Testing Static vs Video vs Carousel Systematically: Advanced Facebook Ad Creative Testing

Quick answer

Run a quarterly format test that places one matched concept into static, video, and carousel formats simultaneously. Most accounts find that 1-2 formats dominate 70%+ of spend but the dominant format shifts every 2-3 quarters. Lock the test, do not let recency bias decide.

Why formats are not interchangeable

Static is cheap to produce, fast to ship, and dominates Feed. Video costs more, takes longer, dominates Reels, and earns higher thumb-stop. Carousel is the most flexible but the hardest to read because attribution spreads across cards. Each format has a different ceiling — and that ceiling moves with audience and category.

The framework, step by step

1. Pick one concept that is currently your control winner.
2. Build three matched executions: a 1:1 static, a 9:16 video (15-30s), and a 5-card carousel. Same headline, same CTA, same offer.
3. Launch in one CBO with three ads. Set placement to All Placements.
4. Run 5-7 days, target 50+ conversions per format.
5. Compare CPA, ROAS, thumb-stop (video only), and CPM. Compute % share-of-spend Meta gave each format.
6. Re-run the test every 12 weeks. Format winners shift seasonally.

Example test matrix

Format	Hook	Body	CTA	Why it might win
Static 1:1	Bold headline	Single image proof	Shop Now	Cheapest CPM, fastest test
Video 9:16	Pattern interrupt	15s demo	Learn More	Highest thumb-stop
Carousel 5-card	Curiosity	Feature/benefit per card	Sign Up	Highest engagement depth

Pitfalls to avoid

• Comparing formats from different campaigns. Auction conditions vary.
• Letting Meta auto-place into Stories only. Force All Placements for fair comparison.
• Reading carousel performance as a single number. Click depth across cards matters.
• Killing video on day 1 because CPM is high. Video CPMs settle by day 3.
• Forgetting that static ads can be repurposed cheaply. Even if video wins, keep statics in the rotation as low-CPM filler.

5 FAQs

Can I compare 1:1 video vs 9:16 video? Yes — that is a separate test (see ratio test post).

Should I test all three every cycle? At minimum static + video. Carousel only when you have enough variants per card to fill it well.

How many cards in a carousel? 5 is the sweet spot. 10 cards split CTR thin.

Are GIF ads considered video? Yes for billing, but they behave more like static. Test them in the static cohort.

Does video always have higher CPMs? Usually yes, but the higher engagement often offsets it on CPA.

All formats from one source

Pix-Vu helps you generate matched static, video, and carousel versions from the same source asset, so format tests are days, not weeks. https://pix-vu.com.