An A-B test is a classic statistical design in which individuals or subjects are randomly split into two groups and some intervention or treatment is applied - one group gets treatment A, the other treatment B. Typically one of the treatments will be a control (i.e. nothing new), and the other will be a treatment of interest. Any difference in outcomes between the two groups is then attributed either to the treatment difference, or to chance. An allowance for chance effects is usually made via statistical inference -- either a confidence interval or a hypothesis test. The term A-B test is typically used in marketing applications of this statistical design.