Creative fatigue is discussed constantly with confident numbers attached, and exactly one defined threshold exists for it. Meta wrote that definition into Ads Manager, and it measures cost rather than frequency or elapsed time.
Once you know the real definition, most of the advice on this subject stops making sense, which is useful.
This is for you if somebody has told you to refresh your adverts on a schedule and you would like to know whether that is based on anything.
The one real definition of creative fatigue
Meta’s own product defines it, which makes this the only threshold with an author.
Ads Manager labels a creative as showing Creative Fatigue when cost per result has reached at least twice its past level. It applies a separate label, Creative Limited, when cost per result is higher than before but has not doubled.
Two things about that are worth pausing on.
It is measured in cost, not in frequency, impressions or days. Nothing about how many times people saw the advert enters the definition.
And doubling is a high bar. A creative producing results at fifty percent above its old cost is not fatigued under Meta’s definition, it is limited. Most advice would have you replace it well before that point.
Why frequency is not the definition
Because the party with the data declined to publish one.
Facebook analysed frequency across 2,439 Brand Lift campaigns between November 2019 and March 2020. It confirmed effectiveness plateaus after a handful of impressions, then published no threshold number.
Meanwhile the industry publishes several. Above 2.5 means fatigue. Cap at three. Refresh every two weeks. None trace to Meta or to any dataset with a disclosed method.
So somebody with thousands of campaigns of evidence would not name a number, and people with none named many. That should settle how much weight those figures deserve.
There is a mechanical reason frequency misleads here too. As frequency rises, impressions grow while reach flattens, so every metric divided by impressions falls automatically. That decline looks exactly like creative fatigue and can be pure arithmetic.
The signature worth learning
Fatigue has a shape, and checking the shape prevents most wrong diagnoses.
Genuine fatigue shows rising cost per result, rising frequency, and falling click-through rate together. All three move, in that pattern, over a period rather than in a day.
Rising cost with flat frequency is something else. Auction pressure, a seasonal shift, a landing page change, or a tracking fault. Replacing the creative will not fix any of those, and plenty of accounts replace it anyway.
Falling click-through rate with flat cost per result is usually the denominator effect described above rather than decay.
So check all three before concluding. The cost is what matters, the frequency tells you whether exhaustion is plausible, and the click-through rate confirms the direction.
Preventing creative fatigue rather than managing it
I would rather it did not arrive than have a process for handling it, and the prevention is not complicated.
Low creative count is the most common cause of fatigue arriving early. When two adverts carry an entire audience, the same people see the same thing repeatedly and exhaustion is inevitable.
Run enough distinct creatives that no single one carries the whole audience, and the problem largely does not develop. That is the same advice as running a proper test, which is convenient, because the same structure solves both.
The word doing the work there is distinct. Six variations of one image do not spread exposure meaningfully, because the audience experiences them as the same advert with different colours.
Prevention is cheaper than diagnosis here, and considerably cheaper than an emergency creative sprint when an account starts sliding.
What adding creatives actually did on one account
The clearest example I have of prevention working retroactively.
An account had been running too few creatives for too long, with cost per lead sitting around $70. The set was starved of options and the audience had seen everything repeatedly.
Adding creatives brought cost per lead down to between $30 and $40 within days.
Not weeks. Days. Nothing else changed: same audience, same offer, same budget, same landing page. The only variable was giving the system more to choose between and the audience something they had not already seen.
That speed is worth knowing because it changes the diagnosis. If adding creatives fixes a rising cost within days, the problem was exhaustion rather than the market or the offer.
Why refresh schedules fail
The calendar is the wrong trigger, and following one produces two opposite errors.
You replace creatives that are still working, because a fortnight elapsed. That throws away accumulated performance and puts the set back into unstable delivery for no reason.
And you keep creatives that stopped working, because the fortnight is not up. Cost climbs while everybody waits for the scheduled review.
Neither happens if the trigger is the measurement. Watch cost per result against its own history, use Meta’s doubling threshold as the outer marker and its Limited label as the early warning, and act when the number says so rather than when the calendar does.
The refresh cadence that matters is production, not replacement. Keep making new angles steadily, and let the performance decide when each one retires.
What creative fatigue is not
Two things get called fatigue and are not, and both waste real money.
A creative that never worked is not fatigued. It failed. Those need different responses, and treating a failure as exhaustion leads to producing more of the same thing.
A whole account declining is not creative fatigue either. When every campaign slides at once, look upstream at tracking, at the landing page, at seasonality or at a competitor entering the auction. Fatigue is creative-specific by definition, so a uniform decline points elsewhere.
Checking whether the decline is universal or specific takes a minute and rules out the expensive misdiagnosis.
What to change this week
Three steps.
Find the Creative Fatigue and Creative Limited labels in Ads Manager and use them, because they are the only defined thresholds available and most people have never looked.
Then count the distinct creatives in each ad set. If any set is running fewer than four, that is your fatigue risk regardless of current performance.
Then drop any calendar-based refresh rule and replace it with the cost trigger, so you stop retiring creatives that are still earning.
Creative benchmarks are mostly invented covers the wider pattern behind the missing frequency threshold, thumbstop rate is three different numbers covers why a falling rate can be pure arithmetic, fund the ad set and read what the algorithm chose is the position underneath, and scaling Facebook ads covers what happens to fatigue as budgets grow. Preventing it properly is a production question, which my Meta Ads work plans around. To have an account checked for genuine fatigue rather than assumed fatigue, book a teardown.