Testing with flags

Testing with feature flags allows you to experiment and gather insights about your feature's performance before fully releasing it. This section covers three main subjects: Goals, Experiments, and How to Use Experiments.

• Goals: Goals in the Bucketeer system help to measure how user behaviors are affected by feature flags within experiments.
• Experiments: Experiments are a way to test new features or changes with a subset of users before a full release. They allow you to collect data and feedback to improve the feature's effectiveness.
• Experiment results: After creating an experiment, you can access the experiment result page. It shows the experiment's current state, evaluation period, and provides components for analysis. The Bucketeer system uses Bayesian inference — with a primary-metric decision, peeking-safe stopping, and a traffic-allocation (SRM) check — to suggest the best-performing variation.

By utilizing goals, running experiments, and analyzing the results, you can make data-driven decisions to optimize features, improve user experience, and achieve desired business outcomes.