Reducing CI Pipeline Time
Cut CI build times from 18 minutes to under 5 minutes to restore developer flow and increase shipping frequency.
The full NCT chain
“Our CI pipeline takes 18 minutes on average. Developers are context-switching during the wait, losing flow, and shipping slower as a result. Our best engineers are most affected because they're iterating fastest. Internal surveys show CI wait time is the #1 developer experience complaint. If we get CI under 5 minutes, we expect measurable improvements in shipping frequency and developer satisfaction.”
Audit and parallelize the three slowest CI pipeline stages
- Profile each CI stage to identify bottlenecks
- Restructure test stages to run in parallel
- Separate unit tests from integration tests
- Benchmark before and after parallelization
Migrate test suite to run against a dedicated test DB instead of shared staging
- Set up dedicated test database infrastructure
- Update test configuration to use test DB
- Migrate test fixtures and seed data
- Verify test isolation and data cleanup
Add pipeline caching for node_modules and build artifacts
- Implement dependency caching layer
- Add build artifact caching between stages
- Configure cache invalidation rules
- Monitor cache hit rates and adjust TTL
When to use this
Context
Use this NCT when developer satisfaction surveys or retros consistently surface CI wait times as a pain point. Especially valuable when you can correlate slow CI with decreased shipping frequency or increased context-switching.
Analysis
Why this NCT works
This Narrative reframes infrastructure work as a developer experience problem tied to shipping velocity — making it easier to prioritize against feature work. The Commitments are technical but specific, and each one independently moves the metric. Engineering leadership can track progress without needing to understand every implementation detail.
Related examples
Building an Observability Stack
Implement structured logging, metrics, and alerting to reduce incident response time and improve system reliability.
Infrastructure & PlatformMigrating a Monolith to Microservices
Decompose a monolithic application into services to improve team autonomy, scalability, and deployment speed.
Product & EngineeringImproving Mobile App Performance
Reduce app startup time and eliminate jank to improve mobile user retention and app store ratings.
See how these teams use NCT
Ready to build your own NCT?
Start with a Narrative. Add Commitments. Break them into Tasks. Free forever to start.
Free forever. No credit card needed.