You launched new creatives. Your CPM was still up. You refreshed the audience. Still up. You duplicated the campaign. Nothing changed. By the time most ecom brands contact us, they've already spent two or three weeks doing what every guide on the internet told them to do — and the CPM is exactly where it was, or higher.
A sudden CPM increase on Facebook ads means the cost per 1,000 impressions has risen sharply — often doubling or more — without a corresponding change in your budget, creative, or targeting. Based on our work diagnosing performance issues across hundreds of ecom accounts, roughly half of all persistent CPM spikes that don't respond to the usual creative fixes have an account-level or internal cause that creative changes simply cannot address.
This guide covers the complete diagnosis — three layers of causes in order, the data patterns that distinguish them, what you can fix yourself, and what requires a deeper intervention. We'll also introduce what most advertisers never find out: that Meta's internal scoring system assigns a trust and delivery quality score to your assets, and that score can become the single biggest driver of your CPM without leaving a single visible trace in Ads Manager.
What a CPM spike actually means — and what it tells you
CPM — cost per mille, or cost per 1,000 impressions — is not a number you directly control. It's the outcome of Meta's auction system: how much you end up paying to reach 1,000 people, based on competition, your account's delivery quality, and Meta's internal evaluation of how much it should cost to show your ads.
A healthy ecom CPM in 2025 typically sits between €10 and €20 depending on niche, seasonality, and placement. Q4 pushes this higher for everyone due to auction competition — global CPM data shows a consistent spike every November, peaking around €25 before resetting in January. This is normal and expected. What isn't normal is a CPM that doubles or triples outside of peak season, holds at that level for weeks, and doesn't respond to creative changes.
That pattern — elevated CPM, unresponsive to creative refresh, persisting across multiple campaigns — is the diagnostic signal that tells you the cause is not in the ad. It's in the account or the infrastructure around it.
Layer 1 — External causes: rule these out first
Before going deeper, rule out the obvious. These are the causes everyone knows about. If your CPM spike is recent and you haven't checked these yet, start here. If you've already been through all of them, skip to Layer 2.
Creative fatigue
When the same ad has been running long enough that your target audience has seen it multiple times, engagement drops, Meta interprets this as low relevance, and CPM rises. The signal in Ads Manager: frequency climbing above 3–4 on cold audiences, CTR falling, CPM rising in step. The fix is straightforward — new creative. If your CPM recovers within a few days of launching fresh ads, this was the cause.
Audience saturation
Smaller or tightly defined audiences exhaust faster. When you've reached most of the available pool, Meta has to work harder (and charge more) to find new people. Check your audience size relative to your daily reach — if you're hitting 50–60% of a small audience regularly, saturation is likely a contributing factor. Broadening your targeting or introducing new audience sets usually resolves it.
Seasonal auction competition
Between October and December, every advertiser with a Q4 budget enters the auction. CPMs rise for everyone regardless of account health. If your spike coincides with this window, seasonality is at least a partial cause. The same applies around major retail events — Valentine's Day, Black Friday, back-to-school periods — where competition concentrates briefly and then eases. Patience is sometimes the right answer here, not a structural fix.
Budget scaling done too aggressively
Increasing daily budget by more than 20–30% in a short period can force Meta's algorithm out of its optimised delivery pattern and push the campaign back into a learning phase. During this relearning window, CPM is typically elevated and unstable. If the spike followed a significant budget increase, give the campaign 7–10 days to stabilise before diagnosing further.
Layer 2 — Account-level causes: what most guides miss
This is where the standard advice stops. Layer 2 involves signals inside your Meta account that directly affect how much you pay to reach people — not through the auction dynamics of Layer 1, but through Meta's evaluation of your business and page as an advertiser.
Your Facebook feedback score
Your Facebook feedback score is one of the most directly impactful account-level CPM levers — and one of the most overlooked. It's a 0–5 rating Meta calculates based on post-purchase customer surveys about their experience with your business: did the product arrive, was it as described, was service responsive.
