Nandini Jammi remembers the moment she learned she was accidentally destroying the news industry. Like so much else in 2020, it came as an unpleasant surprise.
As the cofounder of a social media campaign called Sleeping Giants, Jammi had perfected the art of pressuring advertisers to stop funding websites that peddle hateful content. In 2017, Sleeping Giants attained national notoriety when its viral protests on Facebook and Twitter helped cause a 90 percent drop in Breitbart News’ ad revenue—kneecapping one of the most important and inflammatory right-wing publications of the 2016 election. The experience left Jammi, then in her late twenties, exquisitely attuned to the central economic fact of internet content: Good, bad, or neutral, it’s all sustained by advertising.
This past March, as the full horror of the coronavirus pandemic began to reveal itself, Jammi watched the news business go into a financial tailspin. Even as traffic to coronavirus coverage soared, advertisers slashed their marketing budgets, starving publications of revenue. Hundreds of publishing-industry layoffs, furloughs, and pay cuts soon followed (including at Condé Nast, which owns WIRED). Jammi had a hunch that the situation was even worse than it looked. Not only were advertisers spending less, but what they were spending, she suspected, wasn’t making its way to news publications. So-called keyword blocking software, designed to keep brands from appearing next to “controversial” topics, might be preventing ads from running on stories about the pandemic. Jammi went into a Slack channel for marketers to see if anyone would leak data to confirm her suspicions. At one point she asked why advertisers cared so much about avoiding unpleasant news topics.
“I mean this with all due respect,” one chat participant replied, “but they care because of you (your org), and it’s broadly your fault.”
Jammi’s shame campaigns may have worked too well. Companies afraid of ending up in Sleeping Giants’ crosshairs, and the advertising agencies that work for them, have increasingly turned to “brand safety” tools offered by the ad tech industry. Keyword blocking purports to keep brands far away from unsafe content, but in practice it often steers advertising away from the news, accelerating the decline of an already beleaguered industry. When the pandemic hit, terms like “coronavirus” and “Covid-19” found their way onto block lists. The results were disastrous. A leaked spreadsheet from one major advertiser showed that software was blocking more than half of the company’s ads from appearing on the sites of dozens of reputable publications, including the Boston Globe, CBS News, and Vox. (For WIRED, the number was “only” 41 percent: nearly 77,000 ads blocked in March.) Many brands were skipping block lists and pulling their ads out of news publications entirely. The news media was getting a smaller slice of a shrinking pie.
Re: Targeting is made possible by Omidyar Network. All WIRED content is editorially independent and produced by our journalists.
“As keyword blacklisting ‘coronavirus’ continues to decimate the news industry, I have had the sinking feeling that Sleeping Giants (a campaign which I co-run) has something to do with it,” Jammi wrote to her newsletter subscribers on April 1. “When Sleeping Giants started tweeting at companies asking them to take their ads off Breitbart, we thought we made it pretty clear why: Breitbart was a media outlet promoting hate speech and bigotry, and advertisers’ dollars were funding it. What we never imagined was that brands would turn off the tap on all ‘NEWS & CURRENT EVENTS’ too.”
Her detractors delighted in the irony. “What’s astonishing is that the cofounder of Sleeping Giants would be so clueless that she ‘never imagined’ this would be the outcome of media boycotts,” wrote the conservative columnist Mark Hemingway. He added, “We’d be better off ignoring scolds and censors such as Sleeping Giants who, by their own admission, are wrecking the news business.”
What Jammi’s critics didn’t know was that she was, by then, recognizing the limits of Sleeping Giants herself and was already plotting her next move. She’d realized that the decline of reputable news, the rise of block lists, and the proliferation of online garbage were all perverse outgrowths of the same rotten, underlying fact: The people in charge of corporate marketing budgets have no idea where their ads are actually running, because they have come to outsource the central task of ad placement to vendors of automated technology. And, like the credit-default swaps and derivatives of an earlier era, that inscrutable, dizzyingly complex digital machinery rarely works as promised. Instead, it spends an unaccountable swath of the world’s ad budgets on a whole galaxy of junk websites and apps built to harvest fake clicks.
