Sample Size Calculator

In an A/B test, the planned sample size depends on the baseline event rate, minimum detectable effect, significance threshold, statistical power and how traffic is split between variants. The calculator helps plan whether a test is feasible before it begins. A smaller effect typically needs more observations to detect reliably than a large effect.

The result is a planning estimate, not a guarantee that the test will produce a significant result or establish causality by itself. Invalid randomization, multiple unplanned comparisons, repeated peeking at conventional tests and biased measurement can undermine conclusions even with many visitors. Select the effect worth detecting based on business relevance rather than picking a convenient small number.

For very low-volume sites, an experiment may take too long to deliver a useful answer. In that case, interviews, usability tests or larger design changes may be more practical. A statistical sample size calculator is also different from a pricing or service quote calculator, which estimates a fee rather than evidence requirements.