An experiment comparing two versions of a piece of marketing content — an ad, a headline, an email subject line — shown to similarly sized, randomly assigned audiences, to determine which performs better on a defined metric.
Why it matters
A/B testing replaces opinion and internal debate about what will perform better with an actual measured result, and is one of the most direct ways a marketing team can steadily improve performance over time through accumulated, evidence-based decisions rather than guesswork.
In a sentence
"We A/B tested two subject lines, and the more specific, benefit-led version outperformed the clever, curiosity-driven one by a wide margin."
Related terms
- Often confused with: Multivariate testing (an A/B test compares two distinct, complete versions against each other; multivariate testing tests multiple individual elements — like a headline and an image — in combination, to see which specific combination performs best).
- Synonyms: Split test.