Jammi’s solution is simple and audacious. If digital marketing has helped wreck online discourse while wasting a ton of corporate money, then perhaps marketers are in the best position to save online discourse. She just needs to teach them to take control of where their ads are running—and in the process, take on the $300 billion digital advertising sector with a business of her own.
Buying an ad used to be pretty simple. A company or its advertising agency would negotiate directly with a publisher to buy a few seconds of airtime, a page in a glossy magazine, or a square of newsprint. The publication was a proxy for the audience you wanted to reach. Selling golf clubs? Buy some inches in a magazine read by old, rich guys.
That’s still broadly how things work in print and on TV and radio. But in the digital realm, the old-school approach has all but disappeared, replaced by something called programmatic display advertising. Programmatic ads are, generally speaking, ones that target users wherever they happen to go on the web, based on all sorts of personal information amassed from financial transactions and inferred from online behavior.
When you visit a site or open an app, it immediately sends out a signal to advertisers: Hey everyone, this (anonymized) person is looking at my page; who wants to show them an ad? Advertisers who want to reach users like you then compete in a real-time auction for your eyeballs. It’s a terribly complex system, with a tangled web of intermediaries and software between you and the advertiser, and it occurs entirely in the time it takes the page to load. (More than one expert has compared it to high-frequency stock trading.) All of which makes it hard for brands to keep track of where their ads are being seen and their money is being spent. You can’t hash things out with an algorithm over a three-martini lunch.
This was the world Jammi entered when she got her start in marketing. Jammi immigrated from India as a kid and grew up in the DC suburbs. After graduating in 2011 from the University of Maryland, where she worked on the campus newspaper, she headed to Europe, supporting herself through remote digital marketing work. That led to a gig at a startup in the United Kingdom where, Jammi says, she became a one-woman direct-marketing team, trying to get companies to buy her employer’s product management software. She was fascinated by the psychology of the job. “It's basically like running a persuasion machine,” she says. “You’ve got to convince them to do something that they otherwise wouldn't have done.”
Eventually, her boss put her in charge of the company’s first programmatic ad campaign, with a budget of a few thousand dollars. Like the overwhelming majority of small businesses, Jammi used Google’s advertising platform. “I was very careful with it,” Jammi says. “I wanted to know how the money was being spent.” So she went into the back end to check where her ads were being seen. The sites were mostly ones she hadn’t heard of. The only one she recalls recognizing was Zero Hedge, a libertarian blog that dabbles in conspiracy theories and would later be suspended by Google Ads for publishing racist content. “Is this how it’s supposed to work?” she wondered. “Why would these ads be served on these sites that I have not heard of? Why are they not being served on a publication I have heard of?”
Jammi asked her boss for help figuring out what had happened to the money they spent. As far as they could tell, the whole campaign resulted in a measly two people signing up for free trials of their product. Jammi was still a marketing newbie, but by taking the unusual step of trying to audit her own campaign, she had learned something important about digital advertising: If you don’t take careful control up front, your ads can end up anywhere, including places that make no sense for your brand.
The experience was fresh in her mind a few months later when, shortly after the 2016 election, Jammi visited Breitbart.com for the first time. The site was riding high, having emerged as a more explicitly nationalist alternative to Fox News. Which is why Jammi was surprised to see a display ad for Old Navy featuring an interracial couple. It occurred to her that Old Navy probably had no idea its ads were running on a site that maintained a “Black Crime” article tag.
Jammi wrote a post on Medium urging marketers to stop running ads on the site. “Just go into your Adwords account now and pull your display and remarketing ads off Breitbart.com,” she wrote. “Breitbart News is fully covered under the 1st amendment — but that doesn’t mean we have to help fund it.”
