Why CPM Without Frequency Is a Vanity Metric

Picture the last meeting of the quarter. The lights are down, a dashboard is up, and near the bottom of the screen one number is glowing: a cost-per-thousand of $4.10. Someone says the word "efficient," the way you'd say it about a German car. There is a small, genuine feeling of victory in the room.

Now ask the only question that matters, and watch the victory evaporate: does anyone out there actually remember the ad?

Nobody knows. Nobody measured it. They measured how little it cost to appear, and quietly assumed that appearing and landing were the same thing. They are not the same thing. They are barely related.

The price of a door, not a room

CPM (cost per mille, cost per thousand impressions) is one of the most trusted numbers in marketing, and one of the most misunderstood. It tells you the price of being present. It does not tell you whether presence turned into anything. It is the cost of the door being open, not the cost of anyone walking through it, and certainly not the cost of anyone staying.

Here's the trap. CPM has a wonderful, seductive property: you can almost always make it go down. Broaden the targeting. Chase cheaper inventory. Buy the 3 a.m. impressions, the below-the-fold slots, the audiences nobody else is bidding on. Do all of this and your CPM will fall like a stone, every chart will point the right way, and you will be (this is the uncomfortable part) optimizing, with real skill and real diligence, for the privilege of being ignored at a discount.

Cheap attention that nobody pays isn't a bargain. It's just cheap.

A man at General Electric, 1972

In 1972, a psychologist named Herbert Krugman, who ran opinion research at General Electric, published a short paper with a title that sounds like a shrug: "Why Three Exposures May Be Enough." It became one of the most quietly influential ideas in the history of advertising.

Krugman's claim was that there is no such thing, psychologically, as a single impression doing much of anything. The mind processes an ad in stages, and it needs to come back. The first time you see something, he argued, the only question your brain is really asking is "What is it?". It's cataloguing a shape, filing a novelty. The second exposure is where the real question arrives: "What of it? Does this have anything to do with me?" And the third is the reminder, the one where the thing finally settles into memory, and, Krugman noted, where attention also begins to politely withdraw.

Read that again, because it's the whole game. The first impression (the one CPM is busy making cheaper and cheaper) is the least valuable one. It's a handshake with someone whose name you forget before they've let go of your hand. Meaning doesn't get made on the first pass. It gets made on the return.

Four years earlier, a Michigan psychologist named Robert Zajonc had shown the mechanism underneath it: show people something repeatedly (a shape, a word, a face) and they will, reliably, come to prefer it, for no reason they can name. He called it the mere-exposure effect. Familiarity, it turns out, isn't a byproduct of advertising. Familiarity is the product.

Same spend. Same CPM. Different universe.

So consider those two campaigns again: identical, beautiful $4 CPM, identical budget. The first spreads itself across a million people, each of whom sees the ad exactly once. Efficient! Enormous reach! A number to put in a deck. The second spends the same money to reach a quarter of a million people four times each.

On the dashboard, these two campaigns are twins. Same cost, same efficiency, same triumphant CPM. In the world (the only place that pays you) they are not remotely the same. The first campaign is a rumor: a million people who couldn't pick your brand out of a lineup an hour later. The second is a reputation: a quarter-million people carried past Krugman's threshold, into the territory where memory actually forms.

The dashboard called them identical. Only one of them worked.

What we count instead

This is why, at 3b11, a naked CPM never leaves the building as a headline number. On its own it answers the wrong question (how cheap was it to show up?) and stays conspicuously silent on the only one worth asking: did we land?

So we fold two dimensions back into the number the moment an impression is served. How often the right person saw it: the frequency. And how long the message actually held them: the dwell time. Multiply those together and you get what we call TAG, Total Attention Grab: not the cost of appearing, but the amount of genuine mindshare a dollar actually bought. Frequency drags a cheap impression toward memory. Dwell time proves someone was home.

A low CPM buys you width. Width alone is a rumor. Frequency and dwell are what buy you depth, and depth is the only thing that compounds.

So the next time a beautiful CPM glows at the bottom of a slide and the room reaches for the word "efficient," ask the older, better question first. Not how cheap was it to appear. Did we land?