Definition
A/B Testing
Also known as: Split Testing, Bucket Testing
A method of comparing two versions of a webpage, notification, or campaign element to determine which performs better based on conversion metrics.
What does A/B Testing mean?
A/B testing splits traffic between two versions of a page or element, one control and one variant, so a team can attribute any change in behavior to the difference between them rather than to noise or seasonality. It is the primary discipline for validating whether a new headline, layout, or trust element actually changes visitor decisions.
Why does A/B Testing matter?
Without testing, teams rely on opinion about what a badge, review widget, or CTA color should do, and opinions are frequently wrong once real traffic is exposed to them. Running a controlled test protects revenue from confident-sounding but unproven redesigns and turns subjective debates about social proof placement into resolved questions backed by visitor behavior.
How is A/B Testing applied in practice?
In practice, a test needs a single changed variable, a large enough sample to reach statistical significance, and a fixed conversion goal defined before the test starts. Tools split visitors randomly, track conversions per arm, and report a confidence level; most CRO teams require at least 95% confidence before rolling a winning variant out to all traffic.
Go deeper
Related terms
Where this fits
A/B Testing sits in the conversion rate optimization part of the NotiProof resource network.
Read the Conversion Rate Optimization guide