A/B Test

An A/B test compares two versions of something — a page, button, or email — by showing each to a random half of users and measuring which performs better.

Business & Tech

A/B testing replaces guesswork with data. By changing one thing at a time and splitting traffic randomly, you can confidently say which version drives more signups or sales.

The key is a fair comparison: same time period, random assignment, and enough visitors that the difference is statistically meaningful rather than luck. Tools like Optimizely and Google Optimize automate this.

Examples

Formula

A: 1000 views, 50 signups -> 5.0%
B: 1000 views, 70 signups -> 7.0%

Output:

B wins (+2.0 points)

See also

Back to the Dictionary