Flags of Flexibility - Unlocking Rapid Iteration
summary
Feature flags power a launch of 20 features launched behind flags enabling parallel development, internal dogfooding, private betas, AB/multivariant experiments and gradual rollouts (5% → 10% → 25% → 100%). Infrastructure runs as a hybrid: Split.io + internal platform so teams click preset pipelines for mirror releases and rollouts. Trade-offs include exponential decision complexity, technical debt from stale flags (platform knocks on teams after 3 and 6 months), performance tuning, and about 95% flag eval success with ~5% blocked by ad‑blockers/firewalls.
AI-written summary · about
chapters (10)
- 0:00Introduction and speaker background
- 4:45What are feature flags?
- 9:20Decoupling deployments from releases
- 12:09Testing in production and internal releases
- 13:52Beta programs and segmented rollouts
- 17:05Feature flag nomenclature and variants
- 23:11Benefits and costs of feature flags
- 30:33Shared ownership and behavioral change
- 33:19Conclusion and strategic impact
- 33:41Q&A session
related talks
59mFaster & less risky releases with feature flags
05mFeature Flags
41mFeature Flags: Separate Deployment from Feature Release
41mImproving feature flag implementation to achieve flow
30mA Tale of Two Feature Flags
44mFlagging features — a DevOps approach to continuous release
56mWhen Feature flags go bad
27mJust Enough Feature Flagging