How to use Ad Frequency and CPM
Frequency caps and audience saturation quietly inflate your CPMs long before ROAS falls off a cliff. Here's how to read frequency data, set the right caps, and time creative refreshes to keep paid acquisition profitable.
Ad Frequency and CPM
The relationship between how often an ad is shown to the same user and the cost to reach that user — as frequency climbs, CPMs inflate.
Ad frequency is the average number of times a unique user has seen your ad over a given window. CPM (cost per thousand impressions) is what auction platforms like Meta and TikTok charge you to deliver those impressions. The two are mechanically linked: as you push more impressions into a finite audience, the auction has to bid harder to reach the people most likely to convert, and quality signals like CTR and engagement decay from fatigue.
The result is a CPM curve that bends sharply upward past a certain frequency threshold — usually around 2.5-3.5 on prospecting audiences and 5-7 on retargeting. Watching frequency lets you predict the CPM spike before it shows up in your ROAS report.
If you manage paid media for an online store, you've seen the pattern: a winning ad set runs flat for three weeks, then CPM jumps 40% in five days and ROAS halves. The trigger isn't usually a platform change. It's frequency.
Frequency is the most leading indicator you have for CPM. CPM is a lagging metric — by the time it spikes in your dashboard, you've already burned a week of budget at degraded efficiency. Frequency tells you the spike is coming 7-10 days in advance, which is enough lead time to ship new creative or rotate audiences.
Why CPM rises with frequency
Three mechanisms drive the inflation, and they compound. The first is auction dynamics. Meta and TikTok auctions optimise for the user most likely to take your conversion event. Once that user has seen the ad two or three times without converting, the algorithm has to widen the bid range to keep delivery flowing.
The second is engagement decay. CTR and thumbstop rate drop roughly 30-50% between frequency 1 and frequency 4 on prospecting. Lower engagement signals lower relevance to the platform, and the relevance penalty translates directly into higher CPM — you're paying for the algorithm's reduced confidence.
The third is audience exhaustion. Your active audience — the people who fit the targeting and are reachable that week — is finite. A 2M-person lookalike sounds large, but daily active reachable users might be 150-300k. Spend €500/day at a €15 CPM and you'll touch every reachable user within two weeks.
Frequency leads CPM by 7-10 days
On most Meta prospecting campaigns, frequency crosses 2.5 about a week before CPM visibly inflates in reporting. If you only watch CPM, you're already late. Add a frequency column to your daily review and you buy back that week.
Reading the frequency-CPM curve
The curve isn't linear. CPM stays roughly flat through frequency 1.5-2.0, drifts up modestly through 2.0-2.8, then inflects sharply. On a typical apparel prospecting campaign, going from frequency 2.5 to 3.5 adds 35-55% to CPM. Going from 3.5 to 4.5 adds another 40-60% on top.
The inflection point is creative- and audience-specific. A high-production video on a broad interest audience holds up longer than a static carousel on a 1% lookalike. But the shape — flat, then bend, then cliff — is universal. Your job is to know where your inflection sits and refresh creative before you reach it.
CPM inflation by ad frequency (Meta prospecting, apparel benchmark)
Prospecting (broad / lookalike)
Retargeting (site visitors, 30d)
Retargeting tolerates higher frequency because the user already has intent — they signed up for the curve when they visited your product page. Prospecting punishes frequency more because you're paying to convince strangers, and strangers tune out faster.
Frequency thresholds by audience and vertical
There's no universal frequency cap. The right ceiling depends on audience size, purchase cycle, and how visually distinctive your creative is. A beauty SKU with a clear before/after hook can run hotter than a fashion brand selling on aesthetics, because the message lands on first or second view.
Use the table below as starting thresholds — the point where you should already have new creative in the queue, not the point at which you've lost. The right move is to refresh at the warning line, not wait for the cliff.
