Every media buyer has a frequency number in their head. Ours used to be 3. Tools set alerts at 3, and a much-shared 2026 post put the prospecting ceiling at 2.5, citing Meta's own data.
So we went looking for that data, and couldn't find it.
01 What is a good Facebook ad frequency?
There isn't a universal one. Meta's help page says frequency may average 1 to 2 per ad set or run much higher, depending on budget, audience size and schedule. Its advice is to watch whether performance drops as frequency rises. A good frequency is any level where your click-through rate and cost per result are holding steady.
Meta calculates frequency as impressions divided by reach. Impressions are counted exactly. Reach is estimated from sampled data, and it counts Meta Accounts rather than people, so someone with an Instagram profile that isn't linked to their Facebook profile counts twice. Reach also doesn't add up across ad sets: if one ad set reaches A and B and another reaches B and C, the campaign reached three accounts, not four.
The date range changes the number too. Impressions pile up every day, while reach only grows when Meta finds someone new, so the same ad set shows a higher frequency over 30 days than over 7. Take an ad set spending $100 a day at a $12 CPM, which buys about 8,333 impressions a day:
| Date range | Impressions | Accounts reached | Frequency |
|---|---|---|---|
| Last 7 days | 58,333 | 30,000 | 1.94 |
| Last 30 days | 250,000 | 70,000 | 3.57 |
Nothing about the ads changed between those two rows. Only the window did. A rule that says keep frequency under 3 means nothing until you know which window it was written for, and most never say. The reach figures are an illustrative example.
02 Where the safe frequency numbers come from
The common thresholds of 2.5, 3 and 3 to 4 show up in vendor and agency blogs without a disclosed method. The strongest-sounding one, a 2.5 ceiling attributed to Meta's own data, traces back through blogs citing blogs to no Meta source we could find. The famous 3 comes from 1970s research, where it was the minimum exposure count, not the maximum.
| Number | Where you'll see it | What's behind it |
|---|---|---|
| 2.5 for prospecting | A widely shared 2026 analytics-vendor post, repeated by other blogs | Attributed to Meta's own data. We found no Meta source, and the same vendor's October 2026 guide says there's no universal cutoff. |
| 3 | Tool alert defaults, glossaries, agency checklists | Usually traced to effective frequency research, where three exposures in a purchase cycle was the floor for peak response |
| 3 to 4 | Vendor glossary pages | No source given |
| 5 to 6 | 2020 coverage of a Facebook IQ study of 2,439 campaigns | Brand-lift surveys. Meta's original page is offline, and the write-ups disagree on whether results tapered off or declined. |
| 1 to 2 per week | Facebook IQ, 2016 | A randomized test across 11 large brand campaigns, measuring ad recall and purchase intent, not sales |
The 3 has an odd history. Herbert Krugman's 1972 paper argued three exposures may be enough, and Michael Naples' 1979 review for the Association of National Advertisers concluded that at least three exposures within a purchase cycle was optimal. Naples also found effectiveness kept building beyond three, at a slowing rate, with no evidence of a decline. Somewhere on the way to Ads Manager, a floor turned into a ceiling.
The most quoted performance chart is AdEspresso's 2018 analysis of 500 campaigns. Compared with a frequency of 1, click-through rate was 23% lower at a frequency of 4 and half as high at 9, while cost per click rose 68% and 161%. It used lifetime frequency and compared different campaigns rather than running a test, so read it as correlation.
03 What controlled research says about repetition
Controlled experiments find that wear-out varies wildly between campaigns. In 30 randomized experiments on Yahoo's front page, a third of campaigns showed no wear-out at all and 13% wore out sharply. Some held up after 30 to 50 impressions and others faded after two or three. With a spread like that, no single frequency number can be right for everyone.
That study, by economist Randall Lewis using 2.8 billion impressions from early 2010, found something more uncomfortable. Simply comparing people who saw an ad many times with people who saw it once overstated wear-out for most campaigns. People who see an ad many times tend to be heavier users. They behave differently before the first impression. Every frequency-versus-CTR chart you build from Ads Manager has the same bias.
The rest of the evidence mostly leans the same way:
- In a Yahoo experiment with a national retailer, ads lifted sales 3.6% against a control group, and doubling average impressions from 17 to 34 over two weeks nearly doubled the effect.
- A 2015 meta-analysis in the Journal of Advertising found brand attitude peaked at around 10 exposures in experimental settings, and recall kept climbing past the eighth.
- Meta's own 2016 test is the exception, and an informative one: with large brands and limited creative, ad recall gains slowed after one exposure a week, while purchase intent kept rising until about one and a half.
None of these studies ran on Meta performance campaigns, so they can't hand you a threshold. They do show that frequency on its own predicts very little.
04 What to watch instead of a frequency number
Watch the trend, not the level. Frequency is the warning light. The diagnosis is whether click-through rate falls and cost per result rises as it climbs, on the same ad and the same audience. Meta's own fatigue labels work this way: they compare an ad's cost per result with your past ads rather than looking at frequency.
