Ad fatigue: how to spot it before your CPA moves

What ad fatigue actually is

Ad fatigue is a drop in response caused by repeated exposure rather than by anything changing in the ad, the offer or the landing page. The creative that worked in week one is the same creative in week four. What changed is that the audience has now seen it several times.

It is worth separating from two things it gets confused with. An ad that never worked is not fatigued, it is a failed test. And an account where costs rise across every ad at once is usually an auction or seasonality problem, not fatigue.

The distinction that matters commercially: fatigue is not a production problem, it is a supply problem. One ad wearing out is normal and expected. Having nothing tested to replace it is the part that costs money.

The order the metrics move in

Fatigue does not hit every metric at once. It works down the funnel, which is what makes it possible to see early:

1. 1Three-second hook rate falls. Fewer people stop scrolling, because the opening frames are now familiar.
2. 2Hold rate follows. The people who do stop drop off sooner, since they already know where it goes.
3. 3Click-through rate softens. A smaller share of a smaller audience acts.
4. 4Frequency climbs, because delivery is pushing the same asset at a narrowing pool.
5. 5CPA rises. This is the last step, and the one most teams are watching.

The gap between step one and step five is where the opportunity sits. Hook rate is a creative signal and it moves first. CPA is a financial signal and it confirms what the creative signals already said.

Check hook rate and frequency at the ad level, not the campaign level. A campaign average hides a single fatiguing winner behind everything else that is running.

Thresholds worth acting on

Exact numbers vary by account, offer and audience size, so treat these as triggers to look rather than laws:

• Hook rate down roughly 20% from the ad's own best week, sustained over three days rather than one.
• Frequency climbing past about 3 on a prospecting audience, with reach flat or falling.
• Hold rate falling while spend on the ad is flat or rising.
• CPA still fine. That combination is the useful one, because it means you are early.

Compare an ad against its own history, not against the account average. A strong ad decaying from a high base can still look above average while it dies.

Why a refresh of the same concept usually fails

The standard response to ad fatigue is a new cut of the winner. New thumbnail, new first three seconds, same concept underneath. Sometimes that buys a week. Often it does nothing, and the reason is that the audience is not tired of the edit, it is tired of the idea.

Ten new assets that are seven variations of one concept is one test, not ten. The number that matters is net-new concepts, not files delivered.

What a real replacement looks like

A replacement has a different reason to work. Not a different colour grade, a different hypothesis about why someone stops and why they buy. If you cannot say in one sentence what the new ad tests that the old one did not, it is a refresh.

Treat format as a diagnosis, not a taste decision

When fatigue is driven by ad blindness, the useful move is a format the feed has no reference point for, because the scroll stops before the brain finishes classifying it as an ad.

Lead Printer had sat between 40 and 50 euros per qualified meeting booked for as long as their CEO could remember, and had stopped expecting it to move. The concept we sent was claymation, which is not an obvious choice for a print business. Same offer, same landing page, same budget behind it.

The point is not claymation. It is that the format was chosen from what the account needed, which was novelty, rather than from what looked good in a portfolio.

The fix is a queue, not a rescue

Every approach to ad fatigue that works comes down to the same thing: having the next test ready before the current one needs it. That is a scheduling problem more than a creative one.

• Read the account weekly at ad level, watching hook rate and frequency rather than cost.
• Write the next hypothesis while the current winner is still winning.
• Produce on a fixed cycle so replacements arrive on a known date, not when someone panics.
• Keep the winner running until something beats it. Fatigue is a reason to test, not a reason to pause.

We run that loop on a ten day cycle, which is where the name comes from. The length matters less than the fact that it is fixed: a queue that only fills when something breaks is not a queue.

Questions we get asked

What is ad fatigue?

Ad fatigue is a drop in response caused by people seeing the same ad repeatedly, rather than by any change to the ad, the offer or the landing page. It shows up first in thumb-stop and hold rate, then click-through, and finally in cost per acquisition.

How do you know if an ad is fatiguing?

Compare the ad against its own best week rather than the account average. A hook rate falling about 20% over three days, with frequency climbing past 3 on prospecting and CPA still acceptable, is the clearest early combination.

What frequency is too high on Meta?

There is no universal number, but frequency above roughly 3 on a prospecting audience, with reach flat or falling, usually means delivery is recycling the same people. Retargeting tolerates far higher frequency because the intent is different.

Does changing the thumbnail fix ad fatigue?

Rarely for long. A new thumbnail or opening cut can buy a week, but if the audience is tired of the underlying idea rather than the edit, the same decay resumes. A genuine replacement tests a different reason to stop and buy.

How often should you replace creative on Meta?

Often enough that a replacement is already tested when the current winner fades, which for most accounts spending consistently means a new set of concepts every one to two weeks. The cadence matters less than it being fixed rather than reactive.