What is GrowthBook?
GrowthBook combines feature flag management, experimentation, and product analytics. Teams can target users, roll out features gradually, and evaluate experiments using SQL-backed metrics from data warehouses such as BigQuery, Snowflake, and Databricks. Its statistical engine includes Bayesian and sequential methods, CUPED, bandits, and sample ratio mismatch checks.
The application also provides shared dashboards, funnel metrics, user journey exploration, and experiment diagnostics that expose raw assignment events. SDKs connect application code to feature flags and experiments, while webhooks, a REST API, and an MCP server support automation. Agent permissions can restrict autonomous actions and require human approval for sensitive changes.
GrowthBook can run through GrowthBook Cloud or be self-hosted with Docker Compose using a public prebuilt image. It follows an open-core licensing model: most code uses the MIT license, while some directories fall under a separate commercial enterprise license.
GrowthBook screenshots
Explore the interface. Open a screenshot to view it at full size.
Source: official GrowthBook project page.
Key capabilities
- Feature delivery Target user groups, manage feature flags, and release changes gradually.
- Experiment analysis Evaluate experiments with SQL-backed metrics, statistical methods, and exposure diagnostics.
- Product analytics Build shared dashboards and explore funnels and user journeys.
- Warehouse connectivity Query warehouse data and inspect which queries, experiments, or users drive warehouse costs.
- Workflow automation Use webhooks, the REST API, and the MCP server to manage features and experiments with scoped agent permissions.
- Operational controls Configure observability webhooks to stop and roll back matching feature ramp-ups when alerts fire.
Where GrowthBook fits
GrowthBook fits into application delivery workflows where teams need to introduce features incrementally and measure their effect on product metrics. It also supports analysis of onboarding funnels and user journeys. Its dbt synchronization tool translates models into GrowthBook fact tables and metrics, helping connect experimentation definitions to existing data modeling workflows.
Before deployment, teams should verify SDK support for their application stack, warehouse connectivity, SQL metric definitions, and exposure-event quality. They should also review commercial licensing boundaries and the availability of required capabilities in their chosen deployment. Interleaving experiments and contextual bandits are described as beta features. Building the Docker image from source requires registry authentication for its hardened base images, unlike using the public prebuilt image.
Install
git clone https://github.com/growthbook/growthbook.git
cd growthbook
docker compose up -dReviews of GrowthBook
Be the first to share your experience with GrowthBook.
Write a review
Loading review form…