Unique CPM is the metric you're missing
There's a stage for every business where growth starts to stall. Not failing or in trouble, just not growing. If that's you, the answer is probably not another attribution tool or a hotshot new hire.
I wrote a few months back that most teams don't have an attribution problem, they have a judgment problem. This is an operational follow-up to that piece. It's also a different from the ad efficiency piece I wrote. That one was about why your ads stopped working today. This one is about how to find the next pocket of growth when you've already stabilized.
The overlooked numbers
Most teams optimize against CPA or ROAS. That's the right place to start. If your conversion events are set up correctly, those metrics tell you what each channel is doing against the immediate goal you set.
What gets missed is what you're paying to reach a unique person. That's a different question, and it shows up in a metric almost nobody actually tracks: unique CPM.
Unique CPM is your platform CPM multiplied by your account-level frequency. If your Meta CPM is 10 dollars and your frequency is 6, your unique CPM is 60 dollars. That's what attention actually costs you per unique person reached. The platform shows you cost per impression, which is a different thing altogether.
Run that across your channels and you'll get something like this:
- Meta at 10-dollar CPM × 6 frequency = 60 dollars effectively
- YouTube at 30-dollar CPM × 2 frequency = 60 dollars effectively
- Remnant TV at 1.50-dollar CPM × 2 frequency = 3 dollars effectively
Suddenly the channel you thought was cheap looks not quite as good, and the channel you thought was expensive isn't all that bad.
CPA or ROAS tells you whether the channel is doing its job today. Unique CPM tells you whether you can keep getting unique reach as you push spend up.
New reach is what ultimately drives growth. Once you've saturated an audience, every additional dollar in that channel is paying for another impression on someone who's already seen you. Unique CPM is the cleanest read on whether you're actually expanding the audience or running it back through the same people.
Both matter, but most teams only watch the first one.
Underpriced is the key word
Cheap CPMs and underpriced attention are not the same thing.
For example, display traffic might have really low CPM and be 5 cents a click. The quality is usually low enough that you're paying full price for what you're getting, even at that cost. Bad inventory is just bad inventory you can buy a lot of.
Underpriced attention means the cost of reaching the right person is low relative to the quality of the attention you're getting in return.
A 60-dollar unique CPM on YouTube might still beat a 3-dollar unique CPM on remnant TV if YouTube viewers convert at meaningfully higher rates. You have to weight it with quality. Same-day demand versus long-term brand lift. Propensity to buy this week versus propensity to buy in three months because the brand stuck somehow. You won't get this exactly right, and that's fine. The point is that you're actually looking at it.
The pullback problem
Pulling back from Meta, for example, to fund a new channel with better unique CPMs shrinks your top line in the short term while the new channel is still ramping. Now you have a smaller business under the same fixed costs, and the pressure shows up everywhere.
The redeployment has to be paired with operational improvements somewhere else in the business. If you can lift conversion rate by 20 or 30 percent through better messaging, sharper product pages, and stronger offers, you can absorb the traffic pullback without losing top line. When the new channel ramps after that, things compound: Meta gets cheaper because the audience is warmer, your conversion rate is higher across the board, and the new channel is buying you reach you couldn't get before.
Real media buying decisions cut across pricing, conversion, retention, and product. Treating media buying like a standalone function is what keeps a lot of teams stuck on the same playbook for years.
When to ignore this
This is a framework for evergreen growth. Steady-state, ongoing decisions about where to put marginal dollars.
There are times when this isn't the right move. Holiday is one example. New product launches are another. Anything time-bound where you need to ramp aggressively in the channels you control most. During those windows, push Meta or whatever channel you have the most leverage in. The framework still applies, but the weighting shifts toward speed and known performance.
Why this isn't popular advice
It requires more work than looking at a platform dashboard.
Marketers, myself included, are pretty good at being lazy. We love to find shortcuts and hacks. The longer route is asking what bet you're actually making with your media spend, where the next pocket of attention might be, and what you'd have to fix in the rest of the business to take advantage of it.
That's judgment work, the kind that doesn't show up in any tool you can buy.
Underpriced attention is also short-lived. Markets find equilibrium, and the pocket you find this quarter probably won't be there in two years. The pockets change. The discipline of looking for them doesn't. If you're always asking where the underpriced attention is, you'll find the next pocket before everyone else rushes in.
For most DTC brands without retail distribution, paid attention is your distribution channel.
P.S. Quick exercise. Pull this for each of your channels:
- CPM
- Account-level frequency
- Net new customers acquired
Calculate unique CPM (CPM times frequency) and true CPA (spend divided by net new customers).
Now compare across channels. Which ones are actually delivering unique reach for your money? Which ones are you paying to put another impression in front of someone who already knows you?
If the answers surprise you, that's the start of the conversation about where the next dollar should go.