Meta uses this score to evaluate your business's trustworthiness as an advertiser. The lower the score, the more expensive and restricted your ad delivery becomes. According to Meta's own advertising standards, when a page's score drops below acceptable standards, Meta applies an official delivery penalty — your ads reach fewer people and CPM rises substantially. Our experience shows the impact begins well before the official threshold.
The critical thing most advertisers don't know: the feedback score is tied to your Facebook page, not your ad account. You can create a new ad account, launch fresh campaigns, and still be paying inflated CPMs because the page connected to those campaigns carries a damaged score. The penalty travels with the page, not the account.
Ad rejection rate and policy warnings
Every time Meta rejects an ad, it registers against your account's health record. A handful of rejections over months is normal. A cluster of rejections in a short period — especially for the same category of issue — signals to Meta's systems that your account may be operating in a way that needs scrutiny. This can trigger increased review times, delivery throttling, and higher CPMs as Meta's systems apply more friction to your ads. Check your Account Quality page for any open policy issues or warnings.
Business Manager reputation score
Since late 2024, Meta has made Business Account Reputation Scores more visible within Business Manager — ratings ranging from Excellent to Poor that affect ad delivery, approval times, and access to advanced ad features. If your BM reputation score has declined, it can manifest as a CPM increase across all campaigns running under that Business Manager, regardless of individual page or ad account health. This is a less commonly known signal but worth checking alongside your feedback score.
Layer 3 — Internal Meta signals: the cause nobody talks about
This is the layer that explains the cases where everything else checks out. Your feedback score is fine. Your account quality page shows no warnings. You've refreshed every creative. You've expanded your audiences. The CPM is still elevated, and nothing moves it.
In our experience, this scenario — persistent unexplained CPM elevation after all surface fixes have been applied — points to internal Meta signals. Specifically, a degraded HiVA score or active delivery throttling on your assets.
What HiVA is — and why it matters more than most advertisers realise
HiVA is an internal score Meta assigns to each advertising asset — your ad accounts, pages, and business managers — to rank their trustworthiness and delivery quality. It is not visible in Ads Manager. It doesn't appear in Account Quality. There is no notification when it degrades. But it directly influences how Meta distributes your ads across its auction system.
Think of it this way: two advertisers with identical budgets, identical creatives, and identical target audiences will not necessarily pay the same CPM. If one has a healthier HiVA score on their assets, Meta's delivery algorithm treats their campaigns as higher quality and delivers them more efficiently — to better audiences, at lower cost. The one with the degraded HiVA score is paying a silent premium on every impression, with no indication in the UI that this is happening.
This is the insight that genuinely surprises most advertisers when we explain it: your internal Meta score can matter more than your creative. A technically excellent ad running on a degraded asset will underperform a mediocre ad running on a trusted, healthy asset. This doesn't mean creative doesn't matter — it does, significantly. But it means creative is not the only lever, and sometimes it's not even the most important one.
What delivery throttling actually does to your account
When Meta's systems flag an asset as lower quality through internal signals, they throttle its delivery. This throttling operates on two levels simultaneously: it reduces the quality of the audiences your ads are shown to (prioritising lower-value inventory), and it reduces the auction competitiveness of your campaigns (your bids win fewer high-quality placements at a given spend level). Both effects raise CPM and reduce ROAS, and both are invisible to you in Ads Manager.
The result looks, from the outside, exactly like a saturated audience or a fatigued creative. CPM is up. Reach is down relative to spend. Performance has degraded. The standard response is to refresh the creative — which doesn't work, because the problem isn't the creative. The creative is running through a damaged pipe.
What triggers internal signal degradation
Internal signals degrade through a combination of factors, and diagnosing the specific cause requires a detailed audit — not just a look at metrics, but a thorough review of the full advertising setup: account structure and history, connected assets, ad strategy, landing page experience, and delivery patterns over time. The triggers we see most commonly across ecom accounts include:
- A declining or penalised feedback score that has persisted for months without correction
- A high rate of ad disapprovals, particularly for repeated policy categories
- Aggressive budget scaling that triggered internal quality flags
- Billing issues or payment failures, even if subsequently resolved
- Poor landing page quality signals — high bounce rates, misleading offers relative to ad claims
- Long-term account history that has accumulated low-quality signals across multiple issues
Crucially, these factors often compound over time. An account might tolerate one without significant CPM impact, but two or three operating simultaneously can push internal scores into a range where throttling becomes active and persistent.
