Meta Raised Ad Prices 14% While Spending More Than It Earns. Anderson, Day and Rangan Call That a Channel-Power Problem.
Meta reported third-quarter results on October 21, and the headline was growth: revenue of $64.6 billion, up 26%. The number that matters sits underneath it. Ad impressions rose 10%. The average price per ad rose 14%. More than half of Meta’s growth came from charging advertisers more for each ad, not from showing more of them. In the same quarter Meta spent $34.9 billion on capital expenditure against $32.8 billion of operating cash flow, swung free cash flow from positive $10.6 billion a year earlier to negative $2.1 billion, and raised its full-year capex guidance to $135 to $145 billion. CFO Susan Li credited AI with “more efficient monetization.”
Table of Contents
Most CMOs will read this as validation: AI is making the platform work better, so the higher price is earned. Most boards will read it as a strong vendor with a strong balance sheet. Both reactions skip the question that matters. When your largest channel raises its price while funding an infrastructure bet larger than its own cash generation, your media budget has quietly become a contributor to someone else’s capital plan. That is a channel strategy problem, and it belongs on the CEO’s desk.
Is Your Biggest Channel a Partner or a Landlord?
In 1997, Erin Anderson, George Day and V. Kasturi Rangan published “Strategic Channel Design” in MIT Sloan Management Review. Their argument was that executives treat channels as logistics when they are strategy, and that channel decisions are driven by three forces: customer needs multiplying, strategic priorities shifting, and channel power moving. Almost every discussion of platform advertising addresses the first two. The third is the one Meta’s quarter illustrates.
Channel power sits with whoever controls the customer relationship, the data, and the price. In the paid social model, the platform holds all three. It sees the customer, it measures the result with its own instruments, and it sets the price in an auction it also designs. A 14% rise in price per ad, justified by better AI-driven targeting that only the seller can fully observe, is channel power being exercised in plain sight. Anderson, Day and Rangan would not call that a partnership. They would ask the question they asked in 1997: was this channel designed from the customer back, or did it simply become the one your company grew dependent on?
What I Learned About Renting Reach
When I co-founded Bolt Goodly and we built it into a nine-figure agency, the pattern I saw most often among clients was concentration by default. Spend flowed to whichever platform reported the best return, and the platform’s own dashboard was the scoreboard. It worked until the platform repriced, at which point the client discovered it had built a growth model on a cost line it did not control. The agency could not fix that, because it was never a media problem. It was a design problem made years earlier, one reasonable budget decision at a time.
At Pinnacle, the 60% cost reduction was only possible because of a prior decision about which relationships the company would own and which it would only rent. Reach could be rented. The customer relationship could not. Every dollar we spent in a channel we did not control had to be justified against what it left us holding afterward: a known customer, a measurable pattern, an asset on our balance sheet rather than an invoice in someone else’s.
The Price Rise Is Not Inflation. It Is Rent.
Here is the position I would defend. Meta’s price increase is rational and probably not even unfair. If the AI is delivering better outcomes, advertisers are being charged for real value. But notice what follows from that. The platform is now setting both the value and the price of that value, and funding a capital program of roughly $140 billion on the strength of your willingness to pay it. The only party at the table with no pricing power is the advertiser. Anderson, Day and Rangan list market learning as a design principle for exactly this reason: a channel is also your best sensor of the market, and a channel that keeps what it learns from you makes you dependent on it.
The conventional executive response is to chase efficiency inside the platform, with better creative, more automation, and tighter bidding. That is optimization of a position someone else is repricing. The more useful response sits at the level of capital allocation. A CFO would never let a single counterparty hold half of the company’s cash without a concentration limit. Most boards have no equivalent rule for the single channel that holds half of the company’s demand generation. Dependence on a supplier is a risk, and risk gets a ceiling. Few CMOs have been asked to set one.
The counterargument deserves a hearing. Reach this efficient is hard to replace, and walking away from the best-performing channel to make a strategic point is its own kind of mistake. I am not arguing for leaving. I am arguing that the decision to stay should be made on purpose, with the dependence priced in, rather than discovered in a quarter when the auction moves again.
The Question Your Board Should Be Asking
Every platform earnings call from here on will carry the same subtext. AI is expensive, the spending is enormous, and the revenue to justify it will come from somewhere. Some of it will come from your line item. The CMOs who matter in five years will not be the ones who squeezed the most return out of a channel that was repricing under them. They will be the ones who designed a channel system in which no single supplier could set their cost of growth.
What share of your company’s growth is rented from a supplier that just posted negative free cash flow and raised its spending guidance, and who on your board owns that number?
Discover more from Steve Wolf
Subscribe to get the latest posts sent to your email.
