Free Ad Frequency & Fatigue Calculator.
Creative fatigue shows up as response falling while the cost of reaching people stays flat, not as a frequency number crossing three. Put two periods of the same ad set in and this compares results per thousand people reached against CPM across both. If response fell and CPM held, the creative has worn out. If response held and CPM rose, the auction got more expensive and the ad is fine. Those need opposite responses, and a frequency figure on its own cannot tell them apart.
Free, no sign up, no email. It runs in your browser, so nothing you type is sent to us or to anyone else.
Two periods of equal length, from the ad set’s own reporting. Its first good week against its most recent one is the usual pair.
| Metric | Earlier | Now | Change |
|---|---|---|---|
| Frequency | 1.60 | 2.40 | +50% |
| CPM | 10.00 | 10.00 | +0% |
| Results Per 1,000 Reached | 2.00 | 1.30 | -35% |
| Cost Per Result | 8.00 | 18.46 | +131% |
| Click Through Rate | 2.00 | 1.00 | -50% |
Response per thousand people reached has fallen while the cost of reaching them held steady. The same money is buying the same attention and getting less back, which is what creative fatigue looks like. New creative, not a higher bid.
The read compares results per thousand people reached against the cost of reaching them, because that pair separates a worn out creative from a dearer auction. Frequency is reported because you will be asked for it, not because it decides anything on its own. A ten percent threshold applies: smaller movements between two weeks are noise. Nothing you type here is sent anywhere.
How this works.
Why frequency on its own decides nothing
The rule everybody quotes is that frequency above three means the creative is spent. It is wrong often enough to be expensive in both directions. A retargeting pool of two thousand people can run at a frequency of eight and keep paying for itself, because the people in it are close to buying and being reminded is the point. A cold prospecting audience of four million can be finished at 1.4, because the people who were ever going to respond did so in the first pass.
Frequency tells you how often the ad was seen. It does not tell you whether being seen again is still working. Those are different questions, and only the second one decides whether you change anything.
The two things that separate fatigue from a dearer auction
Results per thousand people reached is the response measure worth watching, because it does not move when frequency moves. It asks what a thousand people did, not what a thousand impressions did. CPM is what it costs to reach those people, and it moves with competition rather than with your creative.
Put them together and the diagnosis falls out. Response down with CPM flat is a worn out creative: the same money is buying the same attention and getting less back. Response flat with CPM up is the auction: your ad is fine and somebody else started bidding. Both down at once happens, and the creative is the half you can fix this week.
Why cost per result is the wrong first number to look at
Cost per result is what everybody checks first, and it is the number most likely to send you at the wrong problem. It moves when the creative tires, when the auction gets busier, when the audience shifts, when tracking breaks, and when a seasonal buyer disappears. A rising CPA tells you something is wrong and nothing about what.
It is still reported here, because it is what the conversation with a client is about. It is just not what the read is built on.
What to compare against what
Two periods of equal length from the same ad set. The ad set's first good week against its most recent one is the usual pair, and it is a fair comparison as long as nothing else changed: same audience, same optimisation event, same placements, no budget jump in between.
If the budget doubled between the two periods, you are not measuring fatigue, you are measuring what happens when delivery is pushed harder into the same audience. That is worth knowing too, but it is a different question and this will read it as fatigue.
What it will not tell you
Whether the tracking is right. A Meta pixel that started deduplicating differently, or a conversions API that dropped out, produces exactly the pattern of a tired creative and has nothing to do with the ad. Check that the results you are comparing were counted the same way in both periods before you trust anything here.
It also will not tell you what to make next. It tells you whether the thing you have is finished, which is the decision people put off because the evidence for it is usually a feeling.
Questions we get asked.
What is a good frequency for Meta ads?
- There is no single number, which is the point of this tool. A retargeting audience can sustain a frequency of eight or more because the people in it are close to buying. A cold prospecting audience can be exhausted below two. What matters is whether response per thousand people reached is still holding, not what the frequency figure says.
How do I know if my Meta creative has creative fatigue?
- Compare two periods of the same ad set. If results per thousand people reached have fallen while CPM has stayed flat, the creative has worn out: you are paying the same to reach people and getting less back. If CPM rose and response held, the auction got more expensive and the creative is still working.
My cost per result went up. Is that creative fatigue?
- Not necessarily, and cost per result cannot tell you. It rises when the creative tires, when the auction gets more competitive, when the audience shifts, when tracking breaks, and when seasonal demand falls. Splitting response from cost of reach is what separates those.
Should I just raise the budget instead of changing the creative?
- If the read says the auction got dearer and response is holding, more budget is a reasonable answer. If the creative has worn out, more budget buys more impressions of something that has stopped working, which makes the cost per result worse rather than better.
How long should each period be?
- Long enough to carry a useful number of results, and equal to each other. A week against a week is the common pair. Anything shorter than a few days on a low volume ad set is mostly noise, and the ten percent threshold here exists so that noise does not read as a finding.
Is anything I type here stored or sent anywhere?
- No. It all runs in your browser. There is no server behind this tool, no sign up and no email gate.
Ready to move the numbers?
Let us talk about your current goals, what is working, what is not, and where the biggest opportunities may be.