Key points
- A test is only a test if one variable changes and a result could make you stop.
- Fix tracking before the first dollar. A test measured on clicks tells you about clicks.
Set up
- Confirm the Pixel and the Conversions API both report the enquiry event, and that the event is de-duplicated.
- Pick one variable: the opening line, the visual, the offer, or the format. Not all four.
- Write down, before launch, the result that would make you drop the variant.
- Use the same audience, placement and budget for every variant.
Read the result
- Judge on the enquiry event, not on cost per click or on thumb-stop rate.
- Wait for enough enquiries that a single extra one would not change the ranking of the variants.
- Keep the winner, retire the loser, and start the next test with the winner as the control.
- Keep the creative consistent with the brand people will meet on the landing page; the branding service covers that system.
Sources
- About the Meta Pixel — Meta Business Help Center
- About Conversions API — Meta Business Help Center
- Advertising Standards — Meta Transparency Center
About this guide
Published · Updated