The post caught the attention of Matt Rivitz, who had just created the Sleeping Giants account on Twitter and called out the mortgage company SoFi for advertising on Breitbart. (The name Sleeping Giants has no particular meaning; Rivitz just thought it sounded cool.) The two decided to join forces, Rivitz from San Francisco and Jammi from Berlin, and began challenging other well-known companies. By early 2017, hundreds of major brands, including the likes of Kellogg’s, BMW, and Lyft, had blocked the site, leading Breitbart’s executive chair, Steve Bannon, to admit that the campaign had crippled Breitbart’s finances. Sleeping Giants went on to claim more victories, pushing advertisers to abandon Fox News’ Tucker Carlson, who suggested on air that immigrants make the US “dirtier,” and spurring a boycott of alleged serial sexual harasser Bill O’Reilly that helped get the host fired.
The campaign was initially anonymous, but eventually the Daily Caller, a conservative outlet cofounded by Carlson, outed Rivitz. He and Jammi were then profiled in The New York Times, which gave the campaign new visibility. They became a household name within the advertising industry, and not necessarily in a good way. “Sleeping Giants will finger-wag and name and shame for the slightest things,” says Robert Rakowitz, who was head of global media at Mars when Sleeping Giants sicced its followers on the candy giant, prompting it to cut ties with Breitbart. Rakowitz appreciates what Sleeping Giants is trying to do but resents their approach. “That’s their business, calling people out, and it’s not helpful to the situation. They’re only accelerating polarization.”
Sleeping Giants wasn’t the only source of incoming fire that advertisers were facing. In February 2017, journalists at the Times of London revealed that prominent brands were inadvertently subsidizing YouTube videos created by terrorist sympathizers, Nazis, and the like, by running ads on them. In response, major companies like Johnson & Johnson pulled their advertising from YouTube altogether; several European brands likewise boycotted the platform.
As interest in “brand safety” spiked, the ad tech sector saw a business opportunity in the anxiety. “Ad verification” companies like DoubleVerify, Integral Ad Science, and Oracle Data Cloud offered the promise of a technological solution to the Sleeping Giants problem: They compiled lists of keywords that were associated with controversy and promised brands that their advertisements would never appear on pages that included those terms. Advertisers and their media agencies eagerly signed up.
But blocking tools defined “brand safety” in terms far more broad and haphazard than what Jammi had been shaming companies for. In the old days, an airline might call up a newspaper to pull an ad that was slotted to run next to a story about a recent plane crash because the “adjacency” was seen as bad for business. Now, however, ad verification vendors were drawing up lists of blocked terms that could stretch into the thousands. The results could be laughable. The magazine publisher Hearst, for example, has complained that articles about Meghan Markle, the Duchess of Sussex, were being blocked because “sex” appeared on block lists. Less laughable: The definition of “controversial” could smuggle in ugly biases, flagging terms like “lesbian” and “bisexual.”
On the whole, the innovation brought no obvious benefit to advertisers, but it delivered pain to the news business. An analysis of newspaper and magazine publishers in the US, United Kingdom, Australia, and Japan concluded that overblocking deprived them of $3.2 billion of digital revenue last year. Meanwhile, bottom-feeding publishers whose entire business model involved gaming the programmatic ad system—and not, say, faithfully reflecting world events—adapted and found ways around the block lists. Breitbart, for example, appears to have survived the advertiser exodus in part by figuring out how to pose as other, more reputable sites in the automated ad marketplace. Moreover, keyword-blocking tools didn’t even work on “closed” platforms like Facebook or YouTube, even though YouTube videos are what prompted much of the soul-searching.
Jammi kept her day job in digital marketing while she helped run Sleeping Giants, and the campaign kept claiming scalps. In late 2017 it pressured institutional investors to force Breitbart backer and Trump megadonor Robert Mercer to step down as co-CEO and resign from the board of his company, Renaissance Technologies, and in early 2019 it helped get conspiracy theorist Alex Jones’ InfoWars channel removed from the Roku TV platform.
