Your best ad stops working. CTR sags, cost per result creeps up, and nothing in the account settings explains it. That is creative fatigue, and since Meta’s Andromeda rollout it arrives faster than it used to. Timelines that ran six to eight weeks in 2024 now commonly run two to four, and for high-budget accounts on Advantage+ delivery, sometimes less.
The good news is that fatigue is the most diagnosable problem in Meta advertising. It has specific symptoms, measurable thresholds, and a fix ladder that works. This guide covers all three, plus the prevention system that keeps you off the treadmill of panicked Friday-afternoon creative swaps.
What creative fatigue is, and what it is not
Creative fatigue is the performance decline that happens when the same people see the same creative elements too many times. Engagement falls, the delivery system compensates by paying more for the same results, and your costs rise with no settings change to blame.
Three lookalike problems get misdiagnosed as fatigue, and the fixes are different, so separate them first:
- Audience saturation. The pool itself is exhausted. New creative recovers performance for days, then sinks again. The fix is a bigger audience, a new market, or lower spend, and no amount of fresh creative solves it.
- Delivery and tracking issues. Performance falls off a cliff overnight rather than eroding. Check pixel and Conversions API health, event deduplication, and account-level flags before you blame the ad.
- Seasonality and competition. Everything in the account declines together, including brand-new ads. Fatigue is ad-specific by definition. If your whole account dipped in the same week, look outward, not at the creative.
The clean test: replace the suspect ad with a genuinely new concept. If performance recovers within three to five days, it was fatigue. If it does not, you have one of the other three problems.
The diagnostic thresholds
These are the signals I check, in the order they typically appear. The thresholds combine Meta’s own delivery statuses with the working numbers creative analytics platforms like Atria publish, and they match what I see in practice.
| Signal | Threshold | What it tells you |
|---|---|---|
| Hook rate on video | Falling week over week | Earliest warning. The opening has stopped stopping thumbs. |
| CTR | Consistent decline across 7 days | The ad still gets seen but has stopped earning clicks. |
| Frequency | 2.5 to 3.0+ on prospecting | The same people are seeing it repeatedly. On retargeting, tolerances run higher. |
| First-time impression ratio | Dropping steadily | Delivery is running out of new people to show this ad to. |
| Cost per result | Trending up with no settings change | The system is compensating. Fatigue is now costing you money. |
| Meta’s delivery column | “Creative limited” then “Creative fatigue” | Meta’s own confirmation. The fatigue label appears when cost per result runs well past your historical baseline. |
By the time Meta’s delivery column says “Creative fatigue,” you have usually been paying the fatigue tax for a week or two. The column is a confirmation, a lagging one. The earlier rows are where the money is saved. Check hook rate and CTR trends weekly, per ad, and you will catch most fatigue before the cost curve does real damage.
Why Andromeda made fatigue faster
Meta’s Andromeda retrieval system finds your best ad quicker and concentrates delivery on it harder. Concentration is efficiency and exhaustion in one mechanism: the more precisely the system routes your winner to its ideal audience, the faster that audience gets used up. Faster winner identification, faster winner burnout. I unpacked the mechanics in my Andromeda explainer, but the operational consequence fits in one sentence: creative refresh went from a monthly chore to a core competency.
There is a second-order effect worth knowing. Because Andromeda clusters conceptually similar ads together, swapping a fatigued ad for a near-duplicate does less than it used to. The system largely recognizes it as the same idea the audience is already tired of. Which is why the fix ladder below distinguishes hard between variants and new concepts.
The fix ladder
Within 24 to 48 hours
When an ad crosses the thresholds, act on the ad, and only the ad:
- Pause the fatigued ad, never the campaign. The campaign holds learning you want to keep.
- Shift its budget to the healthiest concepts still under frequency 2.5.
- Refresh the hook first. New opening three seconds on video, new headline and first line on statics. Hooks fatigue fastest and are the cheapest thing to rebuild. A new cover frame on the same video buys real time.
- Exclude recent engagers if frequency is the loudest signal and the audience has room to spare.
These moves stop the bleeding. They do not cure anything. Hook swaps and cover changes are variants, and post-Andromeda a variant buys you weeks at best.
The structural fix
The cure is new concepts entering the account on a rhythm, so that no single ad ever has to carry delivery long enough to burn out. Working cadences:
- Direct response prospecting: new concepts every one to two weeks.
- Brand and awareness: every three to four weeks.
- Retargeting on small pools: most aggressive of all, because small audiences hit frequency ceilings in days.
Refresh priorities, in order of how fast each element fatigues: hooks first, visual treatment second, CTA copy last. Audiences get tired of an opening long before they get tired of a button.
The five-step diagnostic workflow
When something in the account looks tired, run this sequence per suspect ad. It takes ten minutes and prevents the two expensive mistakes: refreshing creative that was not fatigued, and blaming creative for a problem that lives elsewhere.
