A/B testing, in the context of CRO (Conversion Rate Optimization), is the method of comparing two versions of a page or element (headline, CTA button, layout) by randomly splitting traffic between the control version (A) and the modified version (B), to statistically determine which variant generates a better conversion rate.
For reliable results, A/B tests need a sufficient volume of traffic and conversions to reach statistical significance, a testing period long enough to cover natural variations (days of the week, payment cycles), and testing a single variable at a time to clearly attribute the impact of the change.
Why it matters
A/B testing turns conversion-optimization decisions from subjective guesses into data-driven decisions, reducing the risk of implementing changes that, while they seem logical, actually reduce performance.