The signs of ad fatigue arrive in a predictable order: hook rate falls, frequency climbs, CPM rises, CTR drops, and only then do CPA and ROAS move. The sign most teams alert on, rising CPA, is the last one in the sequence and the most expensive to wait for. If you build your alerts around the first two, you get roughly one to two weeks of warning before the auction sends you the bill.
Which signs come first, and how much warning each gives
| Signal | What it is telling you | Lead time before CPA moves | Threshold worth alerting on |
|---|---|---|---|
| Hook rate falls | Attention is decaying before the click, on the creative itself | 1 to 2 weeks | 15 to 20 percent below this ad's own first-week hook rate |
| Frequency climbs | The same people are seeing the same version repeatedly | 1 to 2 weeks | 2.5 weekly on Meta prospecting, about 3.0 on TikTok |
| Saves, shares and comments dry up | Organic amplification has stopped subsidising your reach | About a week | Any sustained drop while spend is flat |
| CPM rises with no targeting change | The auction is pricing your creative as less relevant | 3 to 10 days | 15 percent or more above the ad set's own trailing average |
| CTR falls | The creative is now inside the measured decay curve | Days, and often simultaneous | 15 to 20 percent below launch in the first two weeks |
| CPA rises, ROAS falls | Fatigue has already been paid for | Zero, this is the arrival | Anything past target, and by then it is late |
Two things about that table are worth arguing with. The first is that hook rate and frequency are the only rows you can act on cheaply, because at that point the ad is still profitable and you have time to produce. The second is that the last row is not a warning at all. It is a receipt.
Why hook rate and frequency are the early pair
Hook rate is the percentage of impressions that stay past the opening beat, and it is the most creative-specific number you have. It moves before CTR because the decision to keep watching happens earlier than the decision to click, and because a proven ad can keep harvesting clicks from a shrinking pool of viewers for a while. On Meta feed we typically see 15 to 25 percent, and 20 to 30 percent on TikTok in-feed, but the useful comparison is never the benchmark. It is this ad against its own first week.
Frequency is the other half of the pair, and it is the leading indicator with the clearest causal story. Our 2026 fatigue benchmark puts the start of decline at a weekly frequency of 2.5 on Meta prospecting, and cites Meta internal research finding CTR drops 45 percent after a fourth exposure to the same creative. Frequency is impressions divided by reach, so it is also the only fatigue signal you can forecast: if you know your budget, CPM and audience size, you know roughly which day you will cross 2.5, and our ad fatigue calculator will do the arithmetic for you.
Watch them together, because they fail differently. Hook rate falling while frequency is flat usually means the creative was weaker than you thought, or the audience shifted. Frequency climbing while hook rate holds means you have room left, but you should already be producing. Both moving at once is fatigue, and there is no third interpretation worth entertaining.
Why CPA is a lagging indicator
CPA is a ratio computed after impressions, clicks and conversions have all accumulated, so it inherits every delay in that chain plus your attribution window. A creative can shed 15 to 20 percent of its CTR across its first two weeks, which our benchmark report identifies as the normal opening decline, and still post an acceptable CPA because a slightly better conversion rate or a run of returning customers masks it.
Then week three arrives. Our benchmark report describes it as a cliff: 45 to 70 percent below the launch baseline, with a creative averaging 38 percent below its peak by week five. Most creative is effectively dead within three weeks. CPA typically breaks target somewhere in that window, which means the alert fires one to three weeks after the problem started and after you have paid full price for the decline.
There is a second reason not to trust CPA as your fatigue signal: it is not creative-specific. CPA moves when the auction gets more expensive, when a landing page breaks, when a tracking change eats conversions and when a competitor enters your category. Hook rate does not. If you want a signal that means fatigue rather than something-is-wrong, watch the one attached to the video.
What to do at each stage
Hook rate below its own first week: change the opening, not the offer
Rebuild the first three seconds against the same proposition. A new opening line, a different first shot, different on-screen text. This is the cheapest intervention available and the one with the best odds, because the part of the ad that has been learned is usually the part the viewer sees before deciding.
Frequency crossing the ceiling: promote, do not broaden
Move budget to a variation the audience has seen less. Broadening the audience at this point buys about a week of pacing and drags in worse-qualified people. Keep 8 to 20 live variations in rotation on an active campaign so there is always somewhere for that budget to go.
Organic engagement flatlining: the concept is spent, not the cut
When saves and shares stop while spend is flat, people are no longer finding the idea worth passing on. A new hook on the same concept will not fix that. This is the point to ship a genuinely different claim.
CPM up with no targeting change: check the auction before you blame the ad
Rising CPM is ambiguous. It can mean your creative is being priced as less relevant, and it can mean a seasonal auction squeeze that has nothing to do with you. Compare the ad set against the account and against the same placement last month before you spend a production cycle on it.
CTR down 15 to 20 percent: you are on the clock
The decay curve has started. Do not wait for the week-three cliff. Put a new concept live alongside the incumbent now, at low spend, so it has learning behind it when you need to promote it.
CPA past target: stop the bleed, then fix throughput
Pause the highest-frequency ads first rather than the worst-CPA ones, since frequency is the cause and CPA is the symptom. Then look at your production rate honestly. Our benchmark report's conclusion is that throughput, not talent, is the bottleneck: brands shipping 15 to 50 creative variants a month see 3 to 5 times longer campaign lifespan than quarterly refreshers.
The one-minute daily check
Open the ad-level view, sort by spend, and look at four columns for your top three ads: hook rate against that ad's own first week, weekly frequency, CPM against the trailing average, and days since launch. That last column matters more than people expect, because creative fatigue is largely a function of exposure count, and days since launch is a decent proxy for it when your budget is stable.
We built Genyad for the production half of this problem, not the measurement half. It turns footage you already own into new video variations, so a hook rate alert can become three new openings the same afternoon. It does not report on your ads: there are no predicted performance scores, no direct publishing to Meta or TikTok, and no dashboard. Every signal above you read in your own platform reporting.
Frequently asked questions
What is the earliest reliable sign of ad fatigue?
A falling hook rate on a creative whose spend and audience have not changed. It typically moves one to two weeks before CPA and is specific to the creative rather than to the auction. Compare an ad to its own first-week hook rate, not to a benchmark, since a good hook rate for one vertical is a poor one for another.
At what frequency should I refresh creative?
Weekly frequency of 2.5 on Meta prospecting is where decline begins according to our benchmark report, and above about 3 in general marks fatigue arriving. Treat 2.5 as the trigger to start producing rather than the trigger to pause, because production takes longer than delivery does.
Can rising CPM be a sign of ad fatigue?
Yes, but it is the least specific signal in the set. Platforms price less relevant creative higher, so fatigue does raise CPM, and so do seasonal auction pressure, audience changes and competitors bidding into your category. Check the ad set against the rest of your account before attributing it to the creative.
How long do I have between the first sign and real damage?
Roughly one to two weeks if you catch the hook rate and frequency pair. Our benchmark report puts CTR 15 to 20 percent below baseline in the first two weeks and 45 to 70 percent below by week three, so the gap between the early signals and the cliff is about a fortnight.