The bill that never arrives
Look at the phone in your hand and add up what it costs to run. The map that finds your route, the app that carries your messages, the video that plays while you wait, the feed you scroll before sleep: almost none of it is billed to you. You paid for the handset once, and you pay a carrier for the connection, but the services themselves, the ones you actually spend your hours inside, are free to install and free to use. A large share of a modern day now runs through them. People pass hours with screen media that charge nothing at the point of use, and the businesses that provide those hours, the search engines and social networks and video sites, are among the largest companies in the world. A service given away to hundreds of millions of people, by a firm worth hundreds of billions of dollars, is not charity. The money is coming from somewhere, and the question that sends readers to this chapter is a fair one: if the app is free, what is the price, and who is paying it.
The short answer is that you are paying, but not in money. You are paying in attention, the minutes and the notice you give the screen, and in the record of what you looked at while you gave it. That attention is gathered up and sold to someone who wants it, which is the advertiser, and the free service is the bait that gathers it. This is not a recent trick invented by technology companies. It is one of the oldest business models in the modern economy, and it was already running, on paper and for a penny, almost two hundred years ago.
A bargain older than the phone
Before 1833, an American daily newspaper was a thing for the well-off. It cost around six cents, roughly a tenth of a day labourer’s wage, and it was sold by yearly subscription paid in advance, commonly eight or ten dollars at a stroke, which put it beyond most households. These papers leaned on subscriptions and on the patronage of political parties and merchants, and they were written for the readers who could afford them. Then a young New York printer named Benjamin Day tried the opposite arrangement. On 3 September 1833 he launched The Sun and priced it at a single cent, a sixth of the going rate. He sold copies not by subscription but on the street, handing them in lots of a hundred to newsboys who hawked them and kept a cut. Others had tried a penny paper before him and failed; Day’s was the first to last, the first commercially successful penny daily in the country, and within a few years it claimed one of the largest circulations in the world.
The penny barely covered the paper and ink. Day was not making his money from readers, and he did not need to, because a large readership was itself the thing he had to sell. What he sold it to was advertising. A merchant who wanted his notice seen by tens of thousands of ordinary New Yorkers could buy space in The Sun, and the cheaper the paper and the wider it spread, the more that space was worth. The reader bought news; the advertiser bought the reader’s eyes; and the paper lived on the second sale, not the first. This is the mechanism that ad-supported free apps now run on, and it is worth being exact about the roles, because a popular slogan blurs them. It is often said that when a service is free, you are the product. The phrase is a modern one, from the age of television and the internet rather than of Day, and as a slogan it compresses too much. What is precisely true is narrower and older: the service is given away to assemble an audience, and it is access to that audience, its attention and its custom, that is sold to a paying third party. You are not so much the product as the raw material from which the product, a delivered audience, is made.
Everything that followed refined that 1833 bargain rather than replacing it. Cheap magazines carried it into the home; commercial radio, from the 1920s, broadcast a free programme paid for by the sponsor whose name was read between songs; television did the same to the evening, giving away the show and selling the break. Each new medium reached more people more cheaply, and each made the same second sale. The device in your pocket is the latest and by far the most exact of them, but it is doing what a penny paper did in Day’s New York.
The one budget that cannot grow
What changed, over that long run, is not the model but its limit. A penny paper competed for a reader who had a few spare minutes in the day; a modern feed competes for the same reader who now has a hundred such services calling at once. The scarce thing, the resource every one of them is really after, is not money but attention, and attention is fixed in a way money is not. Your income can rise; the hours in your day cannot. Whatever you give to one screen is taken from everything else you might have noticed, and there is no way to earn more of it.
The economist Herbert Simon saw this clearly, and early. Writing in 1971, before the personal computer and long before the smartphone, he pointed out that a flood of information is not a pure gain, because information has a cost that is easy to miss. What it consumes is the attention of the people it reaches, and attention, unlike information, does not multiply.
Simon was describing an office drowning in memos, but the observation is exactly the ground the attention market stands on. Every free service is a claim on a budget that cannot be enlarged, and that fixed budget is what makes the competition for it so fierce and the prize worth so much. When there is more information on offer than anyone can possibly attend to, the thing that becomes valuable, and therefore priced and sold, is the attention itself. That is the currency you spend when you open the app. You feel no money leave your account because none does; what leaves is a slice of a stock you can never top up.
