You have made several versions of a film and launched them. A week later, the numbers are in, and one version has the best click rate. Is it the winner? Not necessarily. Reading a creative test well is a skill, and a few habits make it much easier.

Set up a fair test

Channable’s guide to Meta creative testing suggests one creative per ad set, with the same daily budget for each, and a change of one variable at a time, such as a discount badge versus none. It also warns that budgets are not always split evenly between creatives in a campaign, so a result where one ad gets 80% of spend and another gets 5% can mislead.

Give the test enough data

The same guide notes that Meta commonly recommends around 50 optimisation events per week for conversion-focused ad sets. Small tests that stop early often show noise instead of a winner.

Read the numbers in order

First, check delivery was fair: spend, impressions, reach and frequency. Second, scroll-stop: hook rate is three-second video plays divided by impressions. Third, hold: hold rate is ThruPlays divided by three-second plays. Fourth, click intent such as link clicks and landing page views. Fifth, conversion: add to cart, purchase, cost per result.

What each number tells you about the film

A low hook rate points to the first three seconds. A good hook with a low hold rate points to the middle of the story. Good attention with weak conversion often points to the offer or the landing page, not the film. Judge the winner on conversion metrics, not early clicks alone.

Feed it back into the next film

Because AI production keeps characters, looks and voices as assets, the next round can change just the hook or the setting. That makes each test a step forward, not a restart.

If you would like help planning a test around your next film, write to namaste@teevra.co.