But Jammi and Rivitz weren’t only pissing off right-wing kingmakers and alienating advertisers; they were increasingly at odds with each other. In May 2019, Sleeping Giants won an award at the Cannes Lions International Festival of Creativity, a major advertising industry event. From the outside, it looked like a high point. In fact, it was a breaking point. Rivitz went to Cannes without telling Jammi. She didn’t even know about the award until after the fact. (Rivitz ran the team’s Twitter account, while Jammi ran the Facebook page, so they didn’t necessarily both see the same messages that came in.)
“After we win an award, he tells me by DM that we've won an award,” she says. “I didn't know what a big deal it was. And then all these people start messaging me, and they're like, ‘Hi, like, that's amazing. Why aren't you in Cannes?’”
According to Jammi, this was part of a long pattern of behavior in which Rivitz tried to hog credit for Sleeping Giants and keep Jammi from being identified as an equal partner in the press. The Cannes snub was the last straw. So even as Jammi continued to run the Sleeping Giants Facebook account and identify herself as cofounder in the press, she began planning her next move. (She formally left Sleeping Giants this past June, writing a viral Medium post with the subtitle, “How my white male cofounder gaslighted me out of the movement we built together.” Rivitz issued an abject public apology soon after.)
Jammi was starting to realize that the Sleeping Giants approach—calling out a particular brand for subsidizing a particular site—only went so far, anyway. “At Sleeping Giants, I was playing Whac-A-Mole,” she says. The effort might get that brand to knock it off, but it didn’t fundamentally change how they approached marketing. The bigger problem lay in advertisers handing over their campaigns to technology they didn’t understand.
The full scale of the Whac-A-Mole problem hit home for Jammi when, last September, she met a Twitter acquaintance named Augustine Fou for coffee in Manhattan. (Jammi moved back from Germany in 2019.) Fou is a longtime digital marketer turned ad fraud investigator and strident critic of the industry. “It’s cool that you got Breitbart taken care of,” Fou told her. “But did you know that there's hundreds of thousands of Breitbarts out there?”
Fou’s point was that the rise of online hate was intimately tied to the much broader phenomenon of programmatic ad fraud. In the old days, he explained, someone who wanted to publish racist propaganda would have had a hard time getting anyone to read it and an even harder time paying for it. But social media’s algorithms make it easy to amplify inflammatory content—and programmatic advertising makes it simple to monetize. It’s trivially easy to set up a site and get it accepted into an ad tech company’s network, even if the site is full of hateful or even plagiarized content. These bottom-dweller sites can undercut the prices charged for ad space by more legitimate competitors, so that they not only pull ads away from real journalism but also lower the going rate. And for every Breitbart or Zero Hedge, there are thousands of straight-up fraudulent sites designed simply to attract clicks, either from bots or human users duped into visiting the URL.
Until her meeting with Fou, Jammi had been thinking in terms of which website to target next. Now she realized she needed to think about how to change the system itself. And the fraud angle provided an important clue. A junior-level marketer or media buyer at an agency might or might not care about the health of digital media. But show them that they’re wasting their boss’s money on rampant fraud? That could get someone to pay attention.
Thus was born Jammi’s latest venture, Check My Ads, which she cofounded with Claire Atkin, a fellow marketer with an idealistic streak. Atkin, a 32-year-old Canadian based in Vancouver, was so disturbed by the role of social media in the 2016 US presidential race that she went to Italy for a course in international election observation the next year. Her takeaway: The people in charge didn’t have the answers to online disinformation because they didn’t understand the business of it. “I realized, with much respect to the teachers, that because of my background in technology I actually knew more about the subject than they did,” she says.
The name Check My Ads pretty much sums up the idea: For starters, Atkin and Jammi would help marketers figure out where their ads were really showing up, then help them stop subsidizing offensive material and wasting money on fraud. Ultimately, they would try to spark a revolution in the profession, teaching marketers to break their dependence on automation and rediscover the value of running ads on quality publishers. It’s a scheme whose grandiosity is matched only by the stunning dysfunction of the digital advertising industry itself.