A market that changed its clothes but not its size
How large is the business built on that slice? The clearest long view comes from the United States, where advertising spending has been tracked for more than a century. Two facts stand out, and they pull in different directions. The first is that the total has been remarkably steady when measured against the size of the economy. Through boom and slump, and across every change of medium, American advertising has hovered around two per cent of national output for a hundred years, dipping in recessions and during the Second World War, when output surged and advertising was constrained, but showing no lasting upward or downward march. The attention market, measured as a share of everything the country produces, was about as big at the close of the twentieth century as it had been in the age of the newspaper.
The second fact is that within that stable total, the money moved relentlessly from one medium to the next. Advertising began the century with print as its dominant mass medium, in the newspapers and magazines that were then the only way to reach a mass audience, while much of the rest went to direct mail, outdoor, and local selling. Radio took a growing share from the 1930s, television a much larger one from the 1950s, each claiming a rising share of the whole. The total itself showed no lasting rise across the century even as one medium gave way to the next, so what changed was less the size of the pot than which medium held the most of it.
After 2007, where the chart ends, the same migration ran again, faster and in one direction. Internet advertising, only a sliver of the market in the years the chart covers, passed newspapers around 2010, passed broadcast television in 2013, and passed television of every kind in 2016. Newspaper advertising, which had carried the whole business a century earlier, collapsed: American newspaper ad revenue fell from a peak near $49 billion in 2005 to under $10 billion by 2022. By the mid-2020s, on industry estimates, close to four in every five advertising dollars in the United States were spent online. The sums are large and, unlike the share of GDP, they have grown in absolute terms with the economy: advertisers in the United States alone spend on the order of $390 billion a year on estimates for 2024, and worldwide advertising spending passed half a trillion dollars around 2013 and, by 2024 estimates, reached about $1 trillion. The market for your attention is not a curiosity at the edge of the economy. It is one of the largest single things the economy does.
Two sides of one platform
To see why the app is free, and who the free version is really for, it helps to look at the shape of the business rather than its size. A search engine or a social network sits between two very different groups who each want something from the other. On one side are users, who want a service and will gather in large numbers if it is good and costs nothing. On the other are advertisers, who want to reach those users and will pay to do so. The company in the middle is what economists call a two-sided platform, and its central problem, worked out formally by Jean-Charles Rochet and Jean Tirole in 2003, is that it has to get both sides on board at once. Neither side is worth much without the other: advertisers will not pay to reach an empty service, and the service has no way to pay for itself without the advertisers.
The solution such a platform reaches, again and again, is to charge the two sides very differently. It sets a low price, often a zero price, on the side it needs to grow, which is the users, and it recovers the money on the side that will pay, which is the advertisers. What looks from your seat like generosity is a pricing decision: you are the subsidised side of the market, kept free of charge precisely because your presence, and your attention, are what the paying side is buying. The free service is not the product the company sells. It is the cost the company pays to assemble the product, which is a large, well-understood, reachable audience, and to sell it to the advertiser who is the actual customer.
Two features of the market make it far more powerful than Day’s ever was, and both belong to other chapters of this book, so they are named here and followed there rather than reopened. The first is that these platforms grow more useful, and harder to leave, as more people join them, which is the logic of network effects that lets a single service come to dominate a whole category; that mechanism has its home in the volume on frameworks (Volume I, on why monopolies persist). The second is that a platform standing between the two sides is a middleman, and a middleman in the right position can take a large share of the value that passes through, which is why so much of the money the advertiser spends stops at the platform rather than reaching anyone else; the economics of that position is treated in the volume on frameworks as well (Volume I, on why the middleman captures the value). What the newspaper could never do, and the modern platform does as a matter of course, is measure. Day sold a merchant the vague promise of many readers; a platform can tell an advertiser who saw the notice, what they did next, and which version worked better, because it records the attention it is selling. That measurement is what turns a rough audience into a precise product, and it is why the data you generate is gathered as carefully as the attention itself.
A fair trade, or a rigged one
Two readings of this bargain are worth stating plainly, because they lead to very different judgments of it, and the evidence in this chapter does not settle the choice between them. On the first reading, the trade is a fair one, freely made. You are offered a genuinely valuable service, a map that would once have cost money, a way to reach distant friends, a library of video, and you are asked in return only to look at some advertising, which you are free to ignore and free to leave. Advertising, on this view, is not waste: it carries real information, tells you a product exists and where to find it, and lets a new firm reach customers it could never have found otherwise, work whose value the two-sided-market analysis takes seriously. Nobody makes you open the app, and if the attention it costs were not worth the service it gave, you would close it. A great many people, offered the choice, take the deal every day.
