Free tool
A/B test calculator for ad creatives
Enter the visitors and conversions of two ad creatives, thumbnails or pages. The calculator gives each conversion rate, the uplift, and whether the difference is significant at 95% or could be chance.
The tool loads here. It needs JavaScript, and runs in your browser.
What it is good for
- Check whether a new ad creative really beats the old one.
- Test two thumbnails, headlines or landing pages.
- Decide whether to keep a test running or call it.
- Explain a test result with its p-value.
How it works
- Enter the visitors and conversions of variant A.
- Enter the same for variant B, counted the same way.
- Read the rates, the uplift, the p-value and whether it is significant at 95%.
Facts
| Price | free |
|---|---|
| Where it runs | in your browser; nothing is uploaded |
| Test | two-sided two-proportion z-test, pooled, at 95% confidence (p < 0.05) |
| Formula | z = (rate B − rate A) ÷ √(p × (1 − p) × (1 ÷ visitors A + 1 ÷ visitors B)), where p is the pooled conversion rate |
| Worked example | A: 200 conversions from 1,000 visitors (20%). B: 250 from 1,000 (25%). Uplift +25%, z = 2.68, p = 0.0074: significant at 95% |
| Minimum | 100 visitors per variant; under that it says not enough data yet |
| Mark | none (no image) |
Questions
- What is an A/B test?
- An A/B test shows two versions of an ad, page or email to similar audiences at the same time and compares how often each one gets the result you want, such as a click or a sale. Only one thing should differ between A and B, so you know what caused the difference.
- How do I calculate A/B test significance?
- Work out each conversion rate (conversions ÷ visitors) and the pooled rate p (all conversions ÷ all visitors). Then z = (rate B − rate A) ÷ √(p × (1 − p) × (1 ÷ visitors A + 1 ÷ visitors B)). The two-sided p-value comes from the normal distribution; under 0.05 means significant at 95%.
- What is a good conversion rate for an A/B test?
- It depends on the placement, the audience and the season, so there is no one number to aim for. Compare each variant with the other and with your own past tests. What moves it: the offer, the creative, the audience, the landing page and the time of year.
- How many visitors does an A/B test need?
- It depends on your conversion rate and on how small a difference you want to detect: small differences need many more visitors. This calculator waits for at least 100 visitors per variant, and a test that is not significant yet may simply need more.
- Does a significant result mean B will keep winning?
- No. Significant at 95% means a difference this large would rarely happen by chance if A and B were really the same. Results can shift with the season, the audience or ad fatigue, so keep checking.
Part of
More free tools
Add text to imageAspect ratio calculatorBefore and after makerBlur and pixelateCharacter counterCollage makerCPM calculatorCTR calculatorGIF makerColour picker from imageImage compressorImage converterImage resizerImage splitterInstagram fontsPhone and browser mockupMeme generatorNo-crop borderQR code with logoROAS calculatorSafe-zone checkerSocial image resizerUTM link builderVideo frame extractorWatermark makerYouTube banner maker
Free credits every month
Sign in with Google and get 50 free credits every month. Unused credits do not roll over. Your credits also work in Claude Code, Codex and OpenCode through the BrShot MCP server.
Signing in shares your Google email address with BrShot; the privacy note says what we keep and for how long. Privacy note.