GrowthBook is an open source platform, founded in 2020 by Jeremy Dorn and Graham McNicoll, that unifies feature flagging, A/B testing and experimentation, and product analytics into a single tool teams can self-host for free or run on GrowthBook Cloud.
The company went through Y Combinator's Winter 2022 batch, is based in San Francisco with roughly 40 employees, and raised a Series A round in June 2025 that brought its total funding to about $23.1 million.
Key Features
GrowthBook's defining architectural feature is being warehouse-native: rather than duplicating event data into a proprietary store, it connects directly to a company's existing data warehouse, such as Snowflake, BigQuery, or Redshift, and runs its statistics engine against that data.
The platform combines feature flag management, A/B testing with Bayesian and frequentist statistics, multi-armed bandits, and product analytics like funnels and retention in one interface, with the bulk of the codebase released under the permissive MIT license.
Pricing
GrowthBook offers free tiers on both its cloud and self-hosted paths: a 3-user Starter cloud plan and a fully self-hostable open source edition, each including unlimited feature flags and experiments.
Paid tiers add collaboration and enterprise controls, with Pro at $40 per seat per month adding a visual editor and premium support, and custom-priced Enterprise tiers, cloud or self-hosted, adding SSO, SCIM, audit logs, and dedicated support.
Key Features
Warehouse-native architecture — Connects directly to a company's existing data warehouse instead of requiring a separate proprietary event pipeline, keeping experiment and analytics data in one place.
Feature flag management — Lets engineering teams roll out features gradually, target specific segments, and instantly disable problematic releases without redeploying code.
A/B testing and experimentation — Runs Bayesian and frequentist statistical analysis on experiment results computed directly against warehouse data, including multi-armed bandit testing.
Product analytics — Provides funnel, retention, and metric-tracking dashboards built on the same underlying data model used for experiments and feature flags.
Open source, MIT-licensed core — The bulk of GrowthBook's codebase is open source under the MIT license and available on GitHub, with a large and active self-hosting community.
Visual editor — Available on Pro and Enterprise tiers, the visual editor lets non-engineers set up and modify experiments without writing code.
Enterprise access controls — SSO, SCIM provisioning, advanced permissioning, and exportable audit logs support security and compliance needs on the Enterprise tier.
Flexible deployment — Available as a fully managed cloud service or a self-hosted deployment, with feature parity on the core flagging and experimentation functionality.
Pros & Cons
Pros
Genuinely open source MIT-licensed core with a free, fully functional self-hosting option, not just a limited trial.
Warehouse-native design avoids duplicating event data into a separate proprietary analytics store, reducing vendor lock-in.
Combines feature flags, experimentation, and product analytics in one tool instead of three separate point solutions.
Free tiers exist on both the cloud and self-hosted paths, lowering the barrier to trying the full statistics engine.
Backed by a large and active GitHub community with thousands of stars and thousands of reported companies using the platform.
Cons
The warehouse-native model requires a company to already have a modern data warehouse set up, which can be a barrier for smaller teams without one.
Self-hosting requires engineering resources to deploy, maintain, and upgrade the platform over time.
Advanced features like SSO, SCIM, and audit logs are gated behind Enterprise pricing that requires a custom quote rather than being self-serve.
Pro plan is priced per seat, which can get expensive quickly for larger product and engineering organizations.
As a warehouse-dependent tool, real-time experiment results can be slower than platforms that use a dedicated, purpose-built event pipeline.
Pricing
Starter (Cloud) Free Monthly
Pro (Cloud) $40/seat/month Monthly
Enterprise (Cloud) Custom Custom
Open Source (Self-Hosted) Free N/A
Enterprise (Self-Hosted) Custom Custom
Frequently Asked Questions
Is GrowthBook free to use?
Yes, GrowthBook offers a free, fully functional Open Source edition for self-hosting and a free Starter plan on GrowthBook Cloud for up to 3 users, both including unlimited feature flags and experiments.
What does it mean that GrowthBook is warehouse-native?
It means GrowthBook connects directly to a company's existing data warehouse, such as Snowflake, BigQuery, or Redshift, to run its statistical analysis, rather than requiring event data to be duplicated into a separate proprietary analytics store.
Who founded GrowthBook and when?
GrowthBook was founded in 2020 by Jeremy Dorn and Graham McNicoll, went through Y Combinator's Winter 2022 batch, and is headquartered in San Francisco.
How much does GrowthBook Pro cost?
The GrowthBook Cloud Pro plan costs $40 per seat per month for up to 50 users and adds a visual editor, multi-armed bandits, safe rollouts, and premium support on top of the free Starter plan.
Can GrowthBook be self-hosted?
Yes, GrowthBook can be self-hosted for free using its open source edition, or organizations can license a self-hosted Enterprise tier for advanced access control, SSO, SCIM, and dedicated support.
What are GrowthBook's main competitors?
GrowthBook competes with commercial feature flagging and experimentation platforms like LaunchDarkly, Optimizely, Statsig, and Split, as well as open source feature-flag-focused tools like Unleash and Flagsmith.