Measuring Engineering Productivity

Most teams measure productivity wrong. Here's how to do it right.

Bad Metrics to Avoid

These seem appealing but lead to bad decisions:

• Lines of code — Rewards verbosity
• Commit count — Encourages spam commits
• Hours worked — Measures presence, not output
• Tickets closed — Encourages quick, shallow work

Good Metrics That Work

Based on DORA research and our experience:

Outcome Metrics

1. Deployment frequency — How often you ship
2. Lead time — Commit to production time
3. MTTR — Recovery time from incidents
4. Change failure rate — % of broken deployments

Leading Indicators

1. PR review time — How fast feedback happens
2. Code review coverage — What % gets reviewed
3. Test coverage — Quality indicators
4. Technical debt — Accumulation over time

Implementation Tips

Start Simple

1. Pick one metric that matters to your team
2. Track it consistently
3. Review monthly
4. Iterate

Avoid Gaming

• Measure trends, not absolutes
• Combine metrics for context
• Focus on outcomes, not activities

Conclusion

Good metrics help teams improve. Bad metrics lead to dysfunction. Choose carefully and iterate.