- Rule out the account-level explanations. Did anything decline account-wide this week, including new ads? If yes, stop; it is seasonality, competition, or tracking, and no creative refresh will fix it.
- Pull the ad’s own trend lines for 14 days: hook rate, CTR, frequency, first-time impression ratio, cost per result. Fatigue shows as erosion across several of them, in roughly that order. A cliff on one day is a delivery or tracking event, a different investigation.
- Check the delivery column. “Creative limited” means Meta sees the early stage. “Creative fatigue” means you are late and the ad is already running well past your usual cost per result.
- Check spend share. An ad carrying most of the campaign budget fatigues on a compressed clock. If your top spender is at frequency 2.8 and falling CTR, the question is when to rotate, no longer whether.
- Confirm with substitution. Introduce a genuinely new concept against the same audience. Recovery inside three to five days confirms fatigue. No recovery points back to saturation or external factors, and saves you from burning your creative bench on the wrong problem.
The production capacity math
Rotation systems fail in the production pipeline before they fail in Ads Manager, so size the pipeline deliberately. The arithmetic for a typical direct-response account: seven live concepts, a two-to-three week realistic fatigue cycle at meaningful spend, means roughly 10 to 14 new concepts a month to hold steady, before any variants for winners. At larger budgets the same structure burns faster, because concentration accelerates with spend.
Three ways accounts make that number affordable rather than heroic:
- Vary cost per concept on purpose. A monthly mix of two produced videos, four UGC or founder-style pieces, and six statics or text-led formats covers a wide dimension grid at a fraction of an all-video budget.
- Convert winners across formats. Every proven angle can be re-expressed as a static, a carousel, and a short video. Same insight, three concepts, near-zero new thinking required.
- Maintain the bench. Two or three concepts, already tested at low spend and parked, turn fatigue events from emergencies into Tuesday. The bench is the difference between rotating on your schedule and rotating on the algorithm’s.
Prevention: the rotation system
Accounts that never show up in my inbox with fatigue emergencies share one structure. They run five to ten genuinely distinct concepts at all times, they stagger launches so something new enters every week or two, and they hold two or three tested concepts on the bench, ready to promote the day a starter crosses the thresholds. Nothing fatigues catastrophically because nothing carries the account alone.
Building that bench requires a production pipeline that reliably outputs distinct concepts, which is a creative-strategy problem before it is a media-buying problem. My creative diversity playbook covers what makes concepts genuinely distinct in Andromeda’s eyes and how to plan production math around a two-to-three week refresh cycle. The one-line version: count your concepts, not your ads. Twenty variants of one idea fatigue as one idea.
Frequently asked questions
What frequency is too high for Facebook ads?
For prospecting, treat 2.5 to 3.0 as the warning zone and investigate the trend rather than the snapshot. Retargeting runs hotter by design; there, watch cost per result and CTR decline instead of frequency alone.
How long do Meta ads last before fatigue now?
Post-Andromeda, two to four weeks is typical for direct-response prospecting creative, with high-spend accounts on Advantage+ delivery sometimes seeing meaningful decline inside two weeks. Spend concentration shortens the timeline: the more budget an ad carries, the faster it burns.
Does refreshing the thumbnail reset fatigue?
It helps briefly. A new cover image or hook is a variant, and the delivery system still treats the underlying ad as substantially the same idea. Use it to buy time while a genuinely new concept ships.
Should I duplicate the ad to reset frequency?
No. The duplicate inherits the same conceptual fingerprint, reaches largely the same people, and restarts learning on top of it. This tactic mostly died with the Andromeda rollout.
Is ad fatigue the same as creative fatigue?
The terms get used interchangeably, and Meta’s own delivery column uses “creative fatigue” for the confirmed state. If you want a distinction: ad fatigue is the umbrella for any repetition-driven decline, creative fatigue is the asset-level version you fix with new creative. The diagnosis and fixes in this guide cover both.
Can Advantage+ campaigns fatigue faster than manual ones?
Yes, and predictably so. Advantage+ delivery concentrates spend on winners more aggressively than manual structures, which is good for short-term efficiency and hard on creative lifespan. Feed Advantage+ structures a deeper bench and expect rotation on the faster end of the cadences above.
What should I refresh first when an ad fatigues?
The hook. Opening three seconds on video, headline and first line on statics. It fatigues fastest, costs least to rebuild, and a strong new hook on a proven body regularly buys two or three extra weeks. Visual treatment comes second, CTA copy last; audiences tire of buttons slowest.
Sources: Meta Ads Manager delivery status documentation, Atria 2026 creative fatigue guide, AdExchanger (January 2026), Social Media Examiner on 2026 algorithm changes, Meta Engineering (December 2024).
