"My Pages Amuse Only When They Are Free"
"My Pages Amuse Only When They Are Free"
A review of Tim Hwang's The Subprime Attention Crisis, and what it might mean for online writing
For a Good Deal, Seek Out Secundus
Martial had a problem. Poetry has always been a bad business, but it was especially bad in the ancient world. "Are you really that Martial," a passing fan once asked him, "whose mischief and jests are known to all?" Martial nodded. "Why then is your cloak so ragged?"
"Because I am a bad poet," he said.1
This wasn’t quite true. Martial was a great poet. His satires were adored wherever wit and gossip were valued, which was everywhere in ancient Rome. The problem was that, like most great poets, Martial was a lousy businessman. He was always running out of money, falling out with patrons, and losing out on investments. If he could actually make money selling his books, he’d be fine, but ancient Rome had no royalties for artists and no publishing industry to manufacture and distribute books at scale.
Irene Vallejo, writing about Martial and the Roman literary market in Papyrus: The Invention of Books in the Ancient World, proposes a curious solution to Martial’s problem. Literary historians like Vallejo have always valued Martial because his complaints about Roman bookstores have furnished us with valuable evidence about things like the price of books, the role of booksellers in Rome, and the types of shops they ran. And in several of Martial’s books, he goes so far as to list specific bookstores, even giving directions.
"So you won't fail to know where I am for sale, or wander aimelssly all over town," Martial writes in one dedication, "follow these directions: seek out Secundus, the freedman of learned Lucensis, behind the temple of Peace and the Forum of Pallas."
We know that Martial was constantly broke; we know that giving compliments of any kind to any person was hard; why, then, did he go out of his way to put in a kind word for the sellers of his books? Is it possible, Vallejo wonders, that Martial earned money through paid product placement? This would make him the earliest ad-supported writer that we know of, an achievement far more important to the history of literature than his satires.
The Engine of the Internet
There are only two ways to make money from writing: you can either get a lot of people to buy copies of your work, or you can get a rich patron to sponsor you. This was true in Martial’s time two thousand years ago and it’s true in our time. It’s also overwhelmingly true that the patronage model is more durable and more popular by far. Ever since the New York newspaper The Sun put out its first ad-subsidized issue in 1833, dramatically undercutting the newsstand cost of its competitors, most of the money made in the writing business has come from advertising. TV and radio took this even further, giving out content for free, depending entirely on ad revenue to keep the lights on.
The internet has only supercharged this model of content distribution. If you look at the ten most popular websites in the United States, two of them are ad-subsidized markets (Amazon and Apple) and seven of them (Google, YouTube, Facebook, Reddit, PornHub, Twitter, Yahoo) are supported almost entirely by advertisements. Billions of internet users around the world get news, entertainment, social media, pornography, and more without paying a cent. Whole industries are supported by the largesse of advertisers, and it all comes down to the magic of programmatic advertising.
Most of the online ads that you see are, essentially, an auction. Every time you open a web page, an auction between the host and dozens of advertisers is conducted at, quite literally, the blink of an eye. The thicker your data profile—the accumulated dust of your internet travels, suggesting all kinds of probabilities about where you’ve been and what kinds of products you like—the more “relevant” the ad supposedly is, and the more likely it will result in a “click-through,” taking you to the advertiser for a sale (or sometimes, just another ad-supported website).
Programmatic advertising, or adtech, is the engine of the internet as we know it. Ads support more artists, writers, journalists, filmmakers, musicians, and designers than all the patronage of Martial’s Rome ten times over. By renting out a little corner of our screens, we pay for almost everything.
Without it, the internet would be a ghost town: sure, the pipes all work, but there’s no search, no social media, no porn, no Gmail, very little news. Amazon prices would be higher, and smartphone apps would actually cost money. The entire internet, besides maybe monastic little Wikipedia, hiding away in its cloister, would be slower, emptier, and much more expensive. The damage to our creative economy, at least in the short term, would be catastrophic.
Adtech, as you may have seen, isn’t having a good year. Google’s revenues are slipping. Snapchat lost a third of its value in a month. Mark Zuckerberg is so freaked out by the dismal future of ad revenue for Facebook that wasting billions on a VR Cringe Xanadu seems like a better alternative than business as usual. After his shotgun wedding to Twitter, Elon Musk is now kicking out personnel and raising prices to cover the fact that Twitter has almost never turned a profit from advertising. Across the adtech market, advertisers are increasingly wary of spending their ad budgets on, well, advertising.
