When CPMs climb and ROAS slips, the instinct is to rebuild audiences. Most of the time, that’s the wrong lever. On Meta today, the placement, the audience and the bid strategy are doing their job — it’s the creative that’s exhausted.
Why creative fatigue happens faster now
Advantage+ Shopping campaigns and broad targeting recycle the same audiences across more accounts than ever. A single creative now hits frequency 4+ inside the first week of spend at modest budgets. That used to take a month.
The three-tier funnel, refreshed
A 2026 creative system should produce assets at three distinct levels of intent, with separate metrics for each.
- Top — hook tests. Short, 0–3 second cuts designed only to win the stop. Measure thumbstop rate, not ROAS.
- Middle — proof. UGC, founder pieces, demos, social proof carousels. Measure hold rate and add-to-cart.
- Bottom — offer. Static or short video stating the promise, price and CTA. Measure ROAS and cost per purchase.
How much creative do you actually need?
Aim for four fresh hooks and two proof pieces every two weeks per active account. That’s the volume at which Meta’s algorithm has enough signal to actually test, but it’s also achievable for a small business with one creator and a shared template system.
The tested testing process
Don’t test creatives in your scaling campaign. Run a separate CBO with a $20–50/day budget per asset for 3–5 days, kill anything below your CPA threshold, then graduate winners into your scaling ad set. This protects learning and keeps your main campaign calm.
What to refresh first
If you only do one thing this month: replace the hooks. The first three seconds carry 80% of the cost of failure on Meta. New hooks on existing winning creative will lift performance more than any audience change.