A/B testing compares two versions of a message, page or experience using separate groups to measure their effect on a defined outcome.
How it works in retail
A controlled test changes a specific element and assigns comparable audiences to the variants. The outcome could be completed purchases, sign-ups or another stated action.
A retail example
A retailer tests two product-page headlines while holding price, imagery and the offer constant, then compares purchase conversion.
What to keep in mind
Small samples, overlapping changes and stopping a test too early can produce unreliable conclusions. A higher click rate may not translate into more profitable orders.
Examples are illustrative. Terminology and measurement conventions can vary by business and market.