That “viral” song, movie, influencer, and celebrity drama you scrolled by recently was likely the result of a stealth marketing campaign.
this is so bleak!!!!
some excerpts:
Joe Lim estimates that 90 percent of what you see on the internet is advertising in disguise, and he should know. For three years, Lim ran a company called Floodify, which at its peak operated 65,000 dummy social-media accounts used to drum up attention on behalf of paying clients. On a typical day, he says, Floodify posted 50,000 videos across TikTok, Instagram, YouTube, and X, all of them designed to pass for the unscripted output of ordinary users.
...
The point of this kind of marketing is that nobody is supposed to notice it. But lately, the machinery has started to show. In March, Jesse Coren and Andrew Spelman, co-founders of the digital music-promotion agency Chaotic Good Projects, gave a live interview to a Billboard reporter at South by Southwest in which they breezily described using sock-puppet accounts to manufacture enthusiasm for artists at every level of the music industry, from major-label pop stars to niche indie acts. Spelman called the practice “trend simulation.” His motto: “Everything on the internet is fake.”
Chaotic Good’s interview went viral the old-fashioned way: by making lots of real people mad. Some were appalled by the cynicism of the company’s pitch, others by its client list, which included indie artists whose popularity fans preferred to imagine had spread organically. ...
...
The primary tactic used by companies like Chaotic Good and Floodify (and many, many others) is known as clipping. A record label — or a movie studio, celebrity talent agency, political campaign, or just some bozo with a video podcast — hires a company to turn a song, trailer, interview, stump speech, or whatever into short, social-media-friendly fragments, either by cutting the clips in-house or by farming them out to a network of freelance clippers. Those clips are then posted by normal-looking accounts: a meme page might serve up a quote about relationships with a new pop song playing behind it; a fan page for a horror movie might cut the scariest 20 seconds from the trailer into a loop and post it twice a day; another account might chop the most entertaining exchange from a three-hour podcast and rebroadcast it to people who would never sit through the entire episode. If enough of these clips rack up enough views fast enough, credulous social-media algorithms interpret the spike as an authentic surge of interest and push the videos to real users, who sometimes generate real engagement, prompting the algorithm to push those videos even further.
Clipping’s origins go back at least to 2022, when the influencer Andrew Tate deployed members of his fan club to post clips of his podcast on social media, causing so many people to wonder who he was and why he was clogging up their feeds that he briefly became one of the most Googled people on earth. Since then, and especially over the past year, clipping has gone professional. Dozens of agencies now offer the service to paying customers. Many operate out of public view, inside members-only communities — which I found were not so hard to join — on platforms like Discord and Whop, where they recruit regular people to do the posting. Each community functions as its own marketplace. An agency announces a new campaign, specifying where the clips should run (usually TikTok, Instagram, and YouTube) and what they pay (usually $1 or $2 per thousand views). Members then have a few days to make and upload as many clips as they can, hoping at least one will go viral.
...
Manipulating algorithms is only part of the goal. The other is fooling humans, particularly the dwindling number of journalists, critics, and other gatekeepers who are still capable of conferring legitimacy by paying attention. Livestreamers were among the first to discover that clipping could make them seem more significant than their real statistics would suggest. Two of the most successful are the Groyper-provocateur Nick Fuentes, who’s been banned by most major platforms but remains artificially overrepresented on TikTok thanks to his clips, and Clavicular, the looksmaxxer who was recently charged with a misdemeanor for shooting an alligator on one of his streams and who credits his golden-ratio handsomeness to smashing himself in the face with a hammer. The New York Times recently profiled both of them as figures of great importance — which they are now in the sense that profiles in the New York Times can occasionally make people seem important — even though the live shows that are ostensibly their flagship product usually draw concurrent audiences in the low-to-mid-five figures, less than a fading cable-news show does during a slow hour. Reporters and editors who get their ideas from their social-media feeds — which is most of them, most of the time — can mistake a paid simulation of public interest for the real thing and then make it real by covering it.
The screenwriter William Goldman once famously wrote that “nobody knows anything,” by which he meant that no one in Hollywood has any idea in advance which movies will turn into hits. That line has become true in ways Goldman could never have imagined when he wrote it, in 1983, when in retrospect we actually knew quite a bit. Back then, at least we knew, after the fact, which culture products had found an audience because we still had trusted metrics for cultural success that were tracked and audited and reported by people whose jobs depended on getting them right.
