Applying DORA Metrics to Mobile Development
Article Summary
rolgalan from Glovo tackles a question mobile teams keep asking: how do you actually use DORA metrics when you don't control app distribution? Turns out, the answer isn't to abandon them—it's to adapt them.
DORA metrics were built for backend systems with instant deployments. But mobile development lives in a different world: app store reviews, gradual user updates, and zero control over when fixes reach customers. This article breaks down how Glovo's mobile team adapted each DORA metric to make them meaningful for mobile.
Key Takeaways
- Track throughput (PR merges) instead of deployment frequency for weekly release cycles
- Redefine MTTR as time until fix is available, not full user adoption
- Measure build pipeline duration for hotfixes, not store review times
- Use feature flags and kill switches to recover from incidents without releases
- Combine DORA with CircleCI metrics to cover both production and CI/CD health
DORA metrics work for mobile when you focus on what you control (your pipeline and process) rather than what you don't (store reviews and user update rates).
About This Article
Mobile teams at Glovo found it hard to measure Change Failure Rate in a way that actually mattered. Minor defects didn't need hotfixes, but they still had to be tracked before the next release went out.
Glovo started using Conventional Commits to tag commits by their purpose. This let teams calculate defect-to-commit ratios for each release and spot patterns in how code quality changed over time.
Glovo hasn't rolled out company-wide CFR tracking yet, but teams using this method can now see which releases had the most defects. That visibility helped them make targeted improvements to their process.