A-B Split is a method of examining the effectiveness of media or marketing methods. Using A-B split marketing, a list of target names is divided into two groups on a random basis, with one group selected as a control group and the other one as a test group. The aim of the A-B split is to find out which single variable is the most effective in enhancing response rates to a marketing campaign or accomplishing some other desired outcome.
A-B split has been utilized for mailing campaigns in the history but has also successfully adapted for apply in interactive media, for checking the effectiveness of e-mail blasts and poster advertisements.
For instance, an email campaign by a bulletin publisher may involve a specific "call to action" - like subscribe within 48 hours to get 20% discount - embedded in the message to half the target audience, and no call for action (i.e. no solicitation to subscribe or declare a discount) in the message to another half. This will allow the publisher to know if the "call to action" actually works, and whether the response rate is good enough to rationalize the 20% discount.