A/B Test (Split Test)
Comparing two ad variants to determine which performs better.
An A/B test isolates one variable (usually creative or audience) and runs both versions to comparable groups of users. The platform's native split-test tool reserves separate audiences so the same user never sees both — without that, your 'test' is contaminated.
For statistical significance with conversion campaigns, you typically need at least 100 conversions per variant. Below that, the winner is noise, not signal. In small accounts test creative angles first (biggest lever), then offers, then audiences. Almost never test bid strategies — let the platform decide.