A/B testing tools, also called A/B testing software or experimentation platforms, run controlled experiments on websites, apps and product features. They decide which users see which version, record what each group does and report whether the difference is likely to be real.
Showing 3 of 3 tools
Open-source A/B testing & flags
Open-source, warehouse-native experimentation and feature-flag platform with Bayesian and frequentist stats engines.
Enterprise experimentation
Enterprise digital experience platform offering web and feature experimentation alongside CMS and personalization.
Open-source product analytics suite
Open-source all-in-one product analytics with feature flags, A/B tests, session replay and surveys for fast iteration.
The basics
A/B testing tools split users between two or more versions of a page or feature and compare how each group behaves. Many also manage feature flags and product analytics.
In lean startup terms they turn vanity metrics into actionable ones. A controlled experiment shows whether a change caused a result. A shift in the numbers alone does not show that.
The book Trustworthy Online Controlled Experiments, by Kohavi, Tang and Xu, is a common reference for running tests well.
Low-traffic products often cannot reach a clear result in a reasonable time. Interviews, usability tests or before-and-after comparisons may teach more.
Testing tools earn their place once you have steady traffic on the steps you want to improve. Plan the sample size before the test starts, and keep the test running for its planned length.
The routine it supports
How the listed tools are priced
The software randomly assigns each user to a control or a variant and keeps them in that group for the length of the test. It records a chosen metric for each group, such as sign-ups or purchases, and compares them with a statistical method.
Tools differ mainly in that last step. Some use frequentist statistics with a fixed sample size, some use Bayesian methods, and some use sequential testing that allows checking results early. Ask which method a tool uses and how it guards against stopping a test too soon. The glossary entry on A/B testing gives the short definition.
Client-side tools change the page in the user's browser, often through a visual editor. They suit marketing pages and copy tests, and they need little developer time, but the change can cause a brief flicker and is limited to what the browser can alter.
Server-side tools decide the variant in your own code, usually through feature flags. They suit pricing, algorithms, back-end changes and mobile apps, and they need developers to set up. Many product teams use server-side testing and keep client-side tools for marketing sites.
Free A/B testing tools come in two forms: free plans of commercial tools, usually limited by traffic, tests or features, and open source platforms you host and run yourself. Open source options often read results from your own data warehouse, which keeps data in your hands.
Before choosing, check the cost of the data you will send, the effort to host and update the tool, and whether its statistics are documented. For a product with little traffic, customer discovery or a simple MVP may teach more than any test.
Buying guide
Who it's for: Product, growth and engineering teams that want decisions based on measured behavior rather than opinion.
Trustworthy statistics
Clear handling of sample size, significance and peeking, so results can be trusted.
Feature flags
Release to a small group first, then roll out or roll back safely.
Actionable metrics
Cohorts, funnels and retention instead of totals that only go up.
Privacy and data control
Clear data ownership, consent handling and regional hosting if you need it.
Before you choose
FAQ
Sources
Written by the lean-stack editors following our methodology. Updated October 3, 2026.
For vendors
Submit it for review. Every tool is checked against our listing criteria and researched from public sources. Approved tools join their category in the next published batch.
