(Bloomberg) — As billions of consumers binge-watched TV shows during the pandemic, an insidious form of fraud was playing out right under their noses.
Marketers spend millions of dollars to reach audiences glued to internet-connected TVs, but some of that money is ending up in the pockets of scam artists, who’ve devised a complex and hard-to-detect shell game that mixes up where ads appear and who makes money from their placement. This digital switcheroo diverts ads meant for big, high-definition television screens to less expensive platforms, such as smartphones, with thieves pocketing the price difference.
Joe Barone, a managing partner at ad agency GroupM, saw the scam up close last year in a marketing campaign for a client who wanted their messages to show up on smart TVs. But instead, he later learned, the ad was being viewed on a card-game app that’s typically played on handheld devices. “A huge percentage of their impressions were running on Solitaire,” said Barone. “Very few people are playing Solitaire on a 50-inch big-screen TV.”
Connected TVs are prime marketing real estate because the big screen captures the attention of viewers more effectively than a smartphone with a smaller screen. Some ad spots have to be viewed before the show you want to watch appears, which almost guarantees your attention. The TVs generate growing amounts of data on audience behavior to allow better ad targeting, bringing higher prices, and fatter rewards for those who can cheat the system.
Smart TV advertising grew 7.4% last year to $12 billion and it’s now the biggest area for fraud, according to GroupM, an ad-planning and buying division of London-based WPP Plc.
The spoofing is “akin to getting a pizza delivered to a posh neighborhood, but not necessarily directly to a valued customer’s door,” said Scott Thomson, the chief operating officer of ad verification company Method Media Intelligence. “The supplier of that pizza can only track when a neighborhood guard got the pizza, not when the customer got it.”
All advertisers using machines and algorithms to buy space on connected TVs are affected. Cameron Thorn, a director at Method Media Intelligence, estimated that from 5% to 80% of programmatic TV inventory available in exchanges on a given day is fraudulent. Dollars disappear from the system, and executives fear that many deliver no return on investment.
The number of connected-TV fraud impressions detected by DoubleVerify Inc., a company that authenticates online advertising, more than tripled last year and now numbers in the hundreds of millions, the company said. It found more than 500,000 fake connected TV devices per day.
“We’ve never identified so many different attacks in terms of ad fraud on such a narrow area of media,” said Roy Rosenfeld, a senior vice president of product management at DoubleVerify.
It’s not a victimless crime. Brands waste media budgets and trust in the system erodes, so premium publishers are forced to charge less for content they spent billions of dollars to produce.
The perpetrators are a more varied bunch than your stereotypical cyber criminal. They’ve been spotted everywhere from China to Eastern Europe and the U.S., and it isn’t always a hacker hunched in a dark room.
“Sometimes it’s also a media company that wanted to set up connected TV apps with some unique content, and it didn’t work out, but they still have to make ends meet,” said Rosenfeld. “Quickly it starts to deteriorate into something that’s a little bit gray in the beginning, and turns out being ad fraud after a while.”
Digital ad buying involves six or seven steps where dozens of actors solicit, make, sort and aggregate bids and fire back information about who viewed the ad and for how long. This byzantine supply chain obscures the link between the ultimate buyer and seller, opening a window for fraudsters to game the system.
Also enabling this kind of fraud is the difficulty of figuring out who’s watching what on TV. Counting views on mobile phones is straightforward, since there are generally only two mobile operating systems, Apple Inc.’s IOS or Alphabet Inc.’s Android. By contrast, there are as many as 50 different ways to measure views across TV manufacturers and platforms, according to GroupM. That means it’s easier to hide fake ad requests.
Fraudsters can also misrepresent the identity and location of the viewer, or overstate the number of real views, known in the industry as impressions. One method is to hijack or imitate a process called server-side ad insertion, where commercials are stitched into the show you watch. First you work out the digital tags signifying the TV ad views, then you copy the internet addresses of legitimate users before submitting fake ad requests. Those ads aren’t seen at all.
In the case described by Barone at GroupM, the scammer used a server to send back a false signal that appeared to offer a spot for TV ads. Ad buyers failed at first to notice the deceit, the ad ran on handsets and, amid a succession of payments between a string of intermediaries, the scammer quietly took a cut.
Once scammers find a method that works, they scale up quickly. When digital fraud prevention firm White Ops Inc. discovered a scheme called IceBucket, “almost 30% of all the connected TV traffic we were seeing was actually fraudulent,” said Dimitris Theodorakis, an engineer at the company.
One scam dubbed “ParrotTerra” that was discovered in January was faking 3.7 million devices per day, and was on track to steal as much as $50 million from advertisers and publishers, according to DoubleVerify. That would have made it three times the size of the next biggest they’d seen, StreamScam, identified weeks earlier, which raked in $14.5 million.
The problem is far more complex than an earlier wave of digital ad fraud that arose more than a decade ago, when criminals generated billions of false clicks on online ads. A fake click can be compared to a real one. But ad views are passive, making it harder to spot a phantom one.
Fraud detectors can spot the scams by scanning information, such as where an ad is being viewed or the number of smart TVs online in one home, for unusual signals. If one home is showing 20 connected TVs, the chances are it’s a fraud.
Media companies are also teaming up to try to simplify and standardize the ad signals given off by connected TV platforms and make it harder to mimic tags on metrics such as viewing data.
Anti-fraud company Integral Ad Science Inc. is announcing a new connected TV strategy soon, Chief Executive Officer Lisa Utzschneider told Bloomberg. IAS recently hired a founder of streaming service Hulu, Tom Sharma, and touted his connected-TV expertise.
It’s impossible to stamp out fraud entirely, so the goal is to make it so complex and expensive for criminals to impersonate or inject fake signals that it’s simply not worth it, Theodorakis said.
“Ad fraud is an eternal and extraordinarily technical game of whac-a-mole,” said Pete Kim, CEO and co-founder of digital ad agency MightyHive. “But the stakes have gotten higher.”
For more articles like this, please visit us at bloomberg.com
Subscribe now to stay ahead with the most trusted business news source.
©2021 Bloomberg L.P.