Guide
Nothing broke. The people who had not seen it ran out
Usually because the audience already saw it, not because the ad got worse. Four causes account for nearly all of it. Frequency. A change you made. Auction and seasonal shifts. Measurement drift. This page covers how to tell them apart, the order to check in, and what a replacement costs on the one occasion it is fatigue. By the end you will know whether to commission anything at all.
- Causes, and three are free to rule out
- 4
- Numbers that separate fatigue from everything else
- 2
- Tested creatives per expected winner
- 14
Read frequency and cost per result together, or you will fix the wrong thing
Frequency is average impressions per person. Cost per result is what you are paying for the outcome you care about. Neither is diagnostic alone.
Rising frequency with rising cost is fatigue. The fix is new creative or a new audience.
Flat frequency with rising cost is not fatigue. Something outside the creative moved. Producing new ads will cost you a fortnight and change nothing.
That second case is misdiagnosed constantly. New creative is the satisfying answer. Checking the account is the boring one.
Rising frequency with flat cost is a small audience that is still working. Watch it weekly and prepare a replacement rather than acting today.
What the two numbers say together
Only the first row is a creative problem. The other three cost nothing to check and are wrongly blamed on creative every week.
| What you see | What it means | |
|---|---|---|
| Frequency up, cost up | Same people, repeatedly | Fatigue. Replace the creative or the audience. |
| Frequency flat, cost up | Reach is unchanged | Auction, season, or a change you made |
| Frequency up, cost flat | Small audience, still converting | Not yet. Prepare a replacement. |
| Results vanish overnight | A step change, not a slope | Measurement or a settings change |
The first thing to read is your own change log, and most teams do not keep one
A budget jump, an audience edit, a bid change, a landing page tweak, a checkout change. Any of those moves results without touching the creative.
Without a dated log, the conversation becomes an argument about memory. The loudest theory wins rather than the correct one.
Keep one line per change with a date. It takes seconds and it converts a two hour debate into a lookup.
The landing page is the most commonly forgotten one. It belongs to a different team, and the ad account has no idea it moved.
Check the log before you look at anything else. It is free, it is fast, and it is right often enough to be embarrassing.
The order to check in, cheapest first
Three of these cost nothing. Producing new creative is the expensive answer, so it goes last rather than first.
Your change log
Free, five minutes
Frequency and reach
Free, in the account
Auction and season
Free, compare to last year
Replace the creative
The expensive one
This is the whole editor
Highlight a phrase and a clip lands on those exact words. No timeline, no keyframes, no layers.
Auction pressure and seasonality look identical to fatigue on a chart
Costs rise when more advertisers want the same attention. That happens on a calendar you do not control. It raises cost per result with no change in frequency.
Check whether the whole account moved or only one ad set. An account-wide shift is almost never a creative problem.
Seasonality is the same shape from a different cause. Demand moves under you and the creative gets blamed for a change in what people wanted that week.
Compare against the same fortnight last year rather than against last month. Comparing to last month is what makes every autumn look like a crisis.
Measurement drift is the fourth. A tracking change makes results disappear that are still happening. That produces a step change rather than a slope. Confirm on a second source before producing anything.
A fifth cause is invisible inside your account entirely. A competitor changed their offer, so the same ad now argues against a better deal than it did last month.
Nothing in your reporting shows that. The only way to see it is to look at what competitors are currently running. It takes twenty minutes and almost nobody does it.
Check that before you commission anything. A creative fix for an offer problem is expensive and it reliably fails.
When it is fatigue, the real failure was not having a replacement ready
Fatigue is inevitable. Every creative has a finite pool of first impressions. The question was never whether it would decline.
The failure is that the next one is not ready on the day. Spend keeps running behind an ad everybody knows is finished.
Published benchmarks put the winner share at 5 to 8 percent, per Motion's analysis of 550,000+ Meta ads. That is roughly fourteen tested creatives per winner.
A team producing six a month draws a winner every two or three months. Fatigue arrives faster than that. The gap is what a creative crisis actually is.
The cheapest replacement is a new opening rather than a new ad. When hook rate falls and hold holds, the entrance went stale before the argument did.
In Cutroom that is a transcript action. Delete a line and the cut rebuilds. Change the pace and it re-cuts. Trim the opening and export again at 20 credits per output minute.
Fourteen tested creatives, one expected winner
At the middle of the published 5 to 8 percent range. Fatigue arrives on its own schedule, and this grid decides whether you have an answer.
If the ad never worked, none of this applies and volume will not save you
There is a difference between decline and an ad that had one good week on noise. Check whether the early result had the volume to mean anything before you mourn it.
If nothing in the account has ever worked, the problem is upstream of creative. The offer decides more than every production decision combined. No amount of new video fixes a price nobody wants.
Producing more attempts multiplies whatever the offer already does. Against a weak offer, that buys fourteen expensive confirmations.
Fix the offer, then buy attempts. That order is unpopular because the offer is somebody else's decision. That is exactly why it survives untested for so long.
Once the offer is right, the attempts are the whole game. A batch is 100 credits, an export is 20 credits per output minute, and the replacement is ready before the numbers turn.
Questions people ask
- How quickly should I react to a decline?
- Slower than instinct suggests. Daily numbers on a small budget are mostly noise, so read a rolling week and compare to the previous one. Reacting to a bad Tuesday is how accounts get restructured for no reason.
- Should I raise the budget on an ad that is declining?
- Raising budget accelerates frequency, which accelerates the decline you are trying to fix. If the ad is fatiguing, more spend reaches the same people faster. Replace it or widen the audience instead.
- Can I bring an old winner back later?
- Sometimes, and it is worth trying because it costs an upload. Audiences refresh over months, and a creative that fatigued in March can perform again in September. Treat it as a new test with a real verdict rule.
- Is new creative always the right answer?
- Only when frequency is climbing alongside cost. Flat frequency with rising costs is not a creative problem, and a fortnight spent producing new ads will confirm it expensively. Read the change log, the auction and the tracking first.
Check the free causes before the expensive one. Then keep a replacement standing by, because Cutroom makes that a new opening and a 10 credit export rather than a production schedule.