Frequency thresholds for refresh — Meta Ads, 7-day window
| Audience × Vertical | Refresh trigger | Hard cap (cliff) | Typical CPM impact at cap |
|---|---|---|---|
| Prospecting — Apparel / fashion | 2.6 | 3.5 | +55-80% |
| Prospecting — Beauty / skincare | 3.0 | 4.2 | +45-70% |
| Prospecting — Home & lifestyle | 2.4 | 3.2 | +60-85% |
| Prospecting — Electronics / accessories | 2.2 | 3.0 | +65-90% |
| Retargeting — 30d site visitors | 5.0 | 7.0 | +40-55% |
| Retargeting — ATC / checkout abandoners | 6.5 | 9.0 | +30-45% |
Audience size matters too. On a sub-200k retargeting pool, frequency climbs daily and you should plan a creative rotation every 10-14 days. On a 5M+ broad prospecting audience, the same creative can run 4-6 weeks before frequency forces a refresh.
Acting on frequency data before CPM spikes
The operational rhythm is straightforward. Review frequency daily on every active ad set. When 7-day frequency hits the refresh trigger from the table, you should already have new creative variants in review. When it hits the hard cap, the old creative pauses — not because it's lost, but because every additional impression now costs 1.5-2x what it did at frequency 2.
Frequency caps inside Meta Ads Manager are a blunt secondary tool. Setting a campaign-level cap of 3 per week prevents the worst over-delivery but doesn't replace creative refresh — capped delivery just shifts spend to the next ad in the set. Pair frequency caps with a creative testing pipeline that ships 3-5 new variants every two weeks.
The refresh-or-rotate decision
If CPM is rising but CTR is holding, rotate audiences — your creative still works, the pool is just tired. If CTR is dropping alongside CPM rising, refresh creative — the message is fatigued. Diagnosing which one is happening before you act is the difference between a 20% efficiency gain and burning two weeks shipping the wrong fix.
Ad frequency and CPM — common questions
For prospecting, keep 7-day frequency below 2.5-3.0 depending on vertical. For retargeting, 5-7 is normal and tolerable. Above those thresholds, expect CPM inflation of 40% or more within 5-10 days.
Frequency caps are useful as a guardrail (cap at 3-4 per week on prospecting reach campaigns) but they don't solve audience saturation. Capping just shifts spend to other ads in the set. The real fix is a creative refresh cadence — see our guide on creative testing to lower CPM for the pipeline.
Three other causes are common: auction competition increased in your geo (check seasonality), your relevance score dropped due to CTR decay, or you expanded targeting and the new pool is more expensive to reach. Check CTR trend and audience overlap before assuming frequency is the culprit.
Tie the cadence to frequency, not the calendar. On a 500k-1M prospecting audience spending €300-500/day, that usually means new variants every 10-14 days. On larger broad audiences, 3-5 weeks. Always have the next batch in testing before the current creative hits its refresh trigger.
TikTok has a steeper fatigue curve — frequency above 2.0 on prospecting often inflates CPM faster than Meta because the format rewards novelty so heavily. Expect to refresh TikTok creative 30-40% more often than Meta for the same audience size.
Sometimes, briefly. Widening dilutes frequency mathematically but the new audience usually has lower conversion intent, so ROAS may not improve even if CPM dips. Audience expansion is a delay tactic, not a fix. Refresh creative in parallel.
Use the ratio of weekly reach to audience size. If you're reaching 60%+ of an audience in a 7-day window, you're saturating and frequency will climb fast. Combine with CTR trend — a 25%+ CTR drop over two weeks at stable spend is a saturation signal even if frequency looks moderate.
In Meta, frequency caps set in Reach objective campaigns apply at the campaign level across all ads. For conversion-objective campaigns, you can't set a hard cap — you control frequency through creative rotation and audience management. This is why a healthy creative pipeline matters more than the cap setting.
Frequency leads CPM by 7-10 days; CPM leads ROAS by 3-5 days. So frequency is your earliest warning that ROAS is about to drop. A daily frequency check costs nothing and gives you almost two weeks of lead time to defend efficiency.
No. On retargeting, frequency 5-7 is often where you see the best ROAS because intent users need multiple touches to convert. The problem is unmanaged frequency on prospecting, where the curve bends sharply and you're paying premium CPM to reach people who've already decided not to buy.
Track CAC, channels, and funnel conversion in one place
Metricuno connects ad spend, funnel events, and revenue so you can see CAC by channel, cohort, and campaign — without stitching together five tools.