A good frequency for Facebook ads is not a single number. Frequency is impressions divided by reach, and it is driven mostly by budget, CPM and audience size, so the same value can be healthy in one ad set and a warning sign in another. Read it together with reach growth and click-through rate before you act on it.
Most pages on this topic give a number anyway: 1 to 2, under 3, or "Facebook says 3.4". This page shows where those numbers come from, the formula that predicts frequency before you launch, and a table for when a high frequency is normal.
How Meta calculates frequency
Meta defines frequency as the average number of times each account saw your ad, calculated as impressions divided by reach over the date range you select. Meta also marks it as an estimated metric, calculated from sampled data.
Frequency = Impressions ÷ Reach
Two things follow. Change the date range and the number changes. And because it is an average, it hides its own distribution: a frequency of 3 can mean everyone saw the ad three times, or most people saw it once and a small group saw it a dozen times. Those two situations need opposite responses.
Meta's help page says frequency "may average 1 to 2 per ad set or may be much higher, depending on your budget, audience size and schedule." That is a description of what usually happens, not a target.
Where the common frequency numbers come from
| Number you will see quoted | Where it actually comes from |
|---|---|
| "1 to 2 per week" | Meta's Effective Frequency study: 11 large brand campaigns, January 2015 to February 2016, all reach-optimised. Ad recall lift slowed substantially after a cap of 1 per week; purchase intent lift slowed closer to 1.5 per week. A cap of 1 per week captured up to 80% of the potential ad recall lift, and 2 per week captured 95% of the potential purchase intent lift. These are brand-lift results, not conversion campaigns. |
| "One or two times a week" | Meta's reservation buying (formerly reach and frequency) describes limiting brand campaigns to one or two exposures a week. A planning option for awareness campaigns, not a performance finding. |
| "3.4 is Facebook's tipping point" | Repeated on agency and tool blogs and attributed to "Facebook". It does not appear in the published Effective Frequency study, and we could not find a Meta source for it. Treat it as unsourced. |
| "Keep it under 3" | A practitioner rule of thumb, usually for cold prospecting. Useful as a prompt to look, not as a rule on its own. |
The only number here from a published Meta study is about ad recall and purchase intent in awareness campaigns, measured per week. It says little about a sales campaign optimising for purchases. The same study is explicit that there is no single "magic frequency level" for every brand and campaign.
The formula that predicts frequency before you launch
Frequency is mostly arithmetic. Your budget and CPM set how many impressions you buy; your audience size limits how many people those impressions can spread across.
Impressions = Budget ÷ CPM × 1,000
Frequency = Impressions ÷ people reached
Worked example, using invented inputs to show the arithmetic, not benchmarks:
| Ad set | Budget, 7 days | CPM | Impressions | Reachable people | Lowest possible frequency |
|---|---|---|---|---|---|
| Retargeting pool | $700 | $14 | 50,000 | 9,000 | 5.6 |
| Broad prospecting | $700 | $10 | 70,000 | Millions | Set by delivery, often near 1 to 2 |
The retargeting ad set cannot report a frequency below 5.6 in that week, however good the creative is. There are only 9,000 people to show 50,000 impressions to. A high number there is the budget talking, not a creative problem. If the audience is too small for the budget, the fix is a bigger audience or a smaller budget.
When high frequency is normal: a context table
Three ad sets, each showing frequency 4.0 over 14 days:
| Ad set | Why 4.0 happens | Is 4.0 a problem? |
|---|---|---|
| Cold prospecting, broad, millions reachable | Delivery is concentrating on a small part of the audience | Worth a look. Broad audiences rarely reach 4.0 in two weeks unless delivery is narrowing. |
| Retargeting, 9,000 people | Arithmetic: a small pool and a fixed budget | Probably not. Check whether click-through rate is falling before you act. |
| New brand or new message, short campaign | Planned: more exposure is the strategy | Often intentional. Meta's planning guidance puts new brands, short campaigns and complex messages on the higher-frequency side. |
Same number, three meanings. Meta's own planning factors point the same way: new brands, low market share, short purchase cycles, complex messages and high season call for more frequency; established brands with long purchase cycles and simple messages need less. That is why a single published threshold does more harm than good.
Repetition or something else? Check these first
Rising frequency and falling results happen together constantly without one causing the other. Rule out the cheaper explanations first.
| Check | What it looks like | Cheaper than new creative? |
|---|---|---|
| Auction costs moved | CPM up, click-through rate flat, conversion rate flat | Yes |
| Seasonal competition | CPM up in a known peak, see the Q4 CPM curve | Yes |
| Attribution definitions changed | Reported conversions down, orders flat | Yes |
| Landing page or offer changed | Click-through rate flat, conversion rate down | Yes |
| Tracking broke | Conversions down across all campaigns at once | Yes |
| Audience genuinely saturated | Reach growth flat, frequency rising, click-through rate falling, CPM flat | This is the fatigue case |
Only the last row is fatigue. Its signature is falling click-through rate with flat CPM and flat reach growth. Rising CPM with a flat click-through rate is an auction cost problem, and new creative will not fix it.
The third row matters more in 2026. Click-through attribution now requires a link click, so reported conversions can drop with no change in real sales. See Meta's 2026 attribution changes. Replacing working creative because a reporting definition changed is a common and expensive mistake. For how often an ad labelled fatigued never had enough delivery to wear out, see creative fatigue or creative failure.
Read reach growth, not just frequency
Because frequency is an average, the number you want is not in the frequency column. Two ways to get closer:
Compare several windows. Read 7-day, 14-day and 30-day frequency together. If the 7-day figure is close to the 30-day figure, exposure is concentrated in a recent burst. If the 30-day figure is much higher, exposure is spread out.
Watch whether reach is still growing. If reach has stopped growing while impressions continue, you are showing the ad to the same people again, whatever the frequency figure says. A reach plateau usually shows up before the click-through rate falls.
What to do when it really is saturation
In order of cost:
1. Widen the audience before rebuilding the creative. Usually cheaper than a new production cycle, and it often removes the symptom. 2. Bring back paused assets that have been off long enough to feel new. 3. Change the angle, not the execution. A new colour grade on the same claim is a new file, not a new idea. See creative diversification. 4. Produce new creative last, because it is the slowest and most expensive response. The creative refresh playbook covers the order.
Frequency caps: when they help
Frequency controls are built for reach-focused buying. In reservation campaigns you can set a target weekly frequency, now the default control, or a maximum frequency cap, for awareness and engagement goals.
Use a cap when you have a brand or compliance reason to limit exposure. Do not use one as a performance fix: if the audience is too small for the budget, a cap does not change that. It just means the budget under-delivers instead of repeating.
Check one ad in a minute
The creative fatigue checker is free and scores one ad from its CTR decline, days active and frequency, with the formula published on the page. It treats frequency as a review signal, not proof. For a full account view, a Meta ads audit reads a connected Meta ad account and flags where delivery, cost and results do not add up. AdRiseLab is Meta-only and plans start at $39 a month; see pricing.
Sources
Facebook IQ: Effective Frequency, reaching full campaign potential. Meta Business Help Center: Frequency. Meta Business Help Center: about reservation. Meta Business Help Center: choosing your frequency controls for a reservation campaign.
Related Reading
Creative fatigue explained covers the wider set of signals, and the creative fatigue guide walks through reading them in Ads Manager.