Learn A/B testing
These articles explain how to think about A/B testing: the methodology of hypothesis-driven testing, the Bayesian statistics behind your results, the common ways tests mislead, when to call a test done, how to run several tests at once, how to build a testing program, and how to roll out a winner. For product how-tos, see the documentation.
Bayesian A/B Testing: How Chance to Beat Control Works
How the Bayesian engine behind Chance to beat control and the credible intervals works — the priors, the Monte Carlo sampling, and why it beats p-values.
A/B Testing Mistakes: Common Pitfalls and How to Avoid Them
The most frequent ways A/B tests mislead — peeking, stopping early, multi-variable variants, mid-test edits, ignoring segments — and how to avoid each.
How to Build an A/B Testing Program: The Five Pieces
Move from ad-hoc tests to a sustainable experimentation practice, and run it with Test Plans — an AI-planned program that keeps the backlog and cadence going.
When to Stop an A/B Test and Declare a Winner
When to call an experiment done — the 95% threshold and the sample floor, how to end the test, and how to roll out and apply the winning variant after.
A/B Testing Methodology: Principles and Best Practices
The principles behind compounding A/B testing gains — hypothesis-driven tests, single-variable rigor, sequential discipline, and documenting losses.
Running Multiple A/B Tests at Once: When It's Safe
Run several experiments at once — how assignment works across tests, how the Overlap banner flags shared visitors, and the discipline that keeps data clean.
How to Roll Out an A/B Test Winner at 100%
Serve a winning variant to 100% of targeted traffic while you make the change permanent — serving only, nothing recorded or billed, and you can stop any time.