What is GrowthBook?
GrowthBook is a warehouse-native platform for experimentation, feature flags, and product analytics that lets product teams run rigorous tests and control feature rollouts on top of their own data warehouse. GrowthBook is open source and fully transparent, and it offers both a managed cloud option and a self-hosted deployment that runs entirely on the customer's own cloud or on-prem infrastructure.
Key Features
- Warehouse-native architecture: GrowthBook connects to all of a product's data in the customer's warehouse, so metrics and experiment results are computed on data the team already owns rather than copied into a separate silo.
- Experimentation: A powerful and flexible statistical model supports rigorous tests at scale, with metrics defined in SQL directly on the data warehouse and guardrails that keep experiments consistent as programs grow to thousands of tests per year.
- Feature flag management: Auto-rollbacks and ramp schedules with guardrails provide a safety net for fast shipping, and flags can be added automatically to new features. GrowthBook reports handling 1T+ feature flag lookups per day.
- Product analytics: GrowthBook turns product data into insights so teams can understand user behavior, track trends, and identify opportunities, with an AI Analyst for asking questions of the data and support for exploring data with agents.
- Developer tooling: Ultra-lightweight SDKs make integration easy, and real-time debugging and other dev tools support day-to-day engineering work. A REST API and MCP Server let any AI agent interact with the platform.
- Self-hosting and transparency: The self-hosted edition delivers the same features and product as the cloud version, 100% on the customer's cloud or on-prem, backed by an open-source, fully transparent codebase.
- Role-specific workflows: Purpose-built views serve engineers, data teams, and product managers, covering everything from defining metrics in SQL to controlling rollouts without engineering and sharing real-time experiment dashboards.
- Integrations: GrowthBook is built to adapt to any workflow or tech stack, with a published integrations directory covering the tools teams already use.
Use Cases
- AI-era release management: Teams using AI coding tools and agents to ship faster rely on GrowthBook's auto-rollbacks and guarded ramp schedules so increased release velocity does not increase risk.
- Large-scale experimentation programs: Data science teams at high-traffic organizations run rigorous tests at scale; the Wikimedia Foundation scaled experimentation for 15B+ monthly page views, and Oda has run 400+ experiments.
- Feature rollout control for product managers: PMs control feature rollouts without engineering involvement, while engineers get real-time debugging and automatic flag creation for new features.
- Monetization and growth testing: Companies such as Lingokids (3.5M+ customers benefiting from tests) and Breeze Airways ($1MM+ incremental monthly revenue) use GrowthBook to test changes that affect revenue and conversion.
- Agent-driven analysis: Teams point AI agents at GrowthBook through the MCP Server and REST API, or use the AI Analyst, to explore product data and surface growth opportunities.
Who It Is For
GrowthBook is built for engineers, data teams, and product managers at modern product organizations, including data science teams running experimentation programs and companies that want full control over their data stack. It solves the problem of stitching together separate experimentation, feature flag, and analytics tools by putting all three on top of the customer's own warehouse, with an open-source, self-hostable option for teams that need total control. The platform is trusted by 4,000+ companies worldwide, with named customers including Dropbox, the Wikimedia Foundation, Breeze Airways, The Philadelphia Inquirer, TodayTix, Khan Academy, Oda, Lingokids, and Fyxer.
Pricing
GrowthBook offers a free start through GrowthBook Cloud, which provides instant updates, low maintenance, and warehouse-native operation, alongside a self-hosted edition with the same features that can be downloaded and run on the customer's own cloud or on-prem. The source material does not publish specific price points; teams needing details are directed to book a demo.
Frequently Asked Questions
What is GrowthBook used for?
GrowthBook is used for experimentation, feature flag management, and product analytics. Product teams use GrowthBook to run rigorous tests, roll out features safely with auto-rollbacks and ramp schedules, and analyze product data stored in their own warehouse.
Is GrowthBook open source?
Yes. GrowthBook is open source and fully transparent, and it offers a self-hosted edition with the same features and product as the cloud version, running 100% on the customer's cloud or on-prem infrastructure.
Does GrowthBook work with a data warehouse?
Yes. GrowthBook is warehouse-native: it connects to all of a product's data in the customer's warehouse, and metrics are defined in SQL on that data warehouse. This lets teams analyze experiments and product data without moving it into a separate system.
Who should use GrowthBook?
GrowthBook is aimed at engineers, data teams, and product managers. Engineers get lightweight SDKs, real-time debugging, and automatic flag creation; data teams get a flexible statistical model and SQL-defined metrics; and product managers can control rollouts and share experiment dashboards without engineering help.
Can AI agents interact with GrowthBook?
Yes. GrowthBook provides an MCP Server and REST API for any AI agent, and its AI Analyst lets teams ask questions of their product data. GrowthBook positions these capabilities for teams and agents running tests and analyses at scale.
