Creative Fatigue Is Now Called By A Model
Creative fatigue used to be a judgement a media planner made from a declining chart and a sense that people were bored of the ad. Increasingly it is called automatically. Ad platforms track when engagement on a creative is falling, flag it, and in many setups rotate in a fresh variant without anyone deciding to.
That is efficient. It also bakes in an assumption about what advertising is for.
What the model measures as creative fatigue
Response. Clicks, view-through, engagement, conversions, compared with how the same creative performed when it was new. When those numbers fall, the model concludes the audience has seen enough.
For performance advertising that is usually right. An offer people have already ignored three times is unlikely to work the fourth.
A falling click rate can mean an ad is tired. It can also mean it has finished doing its job.
Where the logic breaks
Brand advertising works partly through repetition. The lines people still remember were rarely the ones swapped out every fortnight; they were the ones that ran long enough to become familiar. A model optimising for fresh response will tend to retire an ad just as it starts building memory.
The result, across a category, is constant novelty and very little that anyone recalls.
What planners are doing
Separating the two jobs. Letting automated rotation run on the performance layer, where freshness pays, while protecting a core brand asset from the same logic and judging it on slower measures such as recall and branded search.
It requires someone to decide, deliberately, that a falling click rate is acceptable for one piece of work. That is an unusual thing to defend in a meeting, which is why the protection is rare.
The same discipline applies to testing. A brand that judges every new idea on its first fortnight of response will keep choosing work that wins quickly, and will rarely find out what a slower idea could have built.
More from The AI Shift.