—How, and why, BuzzFeed News media editor Craig Silverman keeps digging up must-read stories about ad fraud—
Craig Silverman, the Toronto-based media editor for BuzzFeed News, is often credited with being one of the people who coined the term “fake news.”
Long before it became ubiquitous in pop culture, Silverman was reporting on the proliferation of overly sensational, extremist online content that metastasized into intentionally misleading and false content—which he started calling “fake news” five years ago.
How does he feel about that credit today?
“It’s complicated. I don’t love it,” he says. “What I think is the accurate way to say it is, I definitely helped popularize the term ‘fake news.’ I started really using it in 2014 as part of a research project, but ultimately who made ‘fake news’ a global term? It was Donald Trump, there’s no question about that.”
For the last two-plus years, a lot of Silverman’s investigative reporting has focused on how the digital advertising industry is repeatedly being manipulated and exploited by criminals who not only pollute the digital ecosystem with more fake content, but have done so in order to steal billions from marketers around the world each year.
Silverman’s stories should matter to everyone in marketing, advertising and media today. Marketers continue to waste billions on bad digital advertising—either directly on fraudulent impressions, or into a system that helps them spend directly on fraudulent impressions—while virtually all other industry players, from creative to traditional legitimate media, feel increasing pressure to do more with less.
Ahead of his appearance at Ensemble’s “The Future of Truth” event in Toronto Wednesday, The Message spoke with Silverman about his early reporting on the rise of misinformation and fake news, and how it led to his extensive reporting on ad fraud.
“Extreme wins”
Early in the conversation, I ask about the use of the term “fake news” today.
“I try to not use it as much as possible,” he says. “But the conundrum is that it’s now a term that is really ubiquitous, known not just in English language countries but countries around the world—they use the term fake news.
“On the one hand it’s shorthand and it’s popular, which normally would be a very helpful thing for communicating, but it is so weaponized and so loaded, it means different things to different people, that it has become almost meaningless.”
Silverman’s roots in reporting on bad information online date back to 2004, when he started a blog called Regret the Error about media errors and corrections. As the social web took off, he grew increasingly interested in how journalism responded to the growing phenomenon of misinformation being shared and spread online, and from there the explosion of fake content, particularly in the lead up to the U.S. election.
By 2016, he was breaking stories about the now infamous Macedonian teens who were creating hyper-partisan pro-Trump content—not because they were fans of the then presidential candidate, but because they found it was the easiest way to generate traffic they could monetize. It’s taken as given now, but Silverman was among the first media reporters to identify the fundamental problem of the modern, digitally based attention economy:
“In a world where we have an abundance of information, and where people are kind of having it wash over them all the time, it’s a basic question of how do you stand out?” he says. “How do you get someone to pay attention to what you’re doing? And in really short form: extreme wins in that environment. Extreme wins and Trump rode that to the White House.”
But we’ve known for a long time that people like sensational media coverage, I say. Well before the internet transformed news media into what it is today, an axiom of journalism was “If it bleeds, it leads.”
That is true, he says, both in terms of what people wanted and what the media was willing to supply. But those tendencies have been “amplified to an incredible extent.” People are drawn toward sensationalist headlines, and information that aligns with their beliefs, but those human tendencies have been “magnified and weaponized in this media environment and they’ve also been automated and put in a massive scale,” he says.
So part of the problem is human nature, but part of it is also the complex systems that “are deciding who wins,” he says.
The systems are the algorithm-based platforms—upon which most of the modern digital marketing ecosystem is based—that are constantly refined and updated to maximize attention above all else because more attention means more ad sales. The systems do not care if content is good or bad. They only care about how much a person engages with it, and the reality is that sensational content does better—even if it is not true.
Connecting fakes news to ad fraud
Like a lot of journalists, Silverman says he didn’t think much about how media companies actually made money, and hadn’t made the connection between fake news and the larger implications for media business models.
The turning point was the late 2016 New York Times article about MethBot, a Russian based fraud ring that stole as much as $5 million a day from from advertisers paying to show their video ads to more than 500,000 fake internet users on 250,000 fake websites.
“It was so squarely in my beat area, it was a wake up call to me to say I don’t know anything about what the hell is going on with this whole world of ad fraud,” says Silverman.
He spent much of 2017 educating himself, reaching out to experts like ad fraud investigator Augustine Fou and occasionally talking with someone working with an DSP or SSP.
“It blew my mind how much money was involved… how much money was was just being stolen. It’s not waste. It’s not, ‘Well, this was our strategy, we ran the ads here, it just didn’t resonate with our audience, so that just wasn’t a successful campaign.’
“This is just like an absolute criminal stealing the money of the brand, and I couldn’t believe how much money was being stolen and I couldn’t believe how ingrained in the ecosystem. It actually made me really, really angry and outraged,” he says.
