At the end of 2016, the anger trade is a good business to be in. Audiences are reading and sharing fake “news” that stirs partisan ire while hardline outlets like Breitbart can count the president-elect as a fan. Advertising dollars are following those eyeballs, and thanks to the automated nature of online ads, the cash can flow into these sites often without the knowledge of companies doing the actual advertising. As a result, brands run the risk of appearing to endorse political stances or rhetoric that potential customers may find objectionable. Now some are starting to worry.
A growing number of brands, from Kelloggs to Allstate to Warby Parker, have pulled their ads from Breitbart. Meanwhile, at least one brand, Fiat Chrysler, said it adds fake news websites one by one to a blacklist it maintains. Others are scrambling to rejigger their ad policies to account for such fraudulent sites. But the influence of those individual brands may pale next to the power of a handful of companies whose names you probably havent heard: the ad tech companies that funnel ads to those sites in the first place.
These are the companies that run the digital marketplaces where advertisers bid for ad space on sites. They also provide the tech tools advertisers use for ad placement. Last month, one of the biggest, AppNexus, said it would block Breitbart News from using its ad-serving tools because the site had violated its hate speech policy. One week later, two more ad-tech companies, TubeMogul and RocketFuel, followed AppNexus lead.
Breitbarts editor-in-chief calls these decisions an attempt at censorship. But whats really at stake isnt a sites right to say whatever it wants to sayits whether a site can make money saying it. Fake news creators and hardline sites can’t get by if advertisers won’t sponsor their sitesa hard fact of the online publishing business that could make a few obscure tech companies the unlikely arbiters of the future of public discourse.
Last month, a week and a day after the election, the anonymous activists behind Sleeping Giants posted their first tweet: @sofi Are you aware that you’re advertising on Breitbart, the alt-right’s biggest champion, today? Are you supporting them publicly?
The tweet was directed at San Francisco-based Social Finance (SoFi for short), an online personal finance company that helps offers student loan refinancing, personal loans, and mortgages. The company didnt respond publicly. But Sleeping Giants added SoFi to the list of brands it says its confirmed have blocked their ads from appearing on Breitbart. The list currently runs more than300 companies deep. The longer that list gets, the more incentive ad tech companies have to proactively seek to keep their clients ads off sites that might expose them to controversy.
But how is it that advertisers often dont know how or where their ads end up online in the first place? In more traditional advertising, brands designate where their ads will appear in print or on television, and they do so well in advance. But online advertising increasingly involves third-party networks and agencies that place ads across the web based on digital auctions that often happen in near real-timea matter of 100 or 200 milliseconds as a visitor arrives on a site.
The tech behind these auctions relies heavily on targeting and personalization based on what people have searched for and clicked on while surfing the web. If you searched for, say, a pair of boots in the last hour, a brand touting a new line of boots might be served up to you, no matter which website you end up on. Automated buys really chase the audience, not necessarily the context, says John Montgomery, executive vice president for brand safety at powerhouse media buyer GroupM. In other words, those boots might chase you to places on the web the makers of those boots dont want to be seen as supporting.
Yes, marketers can work out a media plan that places ads directly on websites, pre-approving domains as they goa so-called whitelisted environment. But the practice is increasingly falling out of fashion. Josh Zeitz, vice president of communications at AppNexus, says automated media buys have clear advantages. If youre doing just a handful of campaigns and buying against 50 sites, you dont know whether youre reaching the users youre trying to getthey may well be using other apps and other sites, Zeitz says. Automated buying, on the other hand, lets you follow users based on what you know about them, rather than relying on the sites you hope theyll visit.
But Zeitz says ad-tech companies have another big responsibility: brand safety.
You have to be able to count on your ad tech partners to maintain some kind of marketplace quality, Zeitz says. Because if its completely unregulated, your clients ads could show up anywhere. Today, certain standards are well established. Most ad tech companies have policies and filters that block ads from being served against piracy, pornography, and graphic violence. But highly charged political sites and fake news are brand new territory for these companies.
Most ad tech companies punt on the issue of serving ads on Breitbart specifically. They say the decision lies with their clients: advertisers and marketers. GroupM doesnt see itself as a censor, says Montgomery. But what we do is invite our clients to advise us at any time if there are any media properties that the company wishes to avoid, for instance on the grounds of perceived political extremism. Google holds the same view; the search giant still currently serves ads on Breitbart, asserting that the site does not violate its own hate speech policy.
