Ad frequency and fatigue: when good creative stops working

Aug 22
3 min read

Updated: Aug 29

Introduction

Frequency is how many times the average person has seen your advert. Fatigue is what happens when that number climbs and performance falls.

It is the most predictable form of decay in paid media, and the most commonly misdiagnosed — usually as a targeting problem, occasionally as a platform conspiracy.

1. Ad frequency and fatigue show a specific signature

The pattern is distinctive if you know what to look for.

Click-through rate falls while conversion rate holds. People who still click still buy — there are simply fewer of them clicking. That is fatigue.

Compare it with the alternatives: if conversion falls while clicks hold, the problem is the landing page or the offer. If cost rises while both hold, the market changed. Only the first is fatigue, and only the first is fixed by new creative.

2. Why raising the bid makes it worse

The instinct when performance dips is to bid more. With fatigue that is precisely wrong.

A higher bid buys more impressions of an advert people are already ignoring, which raises frequency faster and accelerates the decay. You pay more per result and shorten the advert's remaining life.

Spend that money on a new opening instead.

3. Audience size sets the clock

Frequency climbs at a rate set by budget divided by audience size. A small audience with a steady budget reaches high frequency quickly.

This is why retargeting audiences fatigue soonest and broad prospecting audiences last longest. If an advert is wearing out in a week, the audience is probably too narrow rather than the creative too weak.

Broadening the audience often buys more life than changing the advert.

4. Refresh the opening, not the whole advert

Most of the fatigue effect sits in the first few seconds, because that is what determines whether anyone watches the rest.

So the cheapest refresh is a new opening on the same body. Built modularly — several hooks against one reusable middle section — a refresh costs a fraction of a new production, which is what makes staying ahead of fatigue affordable.

Replacing the entire advert every time is why so many advertisers cannot keep up.

5. Rotate rather than replace

Keep three or four variations live and let the platform distribute between them, rather than running one until it dies.

Rotation slows the rise in frequency per creative and gives you continuous comparison data. It also means there is always something to fall back on when one variation tires, instead of a scramble to produce a replacement.

6. Do not reset the learning unnecessarily

Significant edits to a running campaign restart the platform's optimisation, and performance dips while it relearns. That dip is easily mistaken for further fatigue.

Adding a new creative to an existing ad set is usually gentler than rebuilding the campaign. Change one thing, wait, and read the result before changing another.

7. Accept that some decay is normal

Not every decline needs fixing. An advert that ran well for two months and is now tired has done its job.

The mistake is treating decay as failure and rebuilding everything. The system that works is a steady supply of new openings and a rotation habit — not a search for the advert that never wears out, which does not exist.

8. Track frequency, do not infer it

Most platforms report frequency directly. Look at it rather than guessing from performance.

Watching the number climb tells you a refresh is due *before* results fall, which is the whole point — reacting after the decline means you already paid for a stretch of ignored impressions. Note the frequency at which your own adverts historically start to fade; it differs by business and audience, and once you know yours you can schedule refreshes rather than firefight them.

Conclusion

Diagnose from the pattern: falling clicks with steady conversion is fatigue, and it is the only one of the four patterns that new creative fixes.

Do not raise the bid, because that accelerates it. Widen a too-narrow audience, refresh the opening rather than the whole advert, keep several variations rotating, avoid resetting the learning, and treat decay as a supply problem rather than a failure.