On Thursday 3 February Meta Platforms, the company that owns Facebook, Instagram and WhatsApp, lost more than a quarter of a trillion dollars in market value — the largest one-day drop in history. Mark Zuckerberg, whose personal wealth fell by more than $30bn in a single day, appeared red-eyed and shaken (he claimed to have scratched his cornea) in a call with employees, whom he urged to concentrate on new products.
The plunge was caused by an earnings report in which Facebook acknowledged that its daily users had begun to decline. Zuckerberg’s story is that this doesn’t matter in the long run, because Meta will replace the social media platforms of today with the internet of tomorrow: the metaverse. But for the technology author and researcher Tim Hwang, it points to a deeper crisis in the way internet companies make money — one that could have serious implications not just for Facebook, but for all of the tech giants and for the financial markets that have ballooned on their progress for more than a decade.
The problem goes back to the early years of the internet. In April 1998 Larry Page and Sergey Brin, then Stanford University PhD candidates, wrote a paper presenting Google, a new prototype search engine they had built “to crawl and index the Web”. One crucial point remained unsolved: where the money would come from. “Advertising income often provides an incentive to provide poor quality search results,” Page and Brin wrote in an appendix. In the new meritocracy of the web consumers would weed out technology that arrived with “mixed incentives” because it wouldn’t be as useful.
“When you look at the original pitch decks for Google, they actually didn’t think that they were going to make much money from advertising at all,” Hwang, who is a former Google employee, tells me. “They said, ‘Oh, we’re going to make about a third of our money from licensing our search algorithm, and we’re going to maybe make 10 to 15 per cent from advertising.’”
It quickly became apparent, however, that the web would only grow at massive, world-changing speed if websites were free to use — and that the only people who were going to pay for that growth would be advertisers. This gave Google a problem to solve: an individual website could simply put a car ad on a page about cars, as newspapers and magazines had done for centuries. By 2005, however, Google was serving more than 100 billion searches a year on more than 64 million websites. The question was, “How do you deliver a relevant ad to someone, when on a given day someone might search for anything on Google?”
When Facebook arrived in the mid-Noughties it had a similar problem, says Hwang. “How do you deliver millions, if not billions, of ads to lots and lots of different people looking for different things on the web?”
The answer was through an increasingly complex series of ad networks, which, as Hwang explains in his book Subprime Attention Crisis, commoditised attention into tradable units (the CPM, or cost per impression, and the CPC, or cost per click) and enabled advertisers to bid for the right to display an ad to an individual user at a given time. These auctions take place in tiny fractions of a second, every time you open a page with a network ad. This economy of programmatic advertising “created these incredibly wealthy companies, and also shaped the web as we know it”.
Much has been written about the incentives this system has created for disinformation, populism and polarisation. For Hwang, however, the real question is: “What if this waterfall of money that we’ve discovered on the internet is actually not sustainable?”
Hwang maintains there are good reasons to suspect that it isn’t. “The closer that you look at it, the more it appears that there’s large segments of this marketplace which are faulty — in effect, kind of a scam.”
Swathes of internet traffic are useless to advertisers. More than a third of web-page views are from machines, and roughly a third of human web traffic is ad-free, thanks to ad blocking. The criteria for what the industry considers an “impression” are fantastically accommodating — an ad is considered viewable if half of it has been on screen for one second — but studies suggest that less than half of ads meet them.
“We’ve been sold a vision,” Hwang says, “for a very long time that this system, on a fundamental level, really works.” But the effectiveness of programmatic advertising, and the amount advertisers will pay for an impression or a click, has been falling, forcing Big Tech to undermine its own products as they “force more and more ads into an experience, or into a social media channel, in order to make the same amount of money… and that has the perverse effect of squeezing people away from those platforms”.
Hwang thinks this is why Facebook and Google place such value on ceaseless innovation: because they have no choice but to ruin the experiences they create for the people that use them, just as Brin and Page predicted in 1998. What Meta’s earnings report really showed was that Facebook is driving users away from its most profitable product.
Facebook and Google also face rising conflict with the privacy policies of both Apple and the EU because of the ways users are tracked and targeted. It was suggested last week that Facebook and Instagram might have to withdraw from Europe, a prospect that policymakers appeared to relish. “We would live very well without Facebook,” said the French finance minister, Bruno Le Maire.
For Hwang, the real danger of these policies is not that they will stop ads from working, but that companies will lose access to tracking data “and it turns out the ads are pretty much as effective as they always were”. If Big Tech forfeits its powers to track and target, what should companies do — lose access to hundreds of millions of affluent EU consumers, or admit that the technology behind the multi-trillion-dollar industry of surveillance capitalism is about as effective as an ad in the local paper? “That,” Hwang said, “might really pop the bubble.”
A second dotcom crash would not be confined to a handful of companies. Big Tech has been the driving force behind the longest bull run in history, a ten-year boom in innovation. By 2020 the five largest companies in the US were tech companies — Facebook, Apple, Tesla, Amazon and Microsoft — and accounted for 37 per cent of stock market returns.
Big Tech’s rising tide has not only created a long boom in stocks but innovation across the global economy. “The advertising ecosystem has ended up subsidising a lot of things that we don’t really think about as connected,” from the growing AI industry to entertainment and journalism. “We have delegated so much basic R&D to these companies. For example, the progress of self-driving cars is almost entirely funded through advertising.”
Forget the dotcom bubble. Could this be as big as the 2008 financial crash? Hwang accepts that Big Tech, inflated as it is, doesn’t compare in size to the mortgage market that brought down the global economy 14 years ago. But he feels the comparison is still valid: something obscure and complicated is about to become systemically important. In 2008, he says, “you had this financial asset… that we didn’t really think much about… but as soon as that market downturns it had these ripple effects throughout the economy”.