How do I test ad creative on Meta?
Use an A/B test with creative as the variable. Meta splits your audience into random groups so nobody sees more than one version, compares up to 5 versions, and picks the winner with the lowest cost per result. Change one creative element at a time, give each version the same budget, and avoid testing by switching ads off and on.
A creative test answers one question: which image, video or piece of copy gets you results more cheaply? Meta’s built-in way to answer it is an A/B test with creative as the variable. Meta says an A/B test divides your audience into random, non-overlapping groups so nobody sees more than one version, while informal testing, such as turning ads on and off, can lead to overlapping audiences.
What creative variables can you test?
Meta’s creative variable covers different ad images, text or creative types, and is managed at the ad level. Meta’s examples include:
- a slideshow video against a carousel of images,
- the same image or video with different ad text,
- different product images to see which performs best.
Meta recommends starting with a single variable for the clearest insight. You can test more than one, but you won’t know which change caused the winning result.
How do you set up a creative A/B test?
- Write a hypothesis, for example “a product demo video will beat our best static image”.
- Duplicate the ad you want to test, or pick two existing campaigns or ad sets to compare.
- Change only the creative in the new version.
- Give both versions the same budget; Meta recommends this for a fair comparison.
- Choose the key metric, such as cost per result, that decides the winner.
- Run the test from Ads Manager or the Experiments tool, then review the result.
Meta notes that duplicating a campaign also duplicates its budget.
Why not just turn ads off and on to compare them?
Meta does not recommend informal testing like that. It says switching ad sets or campaigns on and off can lead to inefficient delivery and unreliable results, and audiences can overlap. An A/B test splits audiences evenly and keeps them statistically comparable.
What does Meta’s research say about video versus images?
Meta’s direct response research found that campaigns using both static images and video achieved conversion lift at a 17% higher rate than campaigns using static images alone. For testing simple videos against a still image, Meta suggests:
| Advice | Detail from Meta |
|---|---|
| Video types to try | Basic (animate the still), brand, benefit or demo |
| Motion | Add animation in the first 3 seconds |
| Shape | Create in 1:1 or 4:5 |
| Length | Ideally 5 to 10 seconds |
| Message | One benefit, with a call to action and a consistent end card |
What should you test next?
Meta’s tips after a creative test:
- Run the same test with a different campaign objective, since creative may not perform the same for each.
- Test a different creative element, such as the headline after the image.
- Test creatives within the winning format, for example two videos after video beat image.
FAQ
What if my creative test declares no winner?
Meta suggests checking that your audience is large enough, since A/B tests divide it and small groups can under-deliver, and increasing the budget. Meta also warns that, when testing audience, placement or delivery, ad sets that are too similar may stop it from confidently declaring a winner.
How many creatives should I test at once?
A Meta A/B test compares up to 5 versions. Meta’s Instagram video best practices also suggest starting with up to 10 creatives per ad set, keeping the top 5 as evergreen and running tests for at least 4 days.
Does a creative A/B test cost extra?
No. Meta says there is no additional cost to create a test, and you can view all results in the Experiments tool. Meta notes that lift tests are different: because they hold back a control group that does not see your ads, they can affect the overall CPM of the campaigns in the test.
Sources
- About experiments (Meta, checked 2 Oct 2026)
- Selecting a variable for your A/B test (Meta, checked 2 Oct 2026)
- About A/B testing (Meta, checked 2 Oct 2026)
- Testing more than one variable in A/B tests (Meta, checked 2 Oct 2026)
- Tips for improving your A/B tests (Meta, checked 2 Oct 2026)
- Creative best practices for conversion testing (Meta, checked 2 Oct 2026)
- Best practices for Instagram video ads (Meta, checked 2 Oct 2026)
Drafted with AI and checked claim by claim against the pages linked under Sources. Found a mistake? Tell us via the corrections policy.