◆ RESOURCES · CALCULATOR

A/B Test Sample Size calculator.

This A/B test sample size calculator tells you the visitors per variant you need to detect a real lift — and how many weeks that will take at your current traffic. Enter your numbers below.

01Your test parameters

What are you testing?

Baseline conversion rate

Your current conversion rate for the page or flow under test.

Minimum detectable effect (relative)

The smallest relative lift worth detecting. 10% means a move from 3.0% to 3.3%.

Statistical significance

Confidence the result is not chance. 95% is the standard.

Statistical power

Chance of detecting a real effect if one exists. 80% is the standard.

Weekly visitors (total, split across variants)

Total weekly traffic entering the test, divided evenly between control and treatment.

/WK

◆ YOUR TEST PLAN

SAMPLE SIZE PER VARIANT

53,148

106,296 visitors total across both variants

DETECTING

3.00% → 3.30% at 95% significance, 80% power.

ESTIMATED DURATION

11 weeks

at 10,000 weekly visitors

This is a workable test — plan for roughly 11 weeks and avoid peeking at results before the sample is complete.

Or book a demo →

02The method

How to calculate A/B test sample size.

Sample size for an A/B test answers one question: how many visitors does each variant need before a difference in conversion rate is unlikely to be noise? It depends on four inputs — your baseline rate, the size of the lift you want to detect, how confident you want to be, and how often you are willing to miss a real winner.

Baseline conversion rate (p1)

Your current conversion rate. Lower baselines need far more traffic, because the absolute difference you are chasing is smaller.

Minimum detectable effect (MDE)

The smallest relative lift worth caring about. A 10% MDE on a 3% baseline means detecting a move to 3.3%. Smaller MDEs need much larger samples.

Statistical significance

How sure you want to be that a result is real and not chance. 95% (alpha 0.05) is the convention, giving a two-tailed Z of 1.96.

Statistical power

The chance of detecting a real effect when one exists. 80% is standard, giving a Z of 0.84. Higher power needs more visitors.

◆ THE FORMULA

This calculator uses the standard two-proportion sample size formula. For each variant:

n = (Z_alpha/2 + Z_beta)^2 * [p1(1 - p1) + p2(1 - p2)] / (p2 - p1)^2

p1 is the baseline conversion rate; p2 is the baseline lifted by your MDE — p2 = p1 × (1 + MDE).
Z_alpha/2 is 1.645 for 90% significance, 1.96 for 95%, and 2.576 for 99%.
Z_beta is 0.84 for 80% power and 1.28 for 90% power.
Estimated duration is the total required sample (2 × n) divided by your weekly visitors, rounded up to whole weeks.

When that duration runs long — past two or three months — the issue usually is not your test design. It is traffic. That is the wedge Optimize Pilot's Flight Path is built around: it checks whether you have the traffic to test before it tells you to test, and routes low-traffic pages to SEO and demand work first.

◉ TEST WHAT YOU CAN ACTUALLY MEASURE

Know your sample size. Get started.

Optimize Pilot sizes every experiment against your real traffic and tells you when to test — and when to build traffic first. 90-day money-back guarantee.

Book a 15-min stack audit →

◆ Loading…

How to calculate A/B test sample size.