On the second reading, the trade is not made on equal terms, because the service is engineered to take more of your attention than you would knowingly give. The same measurement that makes the audience a precise product also lets the platform test, endlessly and at scale, which arrangements of colour, sound, and timing keep you scrolling longest, and to optimise the service toward capturing attention rather than serving you. On this view what is sold is not merely your notice but your behaviour, nudged and shaped, and the budget you are spending is one you cannot feel yourself spending, which makes it easy to overspend. The harms that can follow, to privacy as your record is compiled and traded, and to whether attention captured this way is really yours to give, run past the edge of this chapter, and this chapter does not rule on them.
Both readings describe the same machinery; they differ on whether the exchange it runs is balanced. What this chapter can settle is narrower and factual: that an exchange is taking place at all, that it has a price, and that the price is paid in a currency the word “free” is built to hide.
The apps on my phone are free.
Oversimplified Moderate confidence
In the only sense the word usually carries, a price in money, the claim is simply true: you install the app, you use it, and nothing is charged to your card. That is why it is so natural to say, and why the deal feels like a gift. But “free” quietly widens from “costing no money” to “costing nothing,” and the second is false. The service is given away to assemble an audience, and it is your attention, and the record of it, that is gathered and sold to the advertiser who is the paying customer; the ad-supported free version exists for that sale. You are paying, in a currency, attention and data, that you cannot feel leaving your account and cannot earn more of. The claim is not a lie, which is why the verdict is Oversimplified rather than Backwards: no money does change hands, and whether the attention you trade is worth the service you get is a real question on which reasonable people differ. The confidence is moderate rather than high not for want of evidence, which is direct and plain, but because the claim turns on the slippery word “free,” true of the price in money and false of the real one; that gap between the word and the construct it names is what holds the pip to moderate. What the word conceals is that there is a trade at all, and a price, and a customer who is not you.
Sources
- J.-C. Rochet & J. Tirole, “Platform Competition in Two-Sided Markets,” Journal of the European Economic Association 1(4) (2003) — the two-sided-market model in which a platform subsidises one side, the users, to profit from the other, the advertisers; the mechanism by which the service is free to you precisely because your attention is what is sold.
- H. A. Simon, “Designing Organizations for an Information-Rich World,” in M. Greenberger, ed., Computers, Communications, and the Public Interest (Johns Hopkins Press, 1971) — attention as the scarce resource that information consumes, and so the currency the free service actually costs you.
- Coen Structured Advertising Expenditure Dataset (via D. Galbi, galbithink.org); IAB / PwC Internet Advertising Revenue Report; eMarketer; WARC — the scale of the advertising market, roughly 2 per cent of US output for a century and about $1 trillion worldwide on 2024 estimates, which is the size of the business built on the trade.
- T. Wu, The Attention Merchants (Knopf, 2016) — the narrative history from Benjamin Day’s penny Sun of 1833 through radio and television to the feed, used here for the through line rather than for any figure.
- Confidence is moderate under the rubric: the pip scores the weaker of evidence directness and construct match, and construct match binds. The evidence that the service is paid for by advertising, and that your attention is what is sold, is direct and plain; but the claim turns on the fuzzy word “free,” which is true of the money price and false of the real one, so the verdict is Oversimplified because the claim is half true and hides the rest, not inverted.
Where the history ends
The free app, then, is the newest form of a bargain that a New York printer struck in 1833: give the service away, gather the audience, and sell its attention to the advertiser who will pay. What has changed since Day is not the deal but its reach and its precision. The market for attention is now measured in hundreds of billions of dollars, it can record and price the very thing it sells, and it draws on a budget, your finite attention, that no growth in wealth can enlarge. The service is free of money and costly in the one resource you cannot replace, which is why it can feel at once like a gift and like something you have paid too much for.
That is the last of the institutions this book set out to follow. History has taken exchange apart one piece at a time: the money that clears a trade and the banks that make it, the central banks and public debt that stand behind the money, the corporation and the stock market that pool the capital, the credit score and the financial sector that price and move it, and now the attention market that sells the one thing left to sell when everything else is given away. Each began as an answer to a plain problem, and each, once built, reshaped the world it was built into. Where those institutions come from is the question this volume asked; what they do to the people living among them, whose work they organise, whose countries they divide, and whose crises they cause, is the work of the volumes still to come. The history is complete; the reckoning is not.