As 2019 rolled into 2020, Jammi and Atkin remained focused on helping brands to stop subsidizing crappy sites. Then the coronavirus pandemic hit, and the inverse problem—of brand safety technology that blocks ads from appearing on legitimate news sites—suddenly attained crisis proportions. As “covid” and “coronavirus” appeared on block lists—and ad revenue to vital news publications plummeted in the midst of a public health emergency—the insanity of the whole system became painfully clear. Ads were being directed away from the likes of Time and Der Spiegel but were still finding their way to, say, fringe sites that promoted drinking bleach to cure autism. The programmatic advertising machine was broken, and brand safety technology, it seemed obvious to Jammi, was less a fix than a kludge on top of the problem.
The ad-verification vendors vigorously dispute this characterization, pointing to the growing sophistication of their products. DoubleVerify touts its “semantic science solution,” Oracle Data Cloud its “contextual intelligence.” In a blog post, DoubleVerify’s chief operating officer, Matt McLaughlin, says that his company’s keyword tool blocks less than 2 percent of news publishers’ ads. He also reminds advertisers to practice “keyword list hygiene.” While DoubleVerify does block “emerging negative content,” among other categories, it says that brands should remove those terms once the topic is no longer emerging.
Even more destructive than keyword blocking is another increasingly widespread practice among marketers in the era of brand safety: refusing to advertise on news publications at all. In a 2018 survey by Digiday, 43 percent of media buyers said they avoid news content at least sometimes. There is no evidence that this makes business sense; there’s no reason to think someone who sees an ad for Domino’s while reading an upsetting story about, say, police violence will be any less likely to order their pizza. To the contrary, one recent study found that the “halo effect” of advertising on a set of established, respected news sites led to a massive increase in “brand lift,” a measure of warm consumer feeling, compared to ads that ran across the rest of the internet. When a brand blocks these publishers, by contrast, its ads will only show up in less prestigious environments, funneling ad money toward the internet’s long tail of obscure, lower-quality sites.
Ad verification companies may not officially endorse avoiding the news wholesale, but Jammi blames them for sending advertisers in that direction. And she rejects the idea that the corrective to overblocking is to rely even more heavily on the newest software. “You can’t automate your way to brand safety,” she says. Layering on new algorithms to clean up the mess of other algorithms only makes things worse. The vendors, she points out, want to show that their products are working, which gives them an incentive to block a lot of ads. “I feel comfortable antagonizing the ad-verification industry,” she says, “because they’re lying to their customers.”
The difference now is that Jammi has set out to antagonize the ad-verification industry in a way that involves having customers herself. “I consider this the end of me calling out companies publicly,” Jammi says. “I want to now help them to ensure that never happens to them. And I want to save them money.”
Jammi got her chance this spring, when Check My Ads found its first client: Andrew Lissimore, the Vancouver-based co-owner of Headphones.com, an online retailer that specializes in top-end audio equipment. He met Atkin at a Zoom happy hour event in early spring, back when people still agreed to Zoom happy hours. Intrigued by her project, he agreed to have her check out his own ad campaign. The company was spending about $1,200 per day on retargeting ads, the kind that ostensibly follow people who have already shown some interest by visiting the site.
Lissimore was pretty confident that Atkin wouldn’t find anything alarming. “I thought the whole time that there was nothing wrong on my end,” he says. “I just thought their project sounds interesting.”
He was mistaken. Under Atkin’s guidance, he emailed the retargeting vendor, Criteo, to get a list of the publishers where his ads were running. Atkin combed through them and picked out some alarming examples. Ads for Headphones.com were appearing on sites that Lissimore finds personally abhorrent, among them Breitbart and the conspiracy-mongering Epoch Times.
