Your Meta CPM is climbing. Here's why — and whether your offer survives Q4.
By Denzel Geng · Updated October 13, 2026
Facts on this page verified October 13, 2026 against public sources.
Most "high CPM" guides list causes from memory. This one uses Meta's own help pages for every cause, real US CPM data for every Q4 since 2023, and a five-minute check to see whether your Black Friday math still works when the auction gets expensive.
Your Meta CPM is high either because more advertisers are bidding for the same people, or because Meta's system has fewer cheap ways to deliver your ad: a narrow audience, few placements, a learning-phase reset, weak conversion signals, or creative people have stopped engaging with. Q4 is the predictable version of the first cause, and it is concentrated: on Black Friday 2025, Meta's own data put CPMs 48–61% above an October 1 baseline, depending on format, while US monthly CPMs on Birch's chart for November and December 2025 were only 2–3% above September.
See what's winning in your niche- What counts as a "high" CPM on Meta?
- Why is my Meta CPM so high? The 6 causes
- How much do Meta CPMs rise in Q4?
- What do other Q4 CPM datasets show?
- How do you check if your offer still works at Q4 CPMs?
- What should you do if the math only works at October prices?
- When does Q4 auction competition start?
- What can you control when CPM rises?
- When is a high CPM not your problem?
- Stocking up on winners in ViralMojo
- Sources
What counts as a "high" CPM on Meta?
A high CPM is one your margin can't carry, not one above an industry average. Meta defines CPM as "the average cost for 1,000 impressions": spend divided by impressions, times 1,000. Averages move too much to be a target. On Birch's public chart of US Facebook ad costs, September CPM was $11.94 in 2024, $16.02 in 2025 and $22.31 in 2026, so a CPM that looked expensive two Septembers ago is cheap today. Compare your CPM with your own trailing 4–8 weeks and with the CPA your margin allows. A $25 CPM can be fine if your ads convert, and an $8 CPM can lose money if they don't.
If your campaigns optimise for purchases, CPM is an input, not the score. Meta notes that for conversion-optimised ad sets, CPM may not be a good indicator of performance. Watch cost per result first, and use CPM to explain why it moved.
Why is my Meta CPM so high? The 6 causes
Meta's help page on high costs names six things to check. Auction competition is one of them, and it's the one Q4 and every sale event push up. The other five are inside your account, so check them first when only one campaign got expensive.
| Cause | What Meta says | How to check | Fix |
|---|---|---|---|
| 1 · Auction competition (Q4, sale events) | "If your costs are still too high, you might be facing increased auction competition." | Did CPM rise across every campaign on the same dates? Check the Auction Competition Change diagnostic: Meta flags a change as significant "if it is over 20%." | You can't outbid the market. Do the Q4 math and make the offer survive it. |
| 2 · Creative people have stopped engaging with | "We subsidize relevant ads in the ad auction, so more relevant ads often cost less and see more results." | Frequency rising, CTR and hold rate falling, or a "Creative fatigue" status (cost per result at least twice your past ads). | New concepts, not new crops of the same ad. See creative diversity and ad fatigue. |
| 3 · Audience too narrow | "By reducing your audience size, you might prevent the delivery system from pursuing lower cost opportunities outside of your targeting criteria." Meta's pricing page: ad sets with at least 2 million people "often see better performance." | Audience size under ~2M, or stacked interests and exclusions. | Go broader, or switch to Advantage+ audience, so delivery can reach cheaper people. |
| 4 · Too few placements | "By including fewer placements, you can prevent the delivery system from pursuing lower cost opportunities." | Placements manually limited to Feed or Stories only. | Turn on Advantage+ placements, with a 9:16 version of each ad. |
| 5 · Stuck in learning | Learning ends "after about 50 results in the week after the ad set's last significant edit"; until then ad sets "usually have a higher CPA." | "Learning" or "Learning limited" status; recent budget, bid or creative edits. | Fewer, bigger ad sets; batch your edits; stop resetting winners. |
| 6 · Weak event signals | "Issues with your data source setup can lead to a lower volume of attributed events and an increased cost per action." | Events Manager: pixel/CAPI match quality, deduplication, missing purchase events. | Fix tracking before you touch the ads. |
All quotes are from Meta Business Help Center pages, accessed October 13, 2026; each one links to its source.
How much do Meta CPMs rise in Q4?
