A/B Testing
Test method where two or more variants of a campaign run in parallel to determine the better version based on data.
Also known as: Split Testing, A/B Test, Multivariate Testing
A/B testing runs two or more versions of an ad creative, landing page or email in parallel to comparable recipient groups. Performance data then shows which variant reaches the defined goal better. Instead of deciding by gut feeling, the test delivers a measurable answer.
What can be tested
In email marketing, subject lines, sender names, send times, call to action buttons, images and layouts are classic test candidates. On landing pages, headlines, form lengths, colour schemes or placements get compared. Important: only one variable per test, otherwise the effect cannot be attributed clearly.
When results are meaningful
For a test to be truly reliable, you need a sufficiently large number of participants per variant and a clearly defined success metric. Tests on tiny lists deliver random noise. The runtime should also be realistic, because a one hour test says nothing about daily or weekly behaviour.
A/B testing in JD Mail
JD Mail supports any number of parallel variants. You define a small test group, JD Mail sends the variants at the same time and automatically picks the winner based on predefined criteria. The winning variant then goes automatically to the remaining target audience.