A/B Sample Size: How Many Visitors Before You Trust the Result
Running a test on too few visitors gives a winner by luck. A sample-size estimate tells you how long the test honestly needs to run.
What is the A/B Test Sample Size Calculator?
Enter baseline conversion, the relative uplift you want to detect, and the tool estimates visitors per variant using a simplified standard approximation. This is a rough planning estimate, not a substitute for a proper power calculation, so treat the result as a starting point and check it before running a high-stakes test.
How to use it
- Enter baseline conversion rate.
- Enter the relative uplift you want to detect.
- Click Estimate Sample Size.
Worked example
With a 10% baseline and 20% relative uplift, the target is 12%, and the estimate is about 3838 visitors per variant.
Why does a small uplift need a big sample?
Smaller differences are harder to tell apart from random variation, so detecting them takes many more visitors.
Frequently asked questions
Is this estimate exact?
No. It is a simplified approximation for planning. A statistics tool or analyst should confirm sizes for important tests.
What if traffic is low?
Test bigger changes, or accept a longer run. Small uplifts on low traffic can take impractically long to prove.