Less than most guides claim across the quarter, and more than most brands plan for in Black Friday week. CPM isn't high for all of Q4. It spikes in the days around Black Friday, and the spike changes a lot from year to year.
| US Facebook CPM (Birch) | Sep | Oct | Nov | Dec | Black Friday week |
|---|---|---|---|---|---|
| 2023 | $9.29 | $9.39 (+1%) | $11.59 (+25%) | $10.61 (+14%) | $13.79 (+49%) |
| 2024 | $11.94 | $13.39 (+12%) | $15.12 (+27%) | $14.53 (+22%) | $19.23 (+61%) |
| 2025 | $16.02 | $15.93 (−1%) | $16.37 (+2%) | $16.44 (+3%) | $18.40 (+15%) |
Birch figures are from Birch's public "Facebook advertising costs" chart (US campaigns, all objectives), read October 13, 2026. Percentages are versus the same year's September. Birch doesn't publish how many advertisers the chart covers.
| Year | Sep | Oct | Nov | Dec | Black Friday week |
|---|---|---|---|---|---|
| 2023 | $9.29 | $9.39 | $11.59 | $10.61 | $13.79 |
| 2024 | $11.94 | $13.39 | $15.12 | $14.53 | $19.23 |
| 2025 | $16.02 | $15.93 | $16.37 | $16.44 | $18.40 |
What do other Q4 CPM datasets show?
The same pattern: a sharp Black Friday spike, much less across the quarter. Meta's own 2025 holiday benchmarks put Black Friday CPM inflation, "when auction competition peaks," at 48% against an October 1 baseline for purchase-optimised campaigns using 9:16 vertical video with sound, versus 61% for other formats, globally. Biddyco's US accounts (about $6.9M a month in spend, accounts spending at least $30k in October) saw a median Black Friday CPM 57% above each account's own October average. For its consumer brands, BFCM week was up 35% and December 29 to January 4 was 24% below October.
So budget for a range, not one multiplier. In 2025 the month barely moved, but Black Friday week still rose 15–61% depending on the dataset. If you have last year's numbers from your own account, use those.
How do you check if your offer still works at Q4 CPMs?
CPM rises every Q4, and your margin doesn't. Run this check now, not on Black Friday morning. It takes five minutes and one spreadsheet row.
1. Find your break-even CPA
Average order value × gross margin (after product, shipping and payment fees). A $60 order at 60% margin leaves $36 to pay for the customer. That's the most you can spend on a first purchase without losing money on it.
2. Take your current cost per purchase
Use the last 2–4 weeks from your main prospecting campaign, not a lucky retargeting week. Say it's $28.
3. Raise it by the CPM increase you saw last Q4
If click-through and conversion rates hold, CPA moves roughly with CPM. At +15% (Birch, BFCM week 2025) it's $32. At +35% (Biddyco consumer brands, BFCM week) it's $38. At +61% (Meta, Black Friday, formats other than 9:16 video with sound) it's $45.
4. Compare with your break-even
In this example the offer works at a mild spike and loses money at a hard one. Conversion rates often rise during sale week, which helps, but don't plan on it covering a 60% CPM jump.
5. Decide now
If the math only works at October prices, use one of the three fixes below while the auction is still cheap.
| Scenario | Cost per purchase | Clears $36 break-even (60 AOV)? |
|---|---|---|
| Now | $28 | Yes |
| +15% (Birch, BFCM week) | $32 | Yes |
| +35% (Biddyco, consumer brands) | $38 | No |
| +61% (Meta, Black Friday, other formats) | $45 | No — but clears $48 break-even at $80 AOV |
What should you do if the math only works at October prices?
Fix the economics before the auction gets expensive. All three of these take weeks to set up, so they're October work.
Raise order value with bundles
A bundle lifts the order value behind every click, so each customer pays more of their own acquisition cost. In the example, raising AOV from $60 to $80 moves break-even CPA from $36 to $48, enough to absorb even the $45 Black Friday CPA.
Use an offer you can run all year
Not a discount that wipes out your margin. In ViralMojo's BFCM 2025 dataset, the 9 campaigns with a verifiable percentage in their ad copy peaked at 23% to 70% off, with a median of 48%. A 48% discount on a 60% margin product leaves almost nothing to pay for the click. A gift with purchase, a bundle price or a threshold offer can protect margin and still read as a sale. See all 9 campaigns and their discounts.
Stock up on proven ads
Don't test in the most expensive weeks. Every test in BFCM week buys its learning at the year's highest CPM. Find your winners in October, then make several genuinely different versions of them (new hooks and formats on a proven angle) so you're not relaunching one tired ad. Meta's own number points the same way: 9:16 video with sound saw 48% Black Friday inflation versus 61% for other formats.
When does Q4 auction competition start?
Earlier than Black Friday. In ViralMojo's BFCM 2025 dataset, among the 20 verified Black Friday campaigns with clear start and end dates, the earliest went live on October 29, six (30%) were already running by October 31, and the median start was November 6, three weeks before the sale. Sixteen of the 20 kept running past Cyber Monday. If your test budget is still finding winners in November, it's competing with every brand's sale creative.
