

What to track: SLAs, SLOs, incident frequency, alert fatigue
Why it matters: A fast pipeline is meaningless if your system is down or unstable

Summary of DORA 2023 Core Metrics

Metric	Definition	What to Track	Elite Benchmark
Lead Time for Changes	Time from commit to production	PR open to prod deploy	< 1 hour
Deployment Frequency	How often code is deployed to prod	Prod deploys per day/week	Multiple times per day
Change Failure Rate	% of deployments causing incidents	Failed/rolled-back deployments	0–15%
Mean Time to Recovery	Time to restore service after failure	Time from incident to resolution	< 1 hour
Reliability	System’s ability to meet uptime and performance expectations	SLOs, uptime, incident rate	SLO-compliant & stable

Beyond DORA: Metrics That Capture Developer Experience

While DORA gives us delivery health, it doesn’t tell the whole story. Developer Experience (DevEx) has emerged as a critical success factor—and these supporting metrics help illuminate it.

6. Build Time

Definition: Time to compile, test, and package a change.

While not in DORA, build time directly affects lead time and developer flow. Long builds waste time, reduce confidence in the CI system, and lead to risky batch commits.

What to track: CI job duration (cold and hot builds)
Why it matters: Developers won’t iterate quickly if builds punish them for trying

Tip: Watch for variance—flaky builds are worse than slow ones.

7. Queue Time

Definition: Time a CI/CD job waits before it starts executing.

This is the invisible bottleneck in scaling engineering teams. When everyone’s building at once, even the best pipelines slow down.

What to track: Time from job submission to execution start
Why it matters: Queue time reflects infrastructure limits and resource contention

Insight: DORA emphasizes speed, but speed can’t happen when runners are exhausted.

8. Frequency of Broken Main

Definition: How often the main branch is in a non-deployable state.

This is a culture and process smell. A consistently broken main branch means poor test coverage, weak gating, or rushed reviews.

What to track: % of time main is red / incident count from merges
Why it matters: Main should be sacred. If you can’t trust it, everything slows down.

Actionable Tip: Set up merge protection, fast revert paths, and daily "green checks".

9. Engineering Overhead

Definition: Time engineers spend on infra, CI tooling, and environment management instead of feature development.

This metric ties directly to platform maturity and DevEx quality—themes DORA 2023 heavily emphasizes.

What to track: % of sprint hours spent on pipeline bugs, test flakiness, tooling fixes
Why it matters: Engineers drowning in overhead are engineers not shipping value

How to measure: Use retros, dev satisfaction surveys, or tag internal tickets. Patterns emerge quickly.

Closing Thoughts: DORA + DevEx Is the Future

DORA’s metrics provide a science-backed foundation for measuring delivery performance. But they’re not the end of the story.

In 2023, DORA explicitly emphasized:

Developer Experience (DevEx)
Platform engineering
Psychological safety
Sustainable delivery culture

To build elite teams, you must measure what matters and remove what hurts. That includes invisible friction: long waits, flaky builds, overloaded tools, and broken feedback loops.

Your goal isn’t just faster code—it’s a system that enables developers to move confidently, safely, and joyfully.

That’s real velocity.

References:

2023 State of DevOps Report – DORA
Accelerate: The Science of Lean Software and DevOps by Forsgren, Humble, and Kim