But by far the biggest portion of his budget was simply being wasted. A big chunk was going to sites in countries in Latin America where Headphones.com does no business. And an amazing 80 percent of the retargeting campaign was going to Android games. That made no sense. Lissimore’s customers are the type of people who might drop $5,000 on a headphone amplifier; 95 percent of them visit the site on Apple devices. They probably weren’t taking a break from playing Fun Race 3D to make an impulse purchase on their phone. This was almost certainly fraud: bots were visiting Headphones.com, then opening these Android games, where they would get served Headphones.com ads. Then the game developers would get paid. After Lissimore asked Criteo to start blocking the objectionable publishers and useless Android games and limit his targeting to the US and Canada, his daily spending dropped to less than $80 overnight. Sales didn’t change.
Lissimore remains incensed at Criteo. His user dashboard didn’t allow him to see, let alone choose, where his ads appeared. He had to email the company to even get that information, and then again to ask it to block certain sites. “They have a version of their dashboard that includes this crucial information, but they purposely don’t share it unless you ask,” he says. “I’d imagine millions of dollars are flowing to these publishers from companies who wouldn’t be caught dead sending them a single cent. All because Criteo is purposely opaque about where the money is going.” (According to Jessica Meyers, a Criteo spokesperson, the information Lissimore requested is “automatically enabled and visible to all clients in the Management Center in the Americas and Europe.” But Lissimore shared emails and screenshots showing that this wasn’t the case. Meyers also said Criteo had added the Epoch Times, along with the Gateway Pundit, another conspiracy-mongering site, to its default block list.)
The moment couldn’t be much riper for a backlash to programmatic ads. Evidence keeps piling up that the technology isn’t all it’s cracked up to be. A recent study of ad campaigns by major brands by the Incorporated Society of British Advertisers found that ad tech middlemen were capturing fully 50 percent of the money spent, nearly a third of which seemed to have simply disappeared. In the Netherlands, the country’s main public broadcaster recently got rid of microtargeting entirely and saw a big increase in digital ad revenue.
Meanwhile, companies are grappling with the morality of their advertising practices in unprecedentedly public ways. The spate of articles about Covid-related keyword blocking in March brought unwanted attention to brand safety technology. And more dozens of prominent brands joined a temporary Facebook ad boycott, partly organized by Jammi’s former partner Matt Rivitz, designed to force the company to update its policies on hate speech. That’s different from programmatic advertising, but it suggests that advertisers are beginning to recognize the leverage they wield over the ad tech giants that depend on their continued spending.
Even Robert Rakowitz, who got stung by Sleeping Giants when he ran advertising for Mars and now leads the Global Alliance for Responsible Media, describes goals that sound similar to Jammi’s. “If the reality is the whole industry pulls away from news, then what are we left with?” he asks. “State-owned media broadcasters and publishers, and tech platforms. And that’s not a good thing.”
In Rakowitz’s eyes, advertisers have been chasing the false promise of programmatic advertising to find the right customer at the cheapest possible cost. Like Jammi and Atkin, he wants advertisers to start by picking publishers they’d like to advertise with, rather than picking the audience they want to microtarget.
But when asked what exactly the alliance is working on, Rakowitz slips into corporate jargon: “We’re going to be kicking off work later this summer where we start to get into KPIs and developing what we consider goals.” (KPI is short for “key performance indicator,” business-speak for tracking progress.) The wheels of global business alliances grind slowly, if they grind at all. It’s fair to wonder whether a consortium that includes Facebook and Google—the two dominant digital advertising companies—will produce any meaningful change to the status quo.
Jammi and Atkin aren’t willing to wait to find out. Though they won’t disclose their current clients, they have begun working with some prominent multinational corporations, “including a well-known pharmaceutical brand,” according to Jammi.
“This is, like, my biggest marketing campaign ever,” she says. “I’m just doing what I did at that old job, but at a bigger scale. I’m showing them a solution to a problem they don’t know they had.” Sometimes progress looks like going back.
*Updated 8/14/2020 2:50 pm ET: A previous version of this article incorrectly suggested that DoubleVerify adds emerging news topics to brands' block lists by default. *
Do you use social media regularly? Take our short survey.
More Great WIRED Stories