Three campaigns from that dataset show the range. Beautiful Earth Boutique went live on October 29, 2025, a month before Black Friday (ad). Dylan Oaks ran its gift-with-purchase campaign from November 6 to December 27, a 51-day flight (ad). The Ordinary kept one flat 23% discount, its "Slowvember", from November 1 to December 14 and never raised it for Black Friday (ad). The Ad Library doesn't show what any of these earned. It does show their creative was live weeks before the most expensive days.
Testing gets cheaper again after the holidays. On Birch's chart, December CPM stayed close to November in 2025, and Biddyco's consumer-brand accounts were 24% below October in the first week of January. That makes early January the cheap window to test again. Every flight, start to finish, is in our BFCM 2025 timing data.
What can you control when CPM rises?
Mostly the creative. You can't lower the market's bids, but you can change how relevant your ad is, and relevance is part of the auction. Meta's ad auction page says the winner is the ad with the highest total value, built from the bid, the estimated action rate and ad quality, and that "an ad that's more relevant to a person could win an auction against ads with higher bids." Meta also says it subsidises relevant ads, "so more relevant ads often cost less and see more results."
Variety counts as much as quality. Meta's creative diversification guidance says that when ads "look or feel alike, these are seen as variations of the same creative." Ten edits of one concept count as one creative and wear out together. Going into Q4 with three or four genuinely different angles is a cost decision as well as a creative one. The creative diversity playbook covers how many concepts to run and how to tell when they're actually different.
When is a high CPM not your problem?
When the whole market moved. If CPM rose on the same dates across every campaign, and Meta's Auction Competition Change diagnostic shows a jump over 20%, the auction got more expensive and nothing in your account broke. Hold your proven ads, don't reset learning with panic edits, and judge the week on cost per purchase rather than CPM. If CPM rose in one campaign while others held, the cause is in that campaign: audience, placements, learning or creative. Our Facebook ads not converting guide walks through telling the two apart.
Sources
Meta Business Help Center (accessed October 13, 2026): About ad auctions · CPM · Your costs are too high · Learning phase · Creative fatigue · Auction Competition Change · Performance fluctuations.
Meta for Business: Ads pricing · Demystifying Creative Diversification (December 16, 2025) · 2026 Holiday Insights Center: ads benchmarks (data October 1, 2025 to January 15, 2026).
Third-party CPM data: Birch, Facebook advertising costs (US, updated weekly; read October 13, 2026) · Biddyco, Meta ads Black Friday CPM (October 5, 2026). ViralMojo: BFCM 2025 Meta ads study.
Common questions.
There isn't a universal good CPM, because averages move every year. On Birch's chart of US Facebook ads, September CPM went from $11.94 in 2024 to $16.02 in 2025 to $22.31 in 2026. A good CPM is one where your cost per purchase stays under your break-even CPA (order value × gross margin). Compare against your own last 4–8 weeks, not an industry figure.
Check whether it rose everywhere or in one place. A jump across every campaign on the same dates is the auction: a sale event, Q4, or competitors scaling. Meta's Auction Competition Change diagnostic calls a move over 20% significant. A jump in one campaign is usually that campaign's audience, placements, a learning reset after edits, or creative fatigue.
Partly, and not right away. In 2025, US monthly CPM on Birch's chart was about the same in December as in November. Biddyco's consumer-brand accounts were 24% below their October level from December 29 to January 4, and Meta reported CPMs dropping 15–30% in Western European markets in mid-to-late December as advertisers pulled back. Early January is usually the cheapest time to restart testing.
Not on its own. If your campaign optimises for purchases, Meta notes CPM may not be a good indicator of performance. Judge by cost per result. A high CPM with a healthy CPA is fine. A rising CPM together with falling CTR and rising frequency points to creative fatigue.
Use Meta's own list. Broaden the audience (Meta says ad sets of at least 2 million people often perform better), enable more placements with a 9:16 version of each ad, get out of learning by making fewer edits, fix event tracking, and refresh creative with genuinely new concepts. Meta says it subsidises relevant ads, so relevance is the cheapest lever you have.
Launch proven creative before the peak, and don't use the peak to test. In ViralMojo's BFCM 2025 dataset, the median campaign started November 6 and 30% were live by October 31. Finding your winners in October and launching the sale versions in early November lets you scale known ads during the most expensive week instead of paying to discover them.
Do the Q4 math in October. Not on Black Friday morning.
When every impression costs more, you can't afford to test blind.
Same research as above, aimed at your product: live competitor ads, organic winners, and customer language, turned into angles before the auction gets expensive.
- Competitors' live Meta ads, with days-running data
- Your niche's organic TikTok and Reels winners
- Angles and hooks in customers' own words
- Built cold from your product URL
Plans from $99/mo, flat — see ViralMojo pricing.