What this looks like in practice — a real diagnosis
A skincare brand had been running profitably on Facebook for about fourteen months, spending around €25,000/month. CPM had historically sat between €12 and €16. Over approximately six weeks, it climbed to €34 — more than double — with no changes to creative, targeting, or budget.
The team had already done everything correctly by standard advice: four new creative batches, two audience expansions, campaign restructure, budget reduction to force relearning. Nothing worked. They came to us after the sixth week.
The audit revealed three issues operating simultaneously. First, the feedback score on their primary page had been sitting at 2.3 for three months — close enough to the penalty threshold that delivery was already affected, but not low enough to trigger the most visible restrictions. Second, a landing page that had worked well during their growth phase was generating increasing post-click dissatisfaction signals as the brand scaled to new, colder audiences who had different expectations from the ad claims. Third, a billing dispute that had been resolved four months earlier had left a low-quality signal on the account's payment history that had never been addressed.
None of these were visible in Ads Manager as discrete problems. Together, they had degraded the internal delivery quality of the account to the point where CPM had roughly doubled regardless of creative quality.
After addressing the feedback score, updating the landing page to better align with ad claims, and working through the account health intervention, CPM began declining within about ten days and returned to the €14–17 range over three weeks — without any further creative changes.
The CPM Diagnosis Stack — how to work through it in order
The right approach to diagnosing a CPM spike is sequential. Start at the surface, rule out each layer, and only move deeper when the surface causes have been genuinely eliminated. Here is the framework we use:
- Check the external signals first (5–10 minutes)
Review frequency, CTR trend, audience saturation, and whether the timing aligns with seasonal auction pressure. If these explain the spike, fix them. If not, move to step 2. - Audit your account-level health (15 minutes)
Check Account Quality in Business Manager. Review your page's feedback score. Look for policy warnings, rejection clusters, and your BM reputation score. If anything is below healthy thresholds, treat this as a likely contributor. - Review your full setup holistically
Look at your account structure, asset history, ad strategy, landing page alignment, and delivery patterns over the past 3–6 months. This is the layer that catches what metrics alone won't show — and it's the diagnostic step that takes specialist experience to do properly. - Match the cause to the fix
Use the table below. Creative fatigue, audience saturation, and seasonal pressure are self-serviceable. Account-level and internal signal issues require targeted intervention — fixing the wrong thing wastes time and budget.
How to fix a CPM spike — matched to the cause
The fix depends entirely on the diagnosis. The most common mistake is applying a creative solution to an infrastructure problem. Here is the correct fix for each cause type:
"The brands that lose the most time — and budget — on a CPM spike are the ones that keep testing creatives while an account infrastructure problem goes undiagnosed."
When to handle it yourself — and when to get help
Layer 1 causes are entirely self-serviceable. If creative fatigue, audience saturation, or seasonality explains your CPM spike, you have everything you need to fix it. These are the most common causes for newer accounts and brands still in their early scaling phase, where account history is short and internal signals have had less time to build up.
Layer 2 — account-level — sits in the middle. You can identify the issues yourself (check Account Quality, read your feedback score), and for mild cases the fixes are within reach. Improving customer fulfilment, addressing shipping complaints, resolving policy warnings — these are operational fixes you can execute internally. For accounts with more severely degraded scores, or where the score has been low for an extended period, the timeline for self-recovery is long and the CPM impact continues accumulating throughout.
Layer 3 — internal Meta signals — is not something you can diagnose or fix through Ads Manager. There is no setting to adjust, no button to push. It requires a holistic audit of your full advertising setup by someone who knows what they're looking for and how to address what they find. If you've been through Layers 1 and 2 and the CPM spike is still unexplained, this is where you are.
.avif)
