A Shopify fashion brand we work with watched their ROAS drop from 3.9 to 2.1 over 6 weeks. They launched 3 new campaigns during that stretch. None of them reversed the decline. When we audited the account, the problem wasn't targeting, bidding, or budget — it was 4 creatives that had been running unchanged since January. By March, the same buyers had seen them 11 times each. The ads kept spending. The results kept falling.
The frequency number that signals creative fatigue
Most brands set a frequency cap of 3 or 4 and assume that's safe. It's not — not across 6 weeks. Frequency resets when you look at 7-day windows, but the buyer's brain doesn't reset. A buyer who saw your ad 3 times last week and 3 times this week has seen it 6 times, regardless of what your dashboard shows.
The number we track in every account is lifetime frequency on top-spend creatives. When a creative hits 6.0 lifetime frequency across warm audiences and spend is still being pushed behind it, you're paying to remind people why they already decided not to buy.
For the fashion brand, their 4 top creatives were averaging 8.3 lifetime frequency across retargeting and warm prospecting audiences. Meta kept spending because historical CTR was strong. The creative had trained the algorithm to spend — but the audience had already made their decision.
Why new campaigns don't fix an exhausted creative library
Launching a new campaign with the same creative doesn't solve fatigue. It restarts delivery on the same assets with a fresh learning phase — higher CPMs, same worn-out hook hitting the same worn-out audience.
The issue isn't campaign structure. It's that the creative library has no fresh assets to serve. Meta's delivery system will always surface the highest-historical-CTR creative in a given ad set. If the best performer is fatigued and there's nothing new behind it, you're paying premium prices for diminishing returns.
When we see this pattern, the first question we ask isn't "what campaign should we build?" — it's "what's in the creative library right now?" New campaigns only work when there's new creative to put in them.
What we found and what we changed
We audited the fashion brand's creative library and found 4 fatigued top-spenders and 6 secondary creatives that had barely run. The 6 secondary assets had low CTR not because they were bad — but because Meta's algorithm had never had a reason to serve them ahead of the stronger historical performers.
We paused the 4 fatigued creatives entirely. We pushed budget behind 3 of the 6 under-tested assets, structured as isolated ad sets at $40/day each for 5 days. Two of the three outperformed the original winners on thumb-stop rate. ROAS moved from 2.1 to 3.3 in 3 weeks — without producing a single new ad.
The fourth original creative had a strong hook but weak middle. We re-edited it — same opening 3 seconds, new product demonstration, sharper CTA. Relaunched as a fresh asset, it outperformed everything else in the account.
Pull your top 5 creatives right now and check this one metric
Open Meta Ads Manager. Go to your top 5 creatives by lifetime spend. Add the "Frequency" column. If any asset is showing above 5.0 lifetime frequency and still has daily budget behind it, it's actively hurting your ROAS.
Pause it. Don't delete — the performance data is useful. Pause it and move spend to assets with under 3.0 frequency. If nothing qualifies, that's the real problem: the creative pipeline is too thin. The fix is a 2-week production sprint, not a new campaign.