Why a media plan takes longer to price than to sell

The call itself is short. A buyer wants a connected-TV flight for a launch, sends a brief, and asks one question: what's your CPM. The buyer spends twenty minutes on my answer. I spend two days producing it. The twenty minutes are what everyone calls selling. The two days are where the money is decided.

The delay is not the spreadsheet. The spreadsheet is fast. The delay is that "CPM" is the last line of an argument with four premises, and each premise has to be settled with somebody who isn't in the room: a publisher partner who can confirm what premium inventory is actually open in those weeks, an ad-ops lead who knows what this mix has completed at historically, and whoever signs off on how much risk we are willing to carry this quarter.

Four decisions, one number

Here is what a CTV price contains, in the order I resolve it.

Minimum commitment. Volume is what I take to publishers. A commitment large enough to matter gets better terms on inventory; a small one doesn't, and it still carries the same setup work: creative checks, measurement, the trafficking nobody bills for. A small plan is expensive per thousand not because anyone is greedy but because fixed work is being divided by a small number.

Completion guarantee. The buyer wants to pay for completed views, not started ones. Reasonable. But the moment I guarantee a completion rate, I own the gap between what the inventory actually completes and what I promised. Big-screen inventory completes in the high nineties without anyone trying; the second-screen and in-app tail of a mix does not. If the blend completes at ninety percent and I promised ninety-five, I buy the extra impressions to close the gap, and nobody pays me for them. The guarantee is a price line, not a footnote in the terms.

Premium share. What counts as premium is a negotiation of its own: big-screen apps with a known content owner, verified household reach, a sane ad load. It costs more per thousand. It also completes better, which is the part sellers forget to say out loud. A plan with more premium needs less top-up to hit the guarantee, so the two decisions partly pay for each other. Price them separately and you overcharge and lose. Price them together and you can explain why the number is what it is.

Who carries under-delivery. Forecasts miss. The only question is what happens when they do: bonus impressions in the next flight, a pro-rata credit, or a credit plus a penalty. Each option puts a different amount of risk on my side, and risk has a price. A buyer who insists on penalties is not being difficult. They are buying insurance, and insurance is not free.

Slide the four decisions and watch the quote move. Then move the last slider, the discount, and watch what happens to the margin. Ten percent off the price is not ten percent off the margin. In the model above it is about a third of the margin in money, and fifteen off takes close to half, because the cost line doesn't move when the price does. That asymmetry is the whole article. Most sellers learn it from the finance team at the next quarterly review, which is late.

By the time they say yes, the number is old

The second problem starts after the quote goes out. CTV inventory is not stock on a shelf. It is a forecast of what publishers expect to have in a given window, and my plan booked the forecast, not the impressions.

Meanwhile the quote travels: the agency planner, the client's marketing lead, procurement, sometimes legal. A week if everyone is at their desk, three if a holiday sits in the middle. During that time other plans book the same premium slots, because premium is the scarce part and everybody wants the same weeks.

When the yes arrives, usually one of three things is true. The premium share I promised is no longer available, so it quietly becomes something less premium. Or it is available at a higher cost, and I eat the difference. Or the flight shifts and the whole forecast is different. None of these changes the quoted CPM, because the CPM is already in the buyer's deck with a signature next to it. All three change the margin.

What I do about it is unglamorous. A validity date on every quote, and I mean it. A re-run of the plan against live availability before anything is signed, even when the buyer says nothing has changed. And I treat "same terms as last time" as a new calculation, because last time's inventory doesn't exist anymore.

The two-minute test

Now the meeting. The buyer has two quotes on the table and says the other one is cheaper. This is where the two days of pricing either pay off or evaporate.

A seller who can retell the four decisions asks: which one do you want to give up? Drop the completion guarantee and the price moves this much. Lower the premium floor and it moves that much. Commit to more volume and I can move it myself. Every one of those is a trade, and a trade leaves the margin alone.

A seller who can't retell them says "let me see what I can do", steps out, and comes back with ten percent off. The buyer got nothing they asked for except a smaller number, and the seller gave away a third of the margin for the privilege of not explaining themselves. I have done this. It feels like closing. It is the opposite.

A discount is what a seller pays for not knowing their own price.

I am not against discounts. I am against discounts that replace an argument. If the buyer genuinely doesn't need the guarantee, removing it is a better deal for both sides than a discount that keeps it. If they can commit to volume, the publisher terms improve and I pass some of that through. Money is the last thing to give away, because it is the concession that is hardest to take back at the next renewal.

So I make my team do it out loud. Take a plan, explain the price in two minutes, no slides. If they can't, the plan isn't ready to send, whatever the spreadsheet says. The spreadsheet was never the slow part.