The A/B test or split test is a multi-variant test for websites or devices to be tested. When testing websites, viewers are presented with two different test versions of a website for evaluation. Unlike the multi-variant test, the A/B test only changes a single element, such as a graphic or a font size.
In the A/B test, the test group is divided into an A and a B group. One group is presented with the original variant, the other with the changes. The two groups of viewers decide which version is more efficient and generates more attention and response. The evaluation can be based on page views, viewing time or marketing and sales goals and conversion rate. A/B testing can include bounce rate or acquisition rate, costs per lead( CPL) or costs per order( CPO). The website determined by the test group to be the most successful version is then placed on the web.