A/B testing is a method of testing for an advertising campaign that involves two different versions of a web page to see which is more effective. In a basic A/B test, a random visitor is shown either the “A” web page, which is generally the current design or control page, or the “B” web page, which is the new design or the challenger page. The reaction of the visitor is tracked, whether the visitor leaves the page, stays and reads text, or makes a purchase based on what the web page looks like. These reactions are recorded and the results determine which page, page “A” or page “B,” will become the new page for the website. There is also a form of the A/B test that is called a 50/50 A/B Split Test, and that leaves the decision of which page is shown to the visitor to coin toss predictability.

To be effective, an A/B test must reach an audience within the target demographic, or more simply, should be tested on the key audience the website wishes to reach. A/B testing should be used to narrow down how individual changes work for or against the website. Changing too many variables at once will render the test useless, as the reaction of the visitors will not be judging just one change on the web page. To change multiple variations and test them, websites should use a multivariate test.

Of course, A/B testing is not just used to test the web page itself. There are other variables on a web page that can use an A/B test to determine whether they have any effect on the behavior of visitors to the website.

Other variables that might be tested using an A/B test include:
Banner ads
E-mails
Landing pages
Other changeable variables