Roll out features safely. Toggle features for cohorts or individuals to test the impact before rolling out to everyone.
Deliver features safely
Deploy updates behind a feature flag to test it with increasing numbers of users, then rollback easily if there’s an issue.
Breakdown our analytics by feature flag variant to understand the impact on performance.
Target on users’ properties, or their behavior
Test changes based on criteria like plan, geography or LTV, or users’ previous behavior.
“We use feature flags to issue changes to 50% of users and then compare the effect. Experiment, find results, decide where to focus and then iterate.”
Feature flags documentation
- Learning resources
- Demo video
- Creating feature flags
- Implementing the feature flag
- Develop locally
- Roll out the feature flag
- Roll out to a percentage of users
- Filter by user properties
- De-activating properties
- Feature flag persistence
- Multivariate feature flags (alpha)
- Creating a feature flag with multiple variants
- Using multivariate feature flags in your code
- Querying data by multivariate feature flag values