Guide
Fatigue is a curve, and its half-life is a quota
Creative fatigue is a decline caused by repeated exposure, with budget and targeting held roughly constant. Every clause in that sentence rules out a decline the word gets blamed for. Here is how to plot your own curve against creative age, read the half-life off it, and convert that half-life into a monthly quota. You will finish with a number instead of an argument.
- Creatives before your own decay curve means anything
- 12
- Notice the three-second view rate gives you
- 1 week
- Replacements a month at four live slots and a one-week half-life
- 16
Every clause of that definition rules out a decline it gets blamed for
Decline, so you need a baseline. Caused by exposure, so auction and audience effects have to be ruled out first. Budget held constant. A creative that gets worse when you triple its spend is being asked to reach further.
Without that discipline, fatigue becomes the label attached to any decline nobody can explain. That makes it useless for planning.
One consequence is worth stating plainly. Fatigue is a property of the pairing between a creative and an audience. It is not a property of the file.
The identical export shown to a fresh audience performs the way it originally did. That is why the same ad can be exhausted in one market and untouched in another.
Fatigue, and the three declines it gets blamed for
If the answer in the right column is no, the word fatigue is doing no work and a new video will not help.
| What actually changed | Does new creative fix it? | |
|---|---|---|
| Fatigue | The same people have seen it too often | Yes, and a new opening is usually enough |
| A budget increase | You are reaching further into the audience | No. The creative is being asked to travel |
| An auction shift | Other advertisers raised the price of attention | No. It is a margin decision |
| A spent pool | The people who would buy have bought | No. Widen the audience or change the offer |
Plot against creative age, not the calendar, and the shape appears
One change makes fatigue visible. Stop plotting metrics against the calendar. Plot them against days since the creative launched.
Give every creative a day zero, then chart cost per acquisition, click-through rate and three-second view rate against creative age.
Ten creatives that launched on ten different days can now be laid on top of each other. The shape of your own decay appears.
Normalise each metric to that creative's own first-week average so a strong ad and a weak one sit on the same scale. You are looking for the shape, not the level.
A dozen creatives is enough to see it. Fewer and one odd launch week bends the median far enough to mislead you. A misleading curve is worse than no curve.
- Day zero per creative, then measure age in days.
- Normalise each metric to that creative's first-week baseline.
- Overlay the batch. The median curve is your account's fatigue profile.
- Track frequency on the same chart so you can separate exposure from everything else.
This is the whole editor
Highlight a phrase and a clip lands on those exact words. No timeline, no keyframes, no layers.
The half-life converts into a refill rate you cannot argue with
Read off the point where performance has fallen halfway from its opening level towards your kill threshold. Call that the half-life.
For many accounts it lands between a week and a month. The exact figure matters far less than knowing yours.
It converts directly into a production quota. Creatives half spent in three weeks, with three needed live at once, means a replacement roughly every week. That is a refill rate, not an opinion.
It also tells you when to stop optimising. Bidding tweaks on a creative past its half-life are rearranging deck chairs. The budget belongs on the next one.
One hour of work changes the Monday meeting. Nobody debates whether the creative feels tired. You read off a chart that says this one is at day nineteen of twenty-one.
Rebuild the curve every quarter and date it in the filename. A half-life measured in March and still quoted in November is a memory rather than a measurement. Audiences turn over faster than reporting habits do.
Replacements a month, at four live slots
Read your half-life off the curve, then read your quota off this. The short half-lives are where plans break.
- One-week half-life16 a month
- Two-week half-life8 a month
- Three-week half-life5 a month
- Four-week half-life4 a month
The view rate turns about a week before the money does
Attention metrics move before money metrics. That is why cost per acquisition is a poor early-warning system. By the time it moves you have already spent the money that told you.
The usual order. Three-second view rate softens first. Recognition is instant, and scrolling is the cheapest response there is. Then hold rate through the middle. Then click-through. Then cost per acquisition, last and loudest.
Watch the first one. A view rate drifting below the creative's own baseline while frequency climbs is fatigue announcing itself with about a week of notice.
That week is enough time to ship a replacement, if your production can move in a week.
An early-warning system is worth exactly as much as your ability to respond to it. A team on a fortnightly production cycle has bought a smoke alarm for a house with no exit.
Which line turns first, and how much warning each gives
Everybody watches the last box. It is the one that reports money you already spent finding out.
Three-second view rate
Turns first. About a week of notice
Hold rate to the middle
Confirms it is the creative, not the hook
Click-through rate
Slower, and noisier on small budgets
Cost per acquisition
Last and loudest. Too late to act cheaply
A week of notice is only worth having if you can ship in a week
Measurement only pays if you can act on it. A team that diagnoses fatigue precisely and produces one creative a month has bought a decline with a chart attached.
The response to a known half-life is a standing cadence sized to the refill rate the curve implies. One finished ad has to be cheap enough to sustain indefinitely.
Cutroom is built for that cadence. One talking-head take of up to three minutes goes in. A finished 9:16 MP4 comes back, cut, captioned and covered with footage. Every change is made by marking the transcript.
That closes the gap the curve opens. The view rate turns on a Monday. You record on Monday evening. The replacement is live on Tuesday. A two-day editing queue cannot do that, and a shoot day is not even close.
The cheapest replacement is often not a new concept at all. Trim the opening. Delete a line. Change the pace. A proven body gets a fresh first two seconds for 20 credits per output minute.
The facts to plan around. Somebody has to be on camera. Three minutes is the cap. Every export is a 9:16 MP4. Deciding what the replacement should say is still yours, and on a fatigued account that is where the win is.
Questions people ask
- Is rising frequency the same thing as fatigue?
- No. Frequency is exposure, fatigue is the performance response to exposure. Two creatives at the same frequency can decay at different rates, and the difference between them is the most useful creative lesson in your account.
- Can a creative fatigue and then recover?
- Genuine exposure fatigue can partly recover after a rest if the audience has grown or turned over. Recovery is usually partial and short-lived, so treat a rested creative as a bonus rather than a plan.
- Does changing the thumbnail or the first frame reset it?
- It helps more than most people expect, because recognition happens in the opening frames. A new opening on the same body is the cheapest anti-fatigue move you have, and it is worth trying before you commission anything new.
- What if every creative decays at the same rate?
- Then your ceiling is audience size rather than creative quality, and the answer is a wider audience, a new offer or a new market. Producing more videos into a saturated pool is the most expensive way to learn this.
- Who should not bother measuring a decay curve?
- Anyone with fewer than a dozen creatives of history. One odd launch week bends the median. Run the volume first, measure second, and if nobody will film, book a creator and batch their take here.
Plot the curve, read the half-life, set the refill rate. Then meet it, with one take on Monday evening and the replacement live on Tuesday.