Simple AB Test Calculator
Compare conversion rates with a stated statistical model and show sample size and assumptions, without a business promise.
Conversions are the visits that did what you wanted: bought, signed up, or clicked.
Your result
Enter the visits and conversions for both versions. The verdict appears here.
A two-sided two-proportion z-test at the 95% level. It tells you whether the gap is likely real, not how big it will be next month.
How to use it
- Type the visits and conversions for A.
- Type the visits and conversions for B.
- Read the z score and the two rates.
Examples
- Not clear yet100 visits, 20 conversions; 100 visits, 30 conversions 20% vs 30%, +50%, about 90% confidence: not clear yet.
- A clear winner2400 visits, 96; 2380 visits, 131 4.00% vs 5.50%, +37.6%, 98.6% confidence: B wins.
Tool details and privacy
Enter visits and conversions for versions A and B. The page shows both conversion rates, the change, the confidence, and a clear verdict.
A z score is a calculation about these four numbers. It is not a promise that one version will win later.
Options
- A two-sided two-proportion z-test at the 95% level.
- The z score and p-value are shown for anyone who wants the detail.
Limits
- This is not a business promise.
- Conversions cannot exceed visits.
- Each visit count must be from 1 to 1,000,000.
Privacy
What you type stays in your browser. Nothing is uploaded or saved on the server.
Questions
Is this a promise about the business?
No. It says whether the gap is likely real.
What do 100, 20, 100, and 30 give?
Rates of 20% and 30%, z = -1.63, about 90% confidence, so not clear yet.
Can conversions exceed visits?
No.
Is the test uploaded?
No.