What Happens When the Cost to Advertise Gets Too High?
· Articles

Phil here, Director of Paid Social. Yea, another CPM and CPMr article...
Everyone optimises for CPA. Everyone watches ROAS. Nobody talks about the base cost of showing up in the auction in the first place.
The cost to advertise is the price you pay to deliver your ads to an audience. Not what you pay per conversion. Not what you pay per click. The underlying cost of reaching a person with your message. It's made up of two numbers: CPM, the cost per thousand impressions, and CPMr, the cost per thousand unique people reached.
When the cost to advertise rises, everything downstream gets harder. CPA climbs even when your creative is fine. ROAS falls even when your offer hasn't changed. The funnel hasn't broken. The floor has moved.
I pulled 15 months of data across eight accounts and five verticals. Here's what the cost to advertise actually looks like, and why it matters more than most people realise.
Your Category Sets Your Floor
Before you've made a single decision, your vertical has set the base price you're competing at.
Across January 2025 to March 2026, average CPMs across the dataset:
Womenswear: $7.69
Automotive: $9.58
Grooming: $16.34
Menswear: $17.32
Food and Beverage: $20.20
Sleepwear: $22.67
Nutrition Tech: $35.76
A Nutrition Tech brand is paying nearly five times the CPM of a Womenswear brand before either of them has done anything. That's not inefficiency. That's the market. The audience is more valuable, the competition for that attention is higher, and the base price reflects it.
A $35 CPM in health isn't a problem if your offer, messaging, and conversion rate are calibrated for it. A $35 CPM brand that has built its model around that floor can run profitably for years. The problem isn't the number. The problem is when the number moves.
It's the Drift That Breaks Accounts
The category tax is what it is. What breaks accounts is when the cost to advertise starts climbing away from its historical level.
Nutrition Tech in this dataset moved from $20 CPM in April 2025 to $51 CPM in November. That's a 154% swing in the cost of reaching the same audiences in the same category. Not because the category got more competitive overnight, because auction dynamics shifted through the year and peaked in Q4.
More telling: a Grooming account in January 2025 was running at $10.45 CPM with a $61 CPA. By August, CPM had reached $23. CPA had moved to $135. CPM up 122%. CPA up 120%. No dramatic creative failure. No landing page change. The cost of showing up in the auction went up, and CPA followed it almost exactly.
This is the mechanism. When CPM rises and your conversion rate holds flat, CPA has to rise with it. The only question is whether the rise is temporary, seasonal, correctable, or structural.
CPM Is Only Half the Story
CPM tells you what you're paying per thousand impressions. It doesn't tell you how many of those impressions are going to the same people.
CPMr, CPM multiplied by frequency, is what you're actually paying to reach a thousand unique people. And the gap between them in these accounts is the most revealing number in the dataset.
Menswear shows the extreme case. CPM has barely moved across 15 months: $14–$22 in a narrow band, nothing alarming on the surface. But CPMr has ranged from $70 to $228, with frequency hitting 11.99x in October 2025. The same person seeing the ad twelve times in a month. The cost to reach a genuinely new person, someone who hasn't already seen the ad eleven times, is $228 per thousand.
That account is not in a CPM crisis. It's in a CPMr crisis. And it won't show up clearly in a standard reporting view unless you're tracking CPMr alongside CPM.
This is what spending more to reach fewer people looks like at account level. The CPA looks stable because the account is still converting — the warm audiences being cycled repeatedly are eventually buying. But the cost of finding a genuinely new customer is being paid invisibly, in the frequency, not in the headline metric.

Everyone watches ROAS and CPA.
By the time those move, the problem has been building for weeks.
CPMr moves first. It's CPM multiplied by frequency. When it climbs consistently, your reach pool is narrowing and the algorithm is stuck in a loop re-serving ads to the same people.

Why I don't shut up about CPMr
I talk about CPMr a lot. Rising CPMr is one of the earliest signs that something is wrong with an account: before ROAS drops, before CPA climbs, before anyone…

The Mountain Metaphor
There's a version of this problem that's seasonal and predictable, and it deserves its own discussion.
Q4 is when the cost to advertise spikes for everyone. Every account in this dataset saw CPMs rise in November 2025. The range was +16% for Womenswear to +142% for Grooming. The auction clears at a higher price because every brand with a Q4 push is competing for the same attention in the same window.
The thing is, customer intent also rises in Q4. Available cash increases. Conversion rates often improve. So a higher cost to advertise in Q4 isn't inherently bad, the market is paying more because the quality of attention is higher.
The mistake is treating Q4 as the starting line.
Think of it as a mountain. The summit is Q4, the highest point in terms of both cost and opportunity. The brands that do best in Q4 are the ones that started climbing in Q2. They built reach, built frequency with new audiences, built enough warm pipeline that when CPMs doubled in November they weren't starting cold. They were already near the top.
The brands that start climbing in October are scrambling up a steep face in a crowd. They're paying peak CPMs to reach audiences that don't know them yet, building frequency from zero at the most expensive time of year. The cost to advertise is the same for everyone. The difference is where they started.
What This Means for How You Plan
The cost to advertise is partly the market and partly you. The market part, your category floor, seasonal swings, you can't change. The you part, frequency, exclusion hygiene, creative diversity, how much of your budget is reaching the same people for the thirteenth time, you can.
Track CPM and CPMr together, not separately. A stable CPM with a rising CPMr is not a stable account. It's an account where the cost to advertise is growing invisibly, in the frequency column, not the headline one.
Watch for drift. The absolute level of your CPM matters less than whether it's moving. A Nutrition Tech brand at $35 CPM can build a profitable model. A Nutrition Tech brand that was at $35 CPM and is now at $70 has a different problem, and it's not a creative problem.
And if you want Q4 to work, start the climb in spring. The cost to advertise will be higher in November regardless of what you do. The question is whether you're warm enough to absorb it.
Read more about CPMr:https://taikundigital.com/phil-kiel/cpmr-calculation
Book a call with me.