Pairs with the Profit-Engine Map — a ready-to-use strategy tool. Included with a subscription, or $1.99.
You open Instagram to kill thirty seconds in a checkout line. Before your thumb finishes its first scroll, an invisible auction has already run, closed, and paid out: advertisers bid against each other for the right to put something in front of you, Meta's own systems picked a winner in milliseconds, and the price of that single impression was set, charged, and booked. You never saw the auction. You saw a sneaker ad. Multiply that across 3.35 billion people who open one of Meta's apps every day3, and the slivers add up to one of the most profitable businesses in the history of technology - $164.5 billion in revenue, almost all of it from doing exactly that.1
The official story is that Meta sells your data, and that you are the product. It is the most repeated sentence in tech criticism, and it is wrong on the mechanics. Meta does not hand your data to anyone. What it sells is your attention - rented out three seconds at a time to whoever bids highest for it.
“You are not the product: you are the abandoned carcass.”7
Even the critique's intellectual godmother disowns the slogan. Zuboff argues the commodity isn't the user at all - it's the prediction of what the user will do next, sold not as raw data but as a behavioral bet.7 That distinction sounds academic until you read Meta's own filings, where the company states flatly that it makes its money by delivering ads to users, with no long-term commitments from the marketers who buy them.2 The data never leaves the building. The auction does all the work.
The auction nobody can see, running a billion times a minute
Here is the part the slogan gets backwards. An advertiser doesn't buy a list of your interests and walk away with it. It tells Meta who it wants to reach, sets a budget, and enters a continuous auction. Every time an app could show you an ad, Meta's systems run that auction in real time - matching the impression to the highest-value bid, where value is the bid times Meta's own prediction of how likely you are to act. The advertiser pays for the outcome. Meta keeps the data, the model, and the audience. This is why the business has two and only two real growth levers: how many impressions it can serve, and what each one fetches. In full-year 2024, impressions grew 11% and the average price per ad rose 10% - both pulling in the same direction at once.1 More inventory, sold for more. That is the whole engine.
| "You are the product" | The attention auction | |
|---|---|---|
| The thing sold | Your personal data | An impression in front of you |
| Who holds the data | The buyer, supposedly | Meta - it never leaves |
| What the advertiser gets | A file | A bid won against rivals |
| How Meta is paid | Per record | Per impression × price per ad |
With 3.35 billion daily active people across the Family of Apps3, even a modest number of impressions each - sold for an average price that climbed 10% in a single year1 - compounds into $164.5 billion of revenue.1 And because the auction infrastructure was built once and serves each new impression at almost no marginal cost, $69.4 billion of that fell to operating income in 2024 - an operating margin near 42% that looks like software, because it is.8
Why the same eyeball is worth 16x more in Ohio than in Indonesia
Not all attention is priced the same, and the gap is the most revealing number in the whole model. A user in the U.S. and Canada generated roughly $233 in revenue in 2024. A user in the Rest of World segment generated about $14 - sixteen times less for the same scroll, the same app, the same auction.10 The reason is brutally simple: the auction is only as rich as the advertisers bidding in it, and advertisers bid against purchasing power. A sneaker brand will pay far more to reach a shopper who can buy the sneakers. This is why headline user growth flatters and misleads. Meta keeps adding billions of people in Asia-Pacific and beyond, swelling impressions, but each of those impressions clears at a fraction of a North American one. Volume grows the top of the funnel; spending power decides what the funnel is worth. The product isn't attention in the abstract - it's monetizable attention, and geography sets the price.
Isn't this all about to be regulated - or torched on a metaverse bonfire?
The fair objection comes in two parts, and both have teeth. The first is regulatory: Meta itself warns in its 10-K that changes to data practices have already hurt its ability to target ads.2 When Apple introduced App Tracking Transparency in April 2021, requiring explicit user opt-in for cross-app tracking, Meta's targeting got blunter almost immediately — one widely-reported estimate put Facebook's ad-revenue hit at over $8 billion in the second half of 2021 alone11 — and Meta itself has disclosed that such data-practice changes have adversely affected its ad targeting.2 A model built on prediction is exposed precisely where the predictions are sourced. The honest answer is that this pressure is real but has so far slowed the machine rather than stopped it - in the same period, AI-driven ad tooling improved the targeting enough that price per ad kept climbing.1 The second objection is the spending: Reality Labs has burned $17.7 billion in 2024, $19.2 billion in 2025, and well over $80 billion since late 2020 on a metaverse the market has not asked for.56 That looks like a company torching its profits. But look at what funds it. The ad business produced $69.4 billion of operating income in 2024, and revenue grew another 22% to over $200 billion in 2025.8 The metaverse losses are staggering in isolation and a rounding error against the machine paying for them. A money machine this efficient can afford very expensive hobbies.