In Ads Manager, Creative limited means cost per result is above your past ads but less than double. Creative fatigue means it has doubled or more. The labels only appear on ad sets with a single creative, and not on Advantage+ catalog ads, dynamic creative or Advantage+ app campaigns, so most accounts need their own check.
Ours: pull ad-level data for the last 30 to 60 days and look week by week at frequency, click-through rate and cost per result on the ads carrying most of the spend. Then read the pattern.
| Frequency | Click-through rate | Cost per result | What it means |
|---|---|---|---|
| Rising | Steady | Steady | Fine. The audience isn't tired yet, so keep spending. |
| Rising | Falling | Rising | Fatigue. Bring in different creative or widen the audience. |
| Rising | Steady | Rising | Costlier impressions. The auction got pricier, as it does in Q4, or Meta is reaching further into a used-up audience. |
| Steady | Falling | Rising | Not a frequency problem. Look at the offer, the landing page or a competitor. |
Retargeting needs one more step, because in a small pool high frequency is arithmetic, not a creative verdict. Impressions equal spend divided by CPM, times 1,000. Spend $1,600 a week at a $20 CPM on 20,000 recent site visitors and the best possible average is four impressions each, even if Meta reached every one of them. Set retargeting budgets from the pool size and the frequency where your own cost per result turned last time, not from a blog's number.
05 How to bring frequency down when it's hurting
Give the audience something genuinely new to look at, or give the ad more people to reach. On a small pool, cutting the budget does the same job. For fatigued ads Meta recommends a materially different ad, a broader audience or Advantage+ creative. Variations of one idea don't help, because Meta's system groups ads with similar visuals and themes and treats them as the same creative.
- New concepts beat new colors. A different hook, format or angle counts as new. A recolored background doesn't. Meta's creative diversity score tells you whether your ad set looks varied to its system.
- Have the replacement ready before the old ad dies. A creative testing system keeps new concepts coming so one tired ad isn't carrying the account. Meta also suggests keeping the original running while the new one picks up.
- Widen the audience or go broad, so the same budget spreads across more people.
- Cut the budget on small pools. Retargeting that can't absorb more spend just repeats itself at higher cost.
Frequency caps won't save a sales campaign. Meta's help center lists frequency cap and target frequency only for Awareness and Engagement campaigns using the Reach or ThruPlay goals, so on sales campaigns you manage frequency the slow way, through what you show and who you show it to.
06 Where to start
Open your top five ads by spend and set the date range to 7 days, then 30. Note frequency, click-through rate and cost per result for each week. Any ad where frequency has climbed for two weeks while CTR slid and cost per result rose is fatigued, whatever its frequency number.
If CPMs are the thing rising, our Black Friday CPM math shows how much your margin can absorb. If you'd like us to run the check on your account, send it over. Where did the frequency number in your head come from?
Sources: Meta Business Help Center pages on frequency, reach, impressions, data sampling, Meta Accounts, creative fatigue recommendations and reservation frequency controls (read October 21, 2026); Meta for Business, Demystifying Creative Diversification (December 16, 2025); Facebook IQ, Effective Frequency (July 21, 2016); Naples, Effective Frequency (Association of National Advertisers, 1979), citing Krugman (Journal of Advertising Research, 1972); Lewis, Worn-Out or Just Getting Started? (Yahoo front page experiments, 2010 data); Johnson, Lewis and Reiley, Add More Ads? (Yahoo Research, 2013); Schmidt and Eisend, Journal of Advertising 44(4), 2015; AdEspresso frequency analysis (August 27, 2018), vendor data from 500 campaigns; Social Media Today coverage of Facebook IQ's 2020 frequency study (October 27, 2020). The worked examples use illustrative numbers.
Want the full playbook?
The 12 Growth Leaks burning your ad budget — the same internal doc we hand to new clients on day one. One short form, no spam.
Get the playbook →FAQ
What is a good frequency for Facebook ads?
There is no universal number. Meta says frequency may average 1 to 2 per ad set or run much higher, depending on budget, audience size and schedule. A good frequency is any level where click-through rate and cost per result hold steady. Judge the trend in your own account, not a rule of thumb.
What frequency is too high for Facebook ads?
Frequency is too high when click-through rate falls and cost per result rises as frequency climbs on the same ad. Some campaigns tire after two or three impressions and others hold up past 30, so popular cutoffs like 2.5 or 3 have no published evidence behind them for sales campaigns.
How is Facebook ad frequency calculated?
Meta calculates frequency as impressions divided by reach. Impressions are counted exactly, while reach is an estimate of Meta Accounts based on sampled data. Because impressions accumulate faster than new reach, the same ad set shows a higher frequency over 30 days than over 7 days.
Can I set a frequency cap on Meta sales campaigns?
Meta's help center lists frequency cap and target frequency only for Awareness and Engagement campaigns using the Reach or ThruPlay performance goals. On sales campaigns, manage frequency through different creative, a wider audience, or a budget the audience can absorb.
What frequency should retargeting ads have?
In a small retargeting pool, frequency is mostly arithmetic: spend divided by CPM times 1,000, divided by the people in the pool. Spending $1,600 a week at a $20 CPM on 20,000 visitors means at least four impressions each. Set budget from the frequency where your own cost per result turned.