Glossary Term
A/B Testing
Showing two versions of something to randomly split groups of users at the same time, then comparing one metric chosen in advance. It answers "did this change cause the difference?", which a before-and-after comparison cannot.
Example: A team tests a shorter checkout against the current one. After two days the new version is ahead and someone wants to ship it. The test was planned for two weeks to cover weekday and weekend buyers, so they wait. By day 14 the gap is gone, and they have avoided a false win.
Share it