Split Testing (A/B Testing)
Definition
Comparing two variations of a webpage or ad to determine the best performer.
Example Use Case
A nonprofit organization tests two donation page layouts to maximize contributions.
Split testing shows two versions of the same thing to different groups and compares how each performs. The term is often used interchangeably with A/B testing, and the distinction is mostly a matter of emphasis: split testing usually means dividing an audience between two complete variants, while A/B testing sometimes refers to testing a single changed element.
Random assignment is what makes the comparison meaningful, because it ensures the groups differ only by chance rather than by who happened to see which version. Deciding the metric before the test starts protects against the temptation to find an encouraging number afterwards.
The common mistakes are consistent: ending the test as soon as one variant is ahead, changing the test mid-flight, or running it long enough to include a seasonal effect that will not recur. A test is only informative if it was designed to answer a question rather than to produce a winner.
