Most people test creatives wrong. They throw five ads in a campaign, wait a week, and pick whichever has the most purchases. That's not testing. That's hoping.
Here's a systematic approach that actually identifies winners.
The Testing Framework
Separate your campaigns into two types: Testing and Scaling. Testing is where new creatives prove themselves. Scaling is where proven winners get real budget.
Never test in your main campaigns. You'll mess up the algorithm and waste money on unproven creatives.
Setting Up the Test Campaign
Create a dedicated testing campaign with these settings:
- Campaign budget optimization (CBO)
- Daily budget: 3-5x your target CPA per ad
- One ad set with broad targeting
- 3-5 new creatives to test
Broad targeting lets Facebook find the best audience for each creative. Don't handicap the test with narrow audiences.
What Makes a Valid Test
You need statistical significance. A creative with 2 purchases isn't proven better than one with 1 purchase. That's noise, not signal.
Rule of thumb: Each creative needs at least $50-100 in spend or 10,000 impressions before you can judge it. More spend means more confidence.
What to Test First
Test the big stuff first. Creative changes beat copy changes beat audience changes in terms of impact.
Order of testing:
- Creative format (video vs static vs carousel)
- Hook/opening (first 3 seconds of video, main image angle)
- Offer positioning
- Copy length and angle
- Call to action
Reading the Results
Look at these metrics in order:
- Hook rate (3-second video views / impressions)
- CTR (click-through rate)
- CPC (cost per click)
- Conversion rate
- CPA or ROAS
High hook rate + high CTR + low conversion = landing page problem, not creative problem. Low hook rate = creative problem. Diagnose correctly.
When to Kill a Creative
Kill fast. If a creative spends 2x your target CPA with no conversions, it's done. Don't wait and hope.
Exception: If hook rate and CTR are strong, give it more time. The eyeballs are there. Conversion might just need more data.
Graduating Winners
When a creative proves profitable in testing, move it to your scaling campaign. Add it as a new ad in an existing winning ad set.
Monitor for 3-5 days. If it maintains performance, increase budget. If it tanks, the test audience might have been different from your scaling audience. Not every test winner scales.
Keep Testing
Creative fatigue is real. Even your best ads will stop working eventually. Always have new creatives in testing so you're ready when current winners die.
Aim to test 3-5 new creatives every week. Most will fail. That's fine. You only need occasional winners to keep your campaigns profitable.