‘The attacks on us are politically motivated, and theyre specifically because weve been effective.’Alexander Marlow, Breitbart Editor-in-Chief
RocketFuel, meanwhile, put Breitbart on its blacklist in November, meaning its clients have to actively opt in to serving ads on the site. But the companys CEO says the decision hinged on creating a more efficient market by trying to anticipate where its clients likely didnt want to see their adsnot on judging Breitbart itself. Efficient markets require trust and transparency between suppliers and buyers, CEO Randy Wootton says. Thats how commerce happens.
Breitbart editor-in-chief Alexander Marlow calls boycotts of his site by advertisers and ad tech companies absolutely an attempt at censoring Breitbart News.
The attacks on us are politically motivated, and theyre specifically because weve been effective, Marlow said in an interview. Weve changed the news narrative time and time again, and weve changed a lot of hearts and minds in the process. Marlow denied that the boycott had any impact on the websites bottom line, and he said Breitbart has no intention of changing its editorial direction in response to the increasing number of companies blacklisting the site. We have tens of thousands of advertisers. We are a loyal community of 45 million readers a montha very powerful consumer group, Marlow said.
Where to draw the line between free speech and hate speech depends on who you ask. Google defines hate speech as content that incites hatred or promotes violence against individuals or groups based on race or ethnic origin, religion, disability, gender, age, veteran status, or gender identity. Other companies have adopted a similar definition. But how different companies apply those standards varies. Google says it found Breitbart did not violate its standards. AppNexus, meanwhile, says it decided to ban Breitbart after an outcry from employees. There was human judgment involved, says Zeitz. There are plenty of people at AppNexus who voted for Donald Trump, and probably plenty of people who read Breitbart. But thats not why we had to remove it from the marketplace. We determined that it violated hate speech.
Of course, plenty of companies just dont care whether their ads appear on a site like Breitbart. Direct sales ads, for examplethose junky buy now! ads that clutter fake news sitesare only about making that sale, not burnishing a brand. But for those that do, new tools are starting to become available, created by ad-tech companies themselves.
For instance, Integral Ad Science, an ad verification company, approaches the problem as a numbers game. It uses tech to comb through individual pages across the web, then rates them from zero to 1000, depending on how risky its algorithms determine putting an ad up against that content might be. DoubleVerify, meanwhile, has created a new category for its clients called Inflammatory News and Politics, which includes Breitbart as well as other hardcore conservative and liberal sites such as Rawstory.com, IfYouOnlyNews.com, and NewsBusters.org.
A new offering from the company also scans for fake news sites as they come online so advertisers and marketplaces can avoid them. The hallmarks for fake news sites are distinct, DoubleVerify CEO Wayne Gattinella says: no documentation behind stories or cross-indexing of stories; vague, generic bylines; no daily updates; and internal links that go to blank pages. Google still largely relies on a human operation to assess which sites should be considered deliverers of fake news, but its also started on an automated effort that can scale. For now, some have criticized Google for spotty enforcement. But the company said it was important to move slowly on the effort to make sure its new policy would be executed correctly.
Because the realm of fake news and inflammatory political sites is still such new territory for advertisers, best practices have yet to become standardized. Brands can choose to serve ads on these sites, then reverse-engineer a public explanation for their actions. Ad tech companies can excuse themselves from pulling ads immediately because it takes time to enforce a policy, and it takes time to scale the tech. And they wouldnt be patently wrong. But that also means, for now, these sites are raking in the ad dollars and spreading disinformation.
In the end, its up to consumers to apply pressure before the next #pizzagate escalates. Dont like that brands are advertising on a certain site? Let them know. And if software companies really believe tech holds the solutions to tech-generated problems, tell them to prove it. When I think of companies the size of Google, Facebook, and others like them, they certainly have the resources to build whatever technology they make a priority, Gattinella says. Ad tech companies themselves might not be as big, but they could still plausibly do the same. Certainly their clients would value such an effort. It all depends on whether brandsand the public they hope to courtmake it clear that no number of clicks is worth the damage to democracy.