Sample Size Calculator

Calculator for AB Tests and Multi-Variant Experiments

Why Sample Size Matters

Sample size determines how trustworthy your experiment results are. Too small, and random noise can look like a real effect; too large, and you waste precious time. The right sample size gives your test enough power to detect real changes while avoiding false conclusions.

How to use the Calculator

This calculator helps you estimate how many samples you need to run a reliable experiment.

Baseline Conversion Rate

This is your current conversion rate—the percent of people who complete the action you care about (like signing up or clicking). If about 5 out of 100 people convert, enter 5.

Minimum Detectable Effect

This is the smallest change you want to be able to detect. If a 20% relative increase would be meaningful for your team, enter 20. The smaller the change you want to detect, the more people you’ll need.

Number of Variants

Pick how many versions you’ll test, including the control. 2 variants is a simple A/B test. More variants (like A/B/C) mean you’ll need more data to stay confident in your results. This calculator doesn’t adjust the confidence for multiple comparisons, so adding more variants slightly increases the chance of false positives.

Advanced Settings

Hypothesis

Choose One-Sided if you only care about detecting a change in one direction. Choose Two-Sided if you want to know if the new version is higher or lower than control.
Mixpanel supports Two-Sided. One-Sided support is coming soon!

Confidence Level

Confidence Level is the probability your test will correctly avoid a false positive when there’s no real effect. 95% is standard. Higher confidence means you’ll need a larger sample.

Power

Power is how likely your test is to correctly detect a real effect if one exists. 80% is common. Higher power gives stronger results but needs more data.

Closing Tips

These calculations assume equal traffic for each variant. Uneven splits reduce statistical power, so aim to keep allocations even. Stopping tests too early can make results unreliable, so aim to reach the recommended sample size before deciding.

In short: Fill in your current rate, how big a change you care about, and your test settings—the calculator tells you how much data you need for trustworthy results.