This report is being co-published with NewsGuard.
The breach of the Capitol building last week by armed protestors who believed the 2020 election had been stolen from President Donald Trump demonstrated the real-world consequences of misinformation in vivid fashion. The protestors believed false claims about rigged ballots, inaccurate vote counting, and malfunctioning voting machines—falsehoods that have circulated on misinformation websites and social media platforms over the past several months.
Companies, brands, and technology platforms have been quick to condemn the violence, with some going so far as to drop support for those they saw as encouraging or enabling the riots.
But an analysis of programmatic advertising data shows that many of the world’s largest and most trusted brands have been financially supporting websites, including sites funded by the Russian government, that spread election-fraud myths and conspiracy theories—placing thousands of programmatic ads on these sites during the period leading up to and following the 2020 election.
From Oct. 1, 2020 through today, 1,668 brands ran 8,776 unique ads on the 160 sites flagged in NewsGuard‘s Election Misinformation Tracking Center for publishing falsehoods and conspiracy theories about the election.
It is likely that, in many cases, these ad placements were inadvertent and unintentional—that is, the brands did not intend to fund hoaxes and conspiracy theories. But the data demonstrate how the various players involved in programmatic advertising, from ad exchanges to verification companies to agencies, are not providing brands with adequate tools to avoid funding misinformation that threatens democracy.
In this report, we provide data and examples demonstrating the scale at which misinformation is fueled and funded by advertising revenue from top brands—and discuss ways the advertising industry can do a better job of placing ads responsibly on behalf of brands.
Trusted brands supporting untrustworthy websites
Our analysis, which combines data from NewsGuard’s election misinformation tracker with data from Moat Pro, found that of the 1,668 brands that funded election misinformation websites, 697 ran ads on multiple sites from the list.
The brands advertising on the largest number of election misinformation sites were Progressive Insurance, which ran 298 ads across 25 separate election misinformation sites, and Planned Parenthood, which ran 71 ads across 18 of the sites. But brands funding election misinformation spanned a wide range of categories and industries.
For example, the advertisers placing ads on election misinformation included a range of nonprofit groups. The AARP placed 57 ads across 16 election misinformation sites, including OANN.com, ZeroHedge.com, and TheGatewayPundit.com. The American Cancer society placed 50 ads across 11 of the sites, and Sloan Kettering Cancer Center placed 28 ads across 7 sites. Other nonprofits on the list included the Alzheimer’s Association; the Leukemia and Lymphoma Society, the American Lung Association, St. Jude Children’s Research Hospital, and CharityWater.org.
Harvard University ran ads on 10 sites flagged for publishing election myths, including OANN.com, which has claimed falsely that voting machines in many swing states were rigged to count votes intended for President Trump as votes for President-elect Joe Biden instead—along with a range of falsehoods about politics and the COVID-19 pandemic.
Large retailers and consumer goods companies made the list. Walmart ran ads on 11 election misinformation sites, including NOQReport.com, a site that has claimed, among other falsehoods, that the CIA worked with Democrats to change voting machine results and that COVID-19 was planned by Bill Gates and other billionaires.
Procter & Gamble ran ads on multiple election misinformation sites, including TheGatewayPundit.com, a site that has pushed false claims about voting machines and previously published falsehoods about COVID-19. Disney, which prides itself on its apolitical and family-friendly brand, placed ads on multiple sites with election misinformation, including CharlieKirk.com.
Major financial companies and insurance providers were among those placing ads on election misinformation sites. For example, American Express advertised on SputnikNews.com, a site controlled by the Russian government that targets propaganda and falsehoods to U.S. audiences, and AmericanThinker.com, which in addition to falsehoods about the election and COVID-19 has advanced the false claim that President Obama was born in Kenya.
Brands advertising on election misinformation included government agencies funded by taxpayer dollars, including the U.S. Department of Veterans Affairs, which placed ads on seven election misinformation sites and the U.S. Department of Homeland Security, which placed ads on two.
The U.K.’s public broadcaster, the BBC, placed ads on five sites pushing election misinformation.
Most election misinformation sites were repeat offenders
The data demonstrate that efforts by brands, agencies, ad verification companies, and ad exchanges to ensure brand safety and suitability for large brands’ ads have largely failed when it comes to misinformation about the 2020 U.S. presidential election. The problem was especially pronounced for the largest ad exchanges: Google‘s DoubleClick advertising service placed ads on more than 80% of the election misinformation sites while ads served by The Trade Desk appeared on more than half.
But the problem of misinformation and brand safety for advertisers extends beyond the 2020 election—because the vast majority of misinformation publishers are repeat offenders. Of the election misinformation sites for which programmatic advertising data was available, 71% had previously been flagged by NewsGuard for publishing misinformation about COVID-19—and in many cases, for publishing other falsehoods prior to the pandemic.
For example, before publishing election myths and conspiracy theories, ZeroHedge.com promoted false claims that COVID-19 is manmade and parroted widely debunked Russian propaganda narratives. The site earns a trust score of 12.5 out of 100 on NewsGuard’s rating system. But despite its credibility problems, ZeroHedge.com received ad revenue from at least 250 brands over the past three month, including AT&T, Bloomingdales, New Balance and Stanford University.
TheGatewayPundit.com saw ads from 226 brands during the period surrounding the election, including Michaels, Best Western, Best Buy, and Verizon. Prior to publishing election falsehoods, The Gateway Pundit had advanced false claims that face masks are dangerous because they push COVID-19 into the brain; that Dr. Anthony Fauci stands to profit personally from the COVID-19 vaccine; and that Barack Obama is a Muslim.
This can be fixed
While this report names many specific brands and websites, it is important to note again that, in many cases, companies may have little idea where they are placing ads at all. Programmatic advertising systems are so automated, and their processes so byzantine, that marketers for most brands only see topline metrics about ad performance, total impressions served, and the like.
Instead, brands trust their campaign placements to agencies, brand safety and ad verification companies, and ad exchanges, which promise varying forms of protection against ad placements on unsafe or untrustworthy content. Yet despite these measures, our analysis shows that hundreds of top brands have nonetheless ended up placing ads on misinformation and conspiracy theories. While brand safety companies have been able to keep programmatic ads off of sites publishing pornography and using hate speech, their artificial intelligence tools are unable to differentiate between generally trustworthy sites and sites that publish misinformation and hoaxes.
Taken as a whole, the data suggest that brands seeking to avoid funding misinformation could address much of the problem simply by avoiding ad placements on sites that have a history of repeatedly publishing falsehoods, while targeting ads to sites that have a history of publishing credible information and strong transparency practices. Large ad exchanges like Google and The Trade Desk can and should make tools available so that brands or their agencies can easily target ad placements based on trust and credibility, the same way they now can target based on factors like a user’s location, keywords on a page, or other criteria.
NewsGuard offers one such solution, a product for advertisers called BrandGuard. But ultimately, addressing the issue of advertising on misinformation will require companies to recognize the problem and prioritize it as a component of their advertising strategy.
Matt Skibinski is the General Manager of NewsGuard.