“On one level, because I knew that at least some of that money could help support journalism. Legitimate media, not just journalism, but all kinds of legitimate media.”
By late 2017, Silverman was posting deeply reported, investigative articles that carefully detailed the machinations of ad fraud schemes and the players involved (this is the one that caught my eye, and not because it involves an actress who played Bob Saget’s girlfriend on Entourage).
The problem of ad fraud today is complex, with clever criminals figuring out new ways to defraud advertisers. Estimates of the costs range from about $5.8 billion by the ANA and White Ops, to $23 billion by cybersecurity firm Cheq, and a staggering $42 billion by Juniper Research. In 2016, the World Federation of Advertisers warned that without meaningful change, ad fraud could become a $50 billion a year problem by 2025. Last year, Adobe concluded that 28% of all web traffic could be non-human (i.e. fake).
It’s not easy to find fraud breakdowns by market, but earlier this year, GroupM estimated it at $113 million in Canada. Canada is about 1.8% of the global market, so some back-of-the-envelope math based on the WFA forecast paints a bleaker picture: Canadian advertisers could be wasting as much as $900 million per year on ad fraud by 2025 if things don’t improve.
To be sure, industry associations are pushing for changes intended to stop ad fraud—such as IAB’s ads.txt intiative—and the ANA pointed out that it’s $5.8 billion estimate was down from $6.5 billion in 2017.
What is the incentive to change
Earlier this month, Silverman reported how the website City of Edmonton News—part of a small network of similar “news” sites—was registering far more traffic than sites belonging to properties like the Edmonton Journal and Edmonton Sun. At its peak, the network of fake sites “may have earned more revenue from programmatic ads than the leading news outlets in these cities,” he wrote.
Part of the problem is that the criminals keep finding new ways to game the system, but there is also a lack of motivation by those within the system to do anything about it, he says.
“What are the incentives? The incentive is to let this roll, because the only people who are losing money are brands, everybody else makes out. And the brands, if they aren’t applying the pressure, then everyone’s just happy to take the money.”
So why aren’t more brands applying pressure?
CMOs don’t want to look bad in front of the CEO, he says. “I talked to a fortune 50 CMO and I showed them that their sponsored content was getting completely fraudulent views and their partner had bought these views. They thanked me, but then publicly they didn’t actually want to acknowledge it.”
In defence of agencies, I say that many turned to programmatic at a time when they felt extreme pressure from their clients to cut costs and be more efficient.
“I have some sympathy for the agencies and how much they are getting squeezed by clients who are now questioning every line item on a bill, but nobody’s questioning the sites or the apps where their stuff ran,” he says. “I think if you’re going to grind the agency on the billing, but then not, together, grind on where the ads showed up—because there’s money to be saved there—I think that’s a short sighted solution.”
Brand leaders need to be more suspicious of the numbers, he says. In this system, if a brand gets a report that their campaign hit all the metrics, that should be a red flag, he argues.
“I think the brands and the agencies need to be hard asses about figuring out where the spend went, where the ads ran, interrogating all of that—asking tons of questions and making people worried that they are going to get found out if they try to pull a fast one,” he says.
“At the end of the day, there are a lot of people who need to lose jobs, there are a lot of companies that should not exist,” he says.
The companies he’s referring to are the many intermediaries that have sprung up to drive programmatic advertising, each taking a part of the marketer spend. Not that long ago, the Lumascape was held up as a sign of the dynamism and rapidly expanding nature of the digital advertising system, when really it should have been viewed as a warning flag for marketers: a mad rush of middlemen all claiming to play a value-adding role in an imperfect system. Of course the vast majority of intermediaries are operating in good faith, but their huge numbers muddied the waters, making the situation worse for marketers.
“In some cases not only does a brand not know where its ad actually appeared, [it doesn’t] actually know how it ended up in different places and how many entities and people touched it along the way,” says Silverman. “And if you create a system like that, it is just perfect for people to rip you off.”
What to do
Programmatic advertising is not inherently bad, and the advertising products developed by the big platforms are remarkably good at delivering targeted advertising, says Silverman. But he has no doubt the current digital ad system is broken—badly. I asked him if he could snap his fingers and make one change what would it be. His response: more transparency.
“Everyone should know—in as close to real time as possible—where every cent of their dollar, and what every ad associated with that cent of every dollar, ran at any given time… And they should know everyone who took every piece of that dollar and where it went,” he says. “I think if we had that level of true and universal transparency, it would be much harder for fraudsters to hide.”
But until then it might just take a more concerted effort from marketers, and a high-tech problem may require a low-tech solution. “I don’t think the stuff I’m doing is that remarkable,” he says of the work he’s done to uncover ad fraud. “I think that if you were [working] at a brand, and you hired a few people who have good open source investigation skills, who understand digital advertising, if you had a small team of people whose job was to defend your spend and review where it’s going, I think you would save a lot of money and you pay their salaries three or four times over.”