A/B Testing
A method of comparing two versions of something - an email subject line, a landing page, a button - to determine which performs better.
Definition
A/B testing (also called split testing) is a controlled experiment in which two variants of a single element are shown to different segments of an audience simultaneously. The version that achieves the desired outcome at a higher rate – more clicks, sign-ups, purchases – is declared the winner and rolled out more broadly.
Why it matters
Opinions about what works are unreliable. A/B testing replaces guesswork with evidence. Small, iterative improvements compound over time: a 10% lift in email click-through rate from a better subject line, repeated across dozens of campaigns, has a significant cumulative impact on revenue and growth.
Example
An email marketer tests two subject lines – “5 tools every developer needs” vs. “The developer toolkit we actually use” – by sending each to 10% of the list, then sending the winner to the remaining 80%. A SaaS company might A/B test a “Start free trial” button against a “Get started” button to see which drives more signups.