Message match glossary

A/B test

An A/B test splits visitors at random between two versions of a page, the original and a change, and compares what each group does, so the difference can be put down to the change and not to who happened to visit.

Also called: split test, controlled test

On a landing page

For ad traffic, the cleanest test compares visitors from the same ad: half see the original page and half see the changed one. Comparing one ad's visitors with another's mixes up the ad and the page.

Mistakes to avoid

  • Stopping at the first good day. Early results swing, especially on revenue.
  • Not checking the split. If one version gets far more visitors than it should, something other than chance is choosing who sees what.
  • Not checking the change still shows. A site update can quietly stop a test from running while the report keeps counting.
  • Testing small changes on pages that get little traffic. They need far more visitors than you have.

See the gap on your own ads

Get a free report on your own live Meta ads: each promise a landing page leaves out, with the ad’s words beside the page’s. A person on our team checks it before it reaches you.

Get a free ad gap reportCompare Throughline CR