Open-source feature flags and A/B testing
Visit WebsiteReviews onG2
25 reviews trackedThe Bottom Line
Entry price
Free plan available, paid tiers above
Biggest pro
Warehouse-native approach means experiment data never leaves your infrastructure
Biggest con
Requires a data warehouse connection for experiment analysis, not plug-and-play
TL;DR - GrowthBook
- Open-source feature flagging and experimentation platform that analyzes results directly in your data warehouse
- Lightweight SDKs (9KB JS) evaluate flags locally with zero network requests for near-zero latency
- Free tier with unlimited flags and experiments; Pro at $40/user/month adds advanced statistics and visual editor
Pricing: Free plan available
Best for: Growing teams
4.6/5 across review platforms
What is GrowthBook?
GrowthBook is a warehouse-native feature flagging and experimentation platform used by over 2,700 companies. It connects directly to your data warehouse, Snowflake, BigQuery, Redshift, or others, so experiment results are computed on your own data without exporting events to a third party. The platform includes both Bayesian and Frequentist statistical engines with support for CUPED, sequential testing, and multiple metric corrections. Feature flags evaluate locally via lightweight SDKs (the JavaScript SDK is 9KB) with zero network requests at runtime. A visual A/B test editor enables no-code experiments for marketing and product teams. GrowthBook is available as open-source self-hosted, cloud-managed, or with a managed warehouse option.
Available on: Web
Pros & Cons
Pros
- Warehouse-native approach means experiment data never leaves your infrastructure
- Generous free tier with unlimited flags, experiments, and traffic
- Open-source edition supports unlimited users at no cost
- Ultra-lightweight SDKs with local evaluation keep latency near zero
- Enterprise-grade statistics (CUPED, sequential, bandits) included in Pro plan
- Visual editor lets non-engineers create and manage A/B tests
Cons
- Requires a data warehouse connection for experiment analysis, not plug-and-play
- Free cloud plan limited to 3 users with no advanced statistics
- Pro plan at $40/user/month can get expensive for large teams
- Self-hosted deployment needs infrastructure expertise and ongoing maintenance
- Visual editor and advanced testing features are only available on paid plans
Ratings Across the Web
4.6(25 reviews)
Ratings aggregated from independent review platforms. Learn more
Key Features
Warehouse-native experiment analysis on Snowflake, BigQuery, and RedshiftBayesian and Frequentist statistical engines with CUPED and sequential testingLightweight feature flag SDKs (9KB JS) with local evaluation and zero network requestsVisual A/B test editor for no-code experiments on websitesMulti-arm bandit testing for automated traffic allocationSafe rollouts with automatic rollback on metric regressionsSticky bucketing for consistent user experiences across sessionsAdvanced permissioning and approval workflows for enterprise teamsSOC 2, ISO 27001, GDPR, COPPA, and CCPA complianceOpen-source self-hosted option with unlimited users and flags
Pricing Plans
Starter (Cloud)
$0 / month
- Up to 3 users
- Unlimited feature flags
- Unlimited experiments
- Warehouse-native analysis
- Managed warehouse option
- Unlimited traffic (MAU/MTU)
- Community support
Pro (Cloud)
$40 / user/month
- Up to 50 users
- Advanced statistics (CUPED, sequential testing)
- Visual A/B test editor
- Sticky bucketing
- Multi-arm bandits
- Safe rollouts
- Advanced permissioning
- Power calculator
- Premium email and in-app support
- 14-day free trial
Enterprise (Cloud)
Custom
- Unlimited users
- SSO/SAML integration
- Advanced access control and approval flows
- Cross-experiment insights
- Data pipelines and exportable audit logs
- Prerequisite flags and holdouts
- Dedicated support and enterprise SLA
- Team training
Open Source (Self-Hosted)
$0
- Unlimited users
- Unlimited feature flags and experiments
- Warehouse-native analysis
- Unlimited traffic
- Community Slack support
Reviews
4.6/5
Across 25 verified user reviews on G2
Add your hands-on experience to help the next buyer.
Best GrowthBook Alternatives
Top alternatives based on features, pricing, and user needs.
LaunchDarklyPaid
Feature flags and experimentation at scale
StatsigFreemium
Feature flags and product experimentation
UnleashFreemium
Open-source feature flag management
SplitPaid
Feature flags and experimentation platform
FliptFree
Open-source feature flag infrastructure
EppoPaid
Feature flagging and experimentation platform
Still deciding?
Most buyers shortlist 2 or 3 tools before committing. Pull a side-by-side comparison or browse the full alternatives shortlist below.
Explore More
GrowthBook FAQ
What does 'warehouse-native' mean in GrowthBook?
GrowthBook connects directly to your existing data warehouse (Snowflake, BigQuery, Redshift, etc.) to analyze experiment results. Your event data stays in your infrastructure and is never copied to a third-party system.
Is GrowthBook open source?
Yes. GrowthBook's self-hosted edition is open source under the MIT license with unlimited users, feature flags, and experiments at no cost. Cloud and enterprise tiers add managed infrastructure and advanced features.
Do I need a data warehouse to use GrowthBook?
Not necessarily. GrowthBook offers a Managed Warehouse option on the free Starter plan so you can run experiments without connecting your own warehouse. For full flexibility, a warehouse connection is recommended.
How do GrowthBook feature flags affect page load performance?
GrowthBook SDKs evaluate flags locally on the device with zero network requests at runtime. The JavaScript SDK is only 9KB, making it one of the lightest in the industry with negligible impact on page load.
What statistical methods does GrowthBook support?
GrowthBook offers both Bayesian and Frequentist engines. Advanced features include CUPED for variance reduction, sequential testing for early stopping, multi-arm bandits for automated traffic allocation, and corrections for multiple metric comparisons.
How does GrowthBook compare to LaunchDarkly?
GrowthBook is open source and warehouse-native, meaning experiment analysis runs on your own data. LaunchDarkly is a proprietary SaaS focused primarily on feature flags. GrowthBook's free tier is more generous, and its Pro plan includes built-in experimentation that LaunchDarkly charges separately for.
Source: growthbook.io