Technotize

A/B Test Sample Size Calculator

Enter baseline rate, minimum detectable effect, and confidence. Get visitors needed per variant.

Inputs
%
%
%
%
Formula · n = 2 × (z₁₋α/₂ + z₁₋β)² × p(1−p) / Δ²

Result

Fill the inputs to see your result.

What is sample size?

Sample size is the number of visitors per variant required to reliably detect a chosen effect at a chosen confidence and power. Plan it before launching, not after.

The sample size formula.

n = 2 × (z₁₋α/₂ + z₁₋β)² × p(1−p) / Δ², with p the baseline conversion rate, Δ the absolute lift you want to detect, and z values from your confidence and power.

A worked example.

Baseline 4%, MDE 20% relative, 95% confidence, 80% power. Δ = 0.04 × 0.20 = 0.008. n is roughly 9,400 per variant.

How to choose MDE.

Pick the smallest lift worth shipping. Detecting smaller effects costs disproportionately more data; detecting nothing is the worst outcome.

When sample is out of reach.

Test a bigger swing, lower confidence and treat the result as directional, or batch multiple changes into one stronger variant. Do not run a test you cannot power.

Related calculators

Experimentation

A/B Test Significance

Analyze the result

Experimentation

Conversion Rate Calculator

Baseline metric

Experimentation

Landing Page Conversion Rate

Where you test

FAQ

How do you calculate sample size for an A/B test?
The standard formula is n = 2 × (z₁₋α/₂ + z₁₋β)² × p(1−p) / Δ², where p is the baseline rate, Δ is the absolute lift you want to detect, and the z values come from your confidence and power.
What is minimum detectable effect (MDE)?
The smallest relative lift you want to be able to detect. A smaller MDE needs a much larger sample.
What confidence and power should I use?
95% confidence (α = 0.05) and 80% power are the common defaults. Higher confidence and power need more visitors.
Why does the sample feel so large?
Small lifts on small baselines need a lot of data. Halving the MDE roughly quadruples the sample.
What if I cannot reach the required sample?
Aim for a bigger change with a stronger hypothesis, or accept lower confidence and treat the result as directional, not conclusive.

Book a strategy call.