The most defensible advertising businesses don't sell data and they don't sell space - they sell a live auction that only they can run, because only they hold both the audience and the model that prices it. The data staying inside is the moat: a competitor can copy your ad units, but it cannot copy the closed loop where every impression teaches the model that prices the next one. Two cautions, though. First, an auction is only as rich as its bidders, so your revenue is hostage to your users' purchasing power, not just their numbers - count monetizable attention, not raw reach. Second, a model built on behavioral prediction is fragile exactly where it sources its signal; when a platform or a regulator cuts off the data feed, the auction still runs, but it prices blind. Defend the loop, and don't mistake scale for monetization.
Meta makes its money the way a tollbooth on the busiest road in human history makes its money - except the toll isn't charged to the traveler. It's charged to everyone who wants the traveler to look their way, and it's set by an auction that runs a billion times a minute, in the dark, on data that never leaves the booth. "You are the product" was always the wrong frame. You were never sold. Your attention was rented - and the genius was never the targeting or the metaverse moonshot. It was building the one place nearly four billion people return to every day, then standing at the door and charging admission to everyone but you.
Profit-Engine Map
A one-page map that pulls a business apart into the hook that gets the customer in the door and the engine that quietly earns the margin. Use it to see where the real profit lives, how the two halves are wired together, and what breaks if the link is cut. Blank to dissect your own P&L; filled as the worked example of a business whose advertised product is not where it makes its money.
Included with any subscription, or unlock this tool for $1.99. Get it → · See plans →
Sources
Where this comes from — the filings, records, and reporting behind it.
- 1Meta generates substantially all of its revenue from advertising; FY2024 total revenue was $164.50 billion, a 22% year-over-year increase; ad impressions grew 11% and average price per ad rose 10% for full-year 2024.
- 2Meta's FY2024 Form 10-K states the company generates substantially all revenue from advertising, that marketers have no long-term commitments, and that changes to user data practices have adversely affected ad targeting ability.
- 3Family daily active people (DAP) reached 3.35 billion in Q4 2024, up from 3.19 billion in Q4 2023; Family monthly active people (MAP) was 3.98 billion as of December 31, 2023.
- 4Meta's worldwide ARPU was $49.63 in 2024, up from $44.60 in 2023; U.S. & Canada ARPU was approximately $233, roughly 16x higher than Rest of World (~$14).
- 5Reality Labs posted an operating loss of $17.7 billion in FY2024 on $2.1 billion in revenue; the unit has accumulated over $80 billion in total operating losses since late 2020.
- 6Reality Labs FY2025 operating loss was $19.2 billion, bringing the cumulative total close to $90 billion over seven years; Meta laid off ~1,000 Reality Labs employees in January 2026 to shift resources toward AI-powered wearables.
- 7Shoshana Zuboff explicitly rejects the 'you are the product' framing, writing instead that 'you are not the product: you are the abandoned carcass,' arguing that behavioral prediction futures — not users themselves — are the commodity sold to third parties.
- 8Meta's operating income for FY2024 was $69.38 billion (up 48% year-over-year) and net income was $62.36 billion; FY2025 total revenue reached $200.97 billion (up 22%), with operating income of $83.28 billion and a 41% operating margin.
- 9Zuboff writes in The Age of Surveillance Capitalism: 'You are not the product; you are the abandoned carcass. The "product" derives from the surplus that is ripped from your life.'
- 10Meta's US & Canada ARPU was $233.42 in 2024, approximately 16.7x the Rest of World figure (~$14), sourced from Meta's Q4 2024 earnings presentation.
- 11Apple's App Tracking Transparency (ATT), introduced with iOS 14.5 in April 2021, cost Facebook, Twitter, Snap, and YouTube nearly $10 billion in ad revenue in the second half of 2021, with Facebook bearing the largest share of the loss at over $8 billion.