Ad fatigue is the point where a creative stops earning its impressions. Nothing about the targeting changed, the budget is the same, the offer is the same — and the numbers quietly get worse. Most teams notice it two weeks late, after the cost per acquisition has already drifted, and then respond by moving budget around instead of fixing the thing that actually broke: the creative.
What fatigue actually is
Fatigue is not the audience getting bored in some vague sense. It is a measurable decay in the response rate of a specific creative against a specific audience, driven by repeated exposure. The same ad that converted a cold viewer on impression one is wallpaper by impression six.
Two things follow. First, fatigue is creative-specific, not account-specific — one asset can be dying while another in the same ad set is fine. Second, it is reversible only by changing what people see, not by changing how much you pay to show it.
The signals, in the order they appear
Fatigue announces itself in a fairly consistent sequence. Watching it in order is what buys you the extra week.
- Frequency climbs — the leading indicator. On cold prospecting, commonly cited warning thresholds sit somewhere around 2 to 2.5 on a seven-day window, with performance falling away sharply beyond 4. Retargeting tolerates considerably more, since intent is already established.
- CTR falls — the confirmation. A sustained decline over several consecutive days, with nothing else changed, is the clearest single read. A drop of roughly 20% or more held over a couple of weeks is generally treated as settled.
- CPM creeps up — the auction responding to weaker engagement signals, usually without any external event to explain it.
- Conversion rate goes last — by the time CVR moves, you have been paying for fatigue for a while.
Those numbers are starting points, not laws. They vary by vertical, audience size, funnel stage and account. The right thresholds are the ones you derive from your own historical curves — but you need some trigger written down, because the alternative is arguing about it in a Monday meeting while spend continues.
Why cycles are shorter than they used to be
Two forces have compressed creative lifespans. Platform ranking systems now lean harder on creative signals when deciding what to serve, so a weak asset is deprioritised faster than it once was. And broad targeting has become the default, which means a single concept burns through a large audience quickly instead of trickling through a small one. A concept that comfortably ran for six weeks a few years ago may now be visibly tired in two or three.
The mistake: treating fatigue as a media problem
The standard response to a fatiguing ad set is to fiddle with the media plan — cap the frequency, expand the audience, shift the budget, tweak the bid. All of these buy days, not solutions. The audience is not the constraint; the asset is.
The deeper problem is that most teams have nothing ready to swap in. Fatigue is only an emergency when the replacement pipeline is empty — and an empty pipeline turns a predictable creative event into a scramble that ships whatever happens to be finished.
Build the bench before you need it
The teams that handle fatigue well are not better at detecting it. They are better at having somewhere to go when it happens. That means a standing bench of validated concepts — not more variations of the winner, but genuinely different angles — scored before they ever reach the auction.
This is where pre-launch evaluation earns its place. Testing in-market is honest but expensive: every fatigued asset you replace with an untested one is a fresh bet paid for in spend. Scoring creative before launch narrows the field first, so the replacement you promote is one you already have reason to believe in. Our guide to creative testing covers the mechanics, and the 13 dimensions of creative effectiveness covers what a score is actually measuring.
A simple operating rhythm
- Write the trigger down — one frequency threshold and one CTR-delta threshold per funnel stage, agreed in advance.
- Review weekly, at creative level — not at campaign level, where a dying asset hides behind a healthy one.
- Keep three scored concepts in reserve — different angles, not different crops of the same image.
- Rotate on the trigger, not the calendar — killing a creative that is still working is its own kind of waste.
- Log what died and why — fatigue patterns are the cheapest creative research you will ever get.
FAQ
Is high frequency always bad?
No. Frequency is a leading indicator, not a verdict. Retargeting audiences reliably absorb far more exposure than cold ones, and some considered-purchase categories need repetition. Frequency matters when it climbs and engagement falls.
Can I just change the headline instead of the whole ad?
Sometimes, and it is worth trying — copy is cheap to change. But if the visual is what people have stopped seeing, new words on the same image rarely resets anything. See our note on copy testing for where that line falls.
How often should we refresh creative?
Top-of-funnel needs the fastest cadence and retargeting the slowest, but the honest answer is to refresh on the trigger rather than the schedule. A fixed calendar will always be either too early for your winners or too late for your losers.
Does more budget fix fatigue?
It makes it worse. More budget against a fatiguing asset means more impressions to the same people, which is exactly the mechanism causing the decline.
The short version
Fatigue is not a failure — it is what happens to every creative that works. The cost is not in the decline itself but in noticing late and having nothing ready. Set a trigger, watch at creative level, and keep a scored bench standing. You can score an ad in about 90 seconds and find out whether your replacement is actually a replacement.