But none of that applies anymore. Most culture is streamed inside apps that lock their consumption data in a black box, report whichever proprietary metric flatters them most, and refuse to be audited or to convert their stats into anything that can be compared with those of any competitor. Even the artists whose work all this machinery is supposedly serving no longer have a reliable way to know what real audiences actually want, since whatever feedback reaches them may already have passed through the same apparatus built to distort that feedback in the first place. In that vacuum, fakery thrives. When nobody knows what’s actually popular, the appearance of popularity matters more than popularity itself.
...
To be fair, social-media platforms brought this on themselves. The great pivot to video of the past decade was sold to the world as a simple accommodation to user behavior: People didn’t want to read anymore; they wanted to watch. But that was only partially true. Platforms wanted video even more because they could charge more for video ads than they could for the banner ads that used to fund beautiful websites like the one you’re reading right now. So those platforms repaved most of the internet into surfaces that could host video ads, then incentivized users and publishers to roll their cameras. The pivot worked. Meta’s revenue has grown more than tenfold since the mid-2010s, and TikTok’s global revenue is expected to top $30 billion this year. But the same shift that made these platforms rich also created a monster that they couldn’t control. Our feeds now require an almost-infinite supply of short-form video, and clipping helps provide it, but it presents a moderation problem with no good solution. Clipping is hard to trace, hard to tell apart from ordinary posting, and hard to eliminate without killing off much of the engagement that these platforms have come to depend on.
...
The thing that most bothered people about Chaotic Good Projects wasn’t clipping but a related service the company calls “narrative campaigns.” Clipping just puts an artist in front of more eyeballs; narrative campaigns tell those eyeballs what they’re seeing. Chaotic Good co-founder Jesse Coren explained the idea to Billboard at South by Southwest. “A lot of what we do on the narrative side is controlling the discourse,” he said. “Most people see a video or see something about an album that came out and it’s like the first thing that they see, or that first comment that they see, is their opinion even when they haven’t heard the whole album.” In other words, in a world drowning in information, nobody has the time to form an opinion from scratch anymore, so they check captions, comments, and quote tweets to see what people who seem like them have to say. And if everybody is outsourcing their first impressions to the crowd, why not just manufacture the crowd? Co-founder Andrew Spelman gave the example of a musical performance on Saturday Night Live: “The second SNL drops at midnight, you should post a hundred times saying that was the best performance of the year.”
...
According to Keith Presley, the co-founder of Gudea, narrative campaigns are far more common and effective than the public knows. Gudea’s main business is using AI to detect coordinated activity on social media, and Presley says he and his team have observed these tactics being used across a wide range of subjects. “We’ve seen this used for stock manipulation, to promote skin-care brands, to shape conversations around AI, you name it,” he says. Many of Gudea’s clients are large companies looking to defend themselves against what he calls “corporate espionage” — paid narrative campaigns run by smaller competitors designed to damage a larger brand’s reputation just enough to make its customers defect.
...
The same scheme works on people. The dominant technique now isn’t so much inventing a controversy from nothing as choosing which real minor outrage to fuel. Because you can usually find someone on the internet mad about almost anything, the job is mostly to choose which objection to amplify and how loudly. In one case, Gudea tracked a campaign promoting a rumor that the cover of Taylor Swift’s 2025 album The Life of a Showgirl contained Nazi symbolism that started in the fringes of X and Telegram before being amplified by what Gudea calls “non-typical accounts,” until regular users picked the rumor up and ran with it. Who would do such a thing? “Well, who’s trying to take down Taylor Swift so they can be the next Taylor Swift?” said Presley. (He did not offer a more specific guess.) 👀
...
What all of this amounts to isn’t just one problem but a stack of them, each feeding the next. Most people now encounter the world through algorithmic feeds built to warp reality, on platforms with every commercial incentive to keep users scrolling and very little incentive to distinguish genuine interest from astroturfed imitations. Into those feeds flows an unprecedented amount of undisclosed advertising engineered to resemble the improvised enthusiasm of human strangers. The platforms reward it with reach; traditional media picks it up and validates it. Meanwhile, as trust in journalism collapses and most of the actual reporting disappears behind paywalls, readers head straight for the comment sections, which seem more like the voice of the people than anything written by a reporter — except many of those commenters may not be people at all.
The good news is that this will all be over soon, according to Lim, because something worse is coming to replace it. He recently shut down Floodify after trying to scale too fast and falling behind on deliverables. At one point, the company accidentally posted the same video to 7,000 accounts, which got them all banned. But he wasn’t discouraged. When we last spoke, he was building a new company and thinking even further ahead. “All of this nonsense is only going to last three to five more years, because in the future, people will stop trusting what they see on social media.” By then, the job will have moved one layer up. “You’ll have to start distributing your content toward AI agents and then they’ll teach humans what they want.” 😱
