growthbook

Why we included this project

GrowthBook bundles feature flags, A/B experiments, and product analytics into one self-hostable codebase, a practical way to cut the number of paid SaaS tools a team has to juggle. The statistics engine is where it differs from most open-source alternatives: CUPED variance reduction, sequential testing, Bayesian analysis, bandits, and sample ratio mismatch checks are all included, the kind of analysis that normally takes a dedicated data science team to reproduce. It is also warehouse native, so metric definitions and dashboards query BigQuery, Snowflake, ClickHouse, or Redshift directly instead of moving data into a separate store. Product and growth teams running controlled rollouts and serious experiments get SDKs for React, Python, Android, and iOS, which covers the apps most of them already ship.