Sample Size Calculator
This free A/B test sample size calculator tells you exactly how many visitors you need per variant before your results are trustworthy. Enter your baseline conversion rate, the minimum uplift you want to detect (MDE), significance and power, and it returns the sample size per variant. The formula is n = (zα/2·√(2·p̄·(1−p̄)) + zβ·√(p₁·(1−p₁)+p₂·(1−p₂)))² / (p₂−p₁)². Example: a 3.5% baseline CVR at 95% confidence, 80% power, detecting a 10% relative uplift needs ~88,000 visitors per variant.
Set your baseline CVR and MDE to see the sample size you need.
Why sample size matters before you test
The most common A/B testing mistake: stopping a test early because it "looks like" it's winning. Without a pre-calculated sample size, your results are meaningless – you're just watching noise. This is called peeking, and it inflates false positive rates to 25–30%.
Calculate your required sample size before you start. Then commit to running until you hit it – regardless of intermediate results.
Once you know your sample size, use our A/B test duration calculator to turn it into a run length in days, the MDE calculator to find the smallest uplift your traffic can realistically detect, and the statistical significance calculator to check the result once the test is done.
Understanding the inputs
Your current conversion rate. The lower your CVR, the more traffic you need to detect changes. At 1% CVR you need ~4× more visitors than at 4% CVR.
The smallest uplift worth detecting. Setting MDE too low means you need enormous traffic. For most eCommerce tests, 10–20% relative MDE is practical.
At 80% power, if there's a true effect you'll detect it 8 out of 10 times. Higher power = more traffic needed, but fewer missed wins.
Running too many inconclusive tests?
Most brands don't have enough traffic to test everything. We prioritize the highest-impact experiments so every test counts.