An experiment where two similar versions of marketing content are randomly shown to potential customers and tested for effectiveness. For example, a marketer could randomly substitute “half off” for “50% off” in an otherwise identical online ad and statistically test for differences in click-through rates.