In shortFatigue is an audience problem wearing a creative problem’s clothes — frequency above about three means you have run out of people, not out of ideas, and swapping the creative buys time without fixing the reason the same faces keep coming back.
Frequency is the most misread number in Ads Manager, because it looks like a property of the ad and is actually a property of the audience. It is impressions divided by reach: how many times, on average, one person saw this. It climbs for exactly one reason — you are spending more money against the same number of people.
What the average hides
An average of three does not mean everybody saw it three times. Ad delivery is not egalitarian: it favours the people most likely to respond, which means a small group sees an ad many times over while most of the audience sees it once or twice. A frequency of three routinely contains people on their tenth impression.
That matters because those are the people whose behaviour changes first. They stopped reading it around the fourth time. The average is still drifting up gently while the part of the audience that was actually going to convert has already tuned out.
Frequency does not measure how tired your creative is. It measures how small your audience has become.
Why the line sits at three
Three is not a law of nature and nobody should defend it as one. It is the point at which the distribution behind the average reliably contains people who have seen the ad enough times to have made up their minds. Below it, the heavy tail is small enough not to matter. Above it, the tail is where your budget is going.
This is why the confidence attached to the finding rises with the excess rather than sitting flat. A frequency of 3.1 is barely distinguishable from 2.9 and the card says so. A frequency of five is a different claim, and by then the rule is nearly certain.
Replacing the creative is not free
Every new ad re-enters the learning phase, and the learning phase spends money inefficiently on purpose while delivery works out who to show it to. Swap a creative that was not actually fatigued and you have paid that cost for nothing — which is why this finding is worth confirming against frequency rather than acting on a falling click rate alone.
It is also why the proposal names a specific file when it can. A creative that has already been analysed and is not running anywhere else turns a vague instruction into a decision you can approve in one press, and nothing launches until you do.
What actually fixes it
If frequency is high because the audience is genuinely small, a new creative resets the counter and buys perhaps two weeks. The same people will see the new ad too often as well, because the constraint was never the creative.
- Widen the audience, or add a lookalike, so the budget has somewhere new to go.
- Cap frequency at the ad set, which trades some delivery for a longer useful life.
- Lower the budget to match the audience you actually have, rather than pushing spend into people who already decided.
- Rotate the creative — worth doing, but as relief rather than a cure.
When Admetriq raises it
The actual rule, so you can hold this page against what you see rather than take its word for it.
- Fires when average frequency reaches 3.0 or above (FREQ_FATIGUE = 3.0). Frequency is impressions divided by reach — the average number of times one person saw this ad in the window.
- Confidence starts at 0.55 and climbs with the excess, 0.15 per point of frequency above three, capped at 0.90. A frequency of 3.1 is a guess; 5.0 is not.
- A further 0.05 is added once the ad has at least 5,000 impressions, because frequency computed over a few hundred people is arithmetic rather than evidence.
- The proposal is a creative refresh. If the Library holds an already-analysed creative that no active campaign is using, that exact file is named in the proposal and launches on approval — when nothing qualifies, the proposal says to rotate something in rather than inventing a candidate.
Most often mistaken for
Low click-through ratelow_ctr
Read frequency before you blame the creative. Fatigue is frequency climbing past three while the click rate falls behind it. A low click rate at a frequency near one is the opposite finding — the ad never worked, on people seeing it for the first time — and replacing it with another in the same vein repeats the mistake at the cost of a fresh learning phase.