A/B test

An A/B test splits an audience at random between two or more versions of a message and compares the outcome that matters. It is the only dependable way to tell whether a subject line, offer or send time changed behaviour rather than simply coincided with it. Most email tests fail on sample size rather than idea quality: a modest list with a low conversion rate can need far more recipients than one send provides, so work out the sample you need before writing the variant.