Tim Hwang, tech researcher and author of The Subprime Attention Crisis: Advertising and the Time Bomb at the Heart of the Internet, saw this coming back in 2020. His short book makes a well-researched, rigorously argued case for what is, when you stop and think about it, blindingly obvious: adtech is bunk. It always has been. I don’t click on ads, you don’t click on ads, pretty much nobody clicks on ads, and never really has.
The Case Against AdTech
The research here, as cited by Hwang, is robust. While the earliest online banner ads in 1994 had an astonishing click-through rate of 44%, by 2018 it was down to about 0.46%. That's a decline, in two decades, to about one-hundredth of their former effectiveness. And even then, about half of those clicks are thought to be "fat-finger clicks," accidental clicks on tiny touch screens.
Then there’s fraud: depending on who Hwang asks, he finds that between one-tenth and one-third of all ad click-throughs are estimated to be fraudulent, the work of sketchy adtech companies juicing their numbers through bots or armies of "click-farmers" in developing countries. Other companies practice “domain-spoofing,” in which they sell premium-looking ad space that turns out to be in a dead zone. It's as though you paid for a billboard on Times Square, but instead got your ad stenciled on the back of a trash can in Poughkeepsie.
But it gets worse. Even when you factor in the fraud, the fat-finger clicks, and the fact that a tiny minority of internet users are responsible for a vast majority of click-throughs, there is still no proof that programmatic advertisements actually work. One of the best independent studies Hwang cites found that user-targeting programmatic ads were 500% more expensive than non-programmatic ones, but only raised 4% more revenue. Even when adtech works exactly as it’s supposed to, it’s hardly better than a generic digital billboard.
Is There an Advertising Bubble?
How could this flim-flam advertising possibly sustain a multi-billion dollar industry and power the internet as we know it? It's always possible, given how much data Google and Facebook have at their fingertips, that they know something we don't, and are using this proprietary data to hedge their bets.
But Hwang has another theory, as the title of his book suggests: the adtech market is subprime. Just like mortgages in the 2008 recession, adtech is an overvalued asset whose poor value is masked by complexity, obscurity, a few bad-faith bullshitters who know better, and a lot of naïve investors suckered into the hype of better advertisements. Hwang thinks that whole swathes of the market are built on financial quicksand. Much of the investments made will never be recouped, the expected growth will never happen, and the assets investors do hold will turn to ashes.
Is there actually a bubble? As a non-expert, it's hard for me to say. Right now, we seem to be in more of a market correction than a bubble collapse, with Google, Facebook, et al. simply losing a bit of their stock value while investors grumble. But if there really is a bubble, you won't need an obscure literary blogger to point that out: you'll know it when Facebook starts charging you to message your friends, or Google hangs a tip jar off the Daily Doodle. And look, Twitter is now charging $8 a month for verified accounts!
Controlled Demolition
More importantly, and putting my literary blogger hat back on, a real adtech collapse would mean the sudden and painful withdrawal of much writing from the public sphere. Paywalls would proliferate, many publications would collapse entirely, and Google AdSense-powered chum factories would start to clog up the pipes of the internet like cicadas after a die-off. In the short term, many thousands of careers in journalism and publishing would disappear forever. Less visibly, but no less important, many of the indirect beneficiaries of adtech profits, like Google Scholar or Goodreads, might get shuttered as their parent companies retreat. As odious as the current internet might be, a bursting adtech bubble would be a disaster.
Instead, Hwang advocates for "controlled demolition," encouraging users and business to migrate away from programmatic advertising before the market grows large enough to threaten us all with its collapse. Use ad-blockers, avoid ad-powered websites when possible, purchase or subscribe to writing that you truly value. For his part, Hwang has put his chips in with Substack, where he now works as general counsel.
The two funding models for writing will always be with us. There will always be places in the market where pure sales are enough, or where patrons can be found, or where a hybrid model can thrive. But if the early 21st century has seen a strong swing towards advertising patronage in the arts, the pendulum may be shifting back in the other direction.
Martial, possible pioneer of poetic advertisement, wrote in a way that maximized his exposure to audiences, allowing him to sell his works cheaply. "My pages amuse only when they are free," he wrote. It led to a fickle fanbase, broad but shallow, and he was the first to admit it in his many complaints about his readers and patrons. It might be time, as Alan Jacobs writes, for writers to focus less on chasing more eyeballs, and turn more to selling dear words to few friends, rather than cheap ones to strangers.
From the Archives:
I’m going to get back to using this space after posts to share older writing, for the benefit of new subscribers. Here’s an essay from last year on the neglected bastard child of literature, the movie tie-in novel.
All translations here are from Martial’s Epigrams: A Selection, translated from the Latin by Gary Wills. They are quoted in Irene Vallejo’s Papyrus: The Invention of Books in the Ancient World, published by Knopf Doubleday in 2022.