# A/B testing

A/B testing compares two versions of something by showing each to a share of visitors and measuring which performs better.

In e-commerce it is used to compare recommendation strategies, campaign designs, page layouts and search ranking rules.

Visitors are split at random, so that the only difference between the groups is the version they see. The result is judged on a measure chosen in advance, such as conversion rate or basket size.

A test needs enough visitors and enough time to rule out chance. Stopping at the first sign of a winner is a common mistake.

---

Source: https://www.primesolutions.com.tr/en/glossary/ab-testing/
