Skip to main content
← Glossary · Experimentation Strategy

GrowthBook

An open-source feature flag and experimentation platform that can be self-hosted, with a SaaS option, and connects directly to your data warehouse.

What Is GrowthBook?

GrowthBook is the leading open-source experimentation platform. It offers feature flags, A/B testing, and warehouse-native analysis — and because it's open source, teams can self-host for free and own their data, or use the hosted SaaS tier. It has become a popular post-Google-Optimize migration target for cost-sensitive and privacy-conscious teams.

Also Known As

  • Open-source experimentation platform
  • "The GitHub stars one" (in developer circles)
  • GrowthBook Cloud (the SaaS tier)
  • Self-hosted experimentation

How It Works

An engineering team deploys GrowthBook via Docker, connects it to their Postgres (for flag config) and their data warehouse (for metric computation). Developers use the SDK to wrap features in flags. When an experiment ends, GrowthBook issues SQL against the warehouse to compute results using Bayesian or frequentist statistics, depending on configuration.

Best Practices

  • If self-hosting, treat the GrowthBook deployment like any production service — monitoring, backups, upgrade cadence.
  • Use the Bayesian engine when stakeholders want probabilistic framing ("80 percent chance variant B is better"); use frequentist when they want traditional p-values.
  • Pin SDK versions and review changelogs — open-source moves fast and breaking changes happen.
  • Contribute back metric templates and SDKs if you customize them; the ecosystem is small enough that it matters.

Common Mistakes

  • Assuming "open source" means "free" in operational terms. Self-hosting has real ops cost; for small teams, GrowthBook Cloud is usually cheaper in total.
  • Under-investing in metric definitions because the tool is free — metric quality still determines experiment quality.
  • Running the self-hosted version on under-provisioned infrastructure and blaming the tool when results are slow.

Industry Context

GrowthBook is popular with early-stage startups, privacy-regulated industries (healthcare, fintech) that need full data control, and teams that were burned by the Google Optimize sunset. SaaS/B2B adopts it frequently; ecommerce/DTC less so; lead gen depends on team engineering maturity.

The Behavioral Science Connection

Open-source tooling shifts ownership from vendor to team, which changes the psychological relationship with experimentation. Teams using self-hosted GrowthBook tend to be more willing to customize and challenge defaults — the endowment effect works for you when you own the tool.

Key Takeaway

GrowthBook is the go-to open-source experimentation platform — ideal when you need data control, want to avoid vendor lock-in, or have outgrown free tiers of commercial tools.

Related Articles

'Not Significant' Doesn't Mean Your A/B Test Failed — It Means You're Uncertain The most expensive misreading in A/B testing is treating 'not statistically significant' as 'no difference.' It actually ... Adaptive Algorithms vs. Fixed-Sample A/B Tests: Why I Changed How I Run Experiments After 100+ experiments per year, fixed-sample A/B testing's opportunity cost became impossible to ignore. Here's what ad ... A/B Test Repository Architecture: Schema, Tagging, and the Retrieval Time Budget A/B test repositories don't fail because the schema is wrong. They fail because nobody can find what they need fast enou ...
← Browse All Terms