Most people in the industry were convinced that their work was moral because it increased consumer choice and therefore freedom. New technologies were evidence of progress and therefore innately good. And any criticism of the industry’s practices or motives therefore threatened freedom and progress.

When a tech company captures an audience, it gets more than the opportunity to sell products and ideas. It also harvests the discretely quantified and collated bits of individual user data that people hand over, wittingly and unwittingly, as they stare at their computer and smartphone screens. As valuable as this information is for what it reveals about individual consumer habits and preferences, it’s even more precious in the aggregate, as so-called big data, which can be used to predict political shifts, market trends, and even the public mood. Who knows, wins, as the old military adage goes—and this is equally true in the world of business. Watch the video, click the link, fill out the form—this is the labor that tech companies turn into profits. The people who carry out this labor consider themselves customers, but they are also uncompensated workers. The process whereby eyeballs get turned into money is mysterious, but not totally opaque—just discouragingly complicated and boring.

Fraud was the hot topic that year, because digital ad buyers were starting to wise up. More than two decades after the arrival of the commercialized Web, a trade group finally funded a proper scientific study on the problem of online ad fraud. The study found, among other things, that marketers were losing $6.3 billion a year to various forms of fraud, much of it staged by organized criminal networks. In outline, such scams allow fraudsters to siphon the fat from corporate ad budgets by employing bots that pose as genuine consumers to click on ads. The crooks are able to grab a piece of the money advertisers are paying out because online publishers—that is, people who run websites on which the ads appear—receive a cut of the money paid to online ad sellers by the companies that buy ads. The crooks are even able to redirect ad revenue from legitimate publishers to their own fraudulent sites through a process known as “injection,” or to generate bogus clicks by hijacking users’ browsers with automated hacking tools. Several experts at the conference told me the study lowballed its multibillion-dollar estimate of industry-wide fraud losses and that the real figure was multiples higher.

In other words, online advertising—the basis for the attention economy that fueled all speculative investment in digital media, from giants like Google on down to low-rent email marketers—was a racket. In the case of Google, an ad buyer will fill out a form saying what search keywords they’d like to associate themselves with, so that when a Google user types in, say, “soap,” they might see an ad for Irish Spring. On Facebook, it would work a little differently. There, ad buyers are able to specify a certain demographic they want to reach—say, expectant mothers with household incomes of $80,000 a year and up, or people with bachelor’s degrees who drive secondhand cars in the Cleveland, Ohio, metro area. This kind of targeting is the core promise of digital advertising. But during the course of the Ad:Tech talks, I came to see that the promise was a sham. The old knock on print and broadcast advertising was that half of ad budgets were wasted, but no one knew which half—the ads went out to everyone. Online ad targeting was supposed to change that by essentially surveilling users and letting advertisers see who actually viewed their ad and did or didn’t buy their product as a result. In reality, though, the new data collection tools didn’t work nearly so well as was promised. A full half of ad budgets was still getting flushed down the toilet.

“It’s all about getting the chart that goes up,” a disaffected social media marketing expert told me over drinks. “There’s a whole industry devoted to making charts that go up.” To illustrate his point, he pointed me to the Guardian newspaper’s now-defunct “partner zones” program, which afforded large institutional advertisers the opportunity to pay massive sums to the newspaper in exchange for the right to post promotional “news” stories on its website—a form of “sponsored content,” in the industry’s parlance. To create the all-important “chart that goes up,” the clients would then pay Facebook to generate traffic to their advertorials. Ostensibly this traffic came through “organically” promoted links targeting genuine potential customers who are so enchanted by the serendipitous appearance of an online advertorial that speaks to their personal desires that they make a conscious choice to click, read, like, and share—or so the story goes. However, the expert, who was a friend of mine, had noticed that a suspiciously high percentage of the paid traffic came from far-flung, low-wage countries such as Bhutan. So the new model supporting digital media was for floundering corporations to pay to place stories about how awesome they were, which publishers would then promote by buying phantom readers. “Then the advertisers can go to the boss and say, ‘Look, we got an article in the Guardian,’” my friend the marketing cynic said. Those phony measures of success supplied fodder for still more charts that went up—these ones for internal consumption, and used to justify the marketing department budget to higher-ups.