Advertisers that rely on ad tech vendors to protect them from buying ads on shoddy, offensive or deceptive news content are … in for some bad news.
Researchers at Stanford University and Carnegie Mellon University recently set out to determine how often ads are served to sites that publish misleading stories, propaganda or conspiracy theories and whether advertisers accurately understand the issue.
The researchers found that, between 2019 and 2021, two-thirds of advertisers bought ads on sites flagged by media watchdogs NewsGuard and the Global Disinformation Index for repeatedly publishing false or misleading content. In a follow-up survey, only 20% of advertising executives believed their company had placed ads on such sites.
Perception vs. reality
Advertisers have a sense of “false uniqueness,” as the researchers put it in their report, which was published in the scientific journal Nature last week.
In other words, advertisers are aware that monetizing misinformation is a problem but generally believe other brands are the ones funding fake news. In fact, most brands end up buying ads on disreputable sites.
“Most companies don’t have a very good view of exactly where their ads are running [or whether] they have a brand safety issue around misinformation,” said Matt Skibinski, general manager at NewsGuard.
The Stanford/Carnegie Mellon study analyzed data from more than 9 million ads placed on 5,485 sites by 42,595 unique advertisers. The media set for the research included 1,276 sites that watchdogs called out for publishing misinformation.
Researchers pulled the data from Oracle Moat Pro, a searchable database of ad placements that was discontinued in 2021 – hence why the data only goes up to that year.
The point of the study was to examine “the consequences of automation in advertising” and the role programmatic tech plays in funding misinformation, said Wajeeha Ahmad, a researcher at Stanford’s Department of Management Science and Engineering.
The researchers were motivated by media reports about brands and ad agencies not understanding how their ads might be financing online misinformation, Ahmad said. For example, she cited an initiative launched by Home Depot and Omnicom investors and nonprofit watchdog Open MIC back in November 2020.
Most brands buy misinfo
Serving the occasional ad to cynical and false news publishers is a general problem. Although, advertisers in particular verticals bought inventory from these sites at far higher rates.
For example, the report shows at least 70% of brands in the food and beverage, business solutions, education, natural gas and electric, insurance, travel, telecommunications and technology verticals buy ads on shady sites.
Meanwhile, media orgs were most effective at avoiding problematic placements. Only 46% of digital publishing brands, 50% of print publishers and 56% of media and entertainment orgs placed ads on these sites. Score one for media literacy, we guess. But even still, nearly half or more than half of these brands failed to avoid supporting fake news.
Holding companies had by far the worst track record, with 82% buying ads on misinfo sites.
Ad tech’s role
While some specific categories fared better than others, however, the big problem is programmatic. Brands of any sort using programmatic tech ended up buying sketchy inventory about 10 times more often than brands that buy direct or in house.
Among brands that don’t use digital ad tech platforms, only about 8% bought from sites deemed untrustworthy by media watchdogs. But nearly 80% of brands that used digital ad platforms purchased suspect placements.
The researchers also conducted a follow-up survey in 2022 with 442 ad execs to gauge how accurately decision-makers grasped the scale of misinfo monetization. Respondents believed about 44% of programmatic advertisers appear on misinformation sites – a far cry from the 80% found in the data.
An undisclosed number of participants contacted for the follow-up survey reached out to the researchers to see if their company had placed ads on the sites in question. Among this cohort, 81% of advertisers had done so. And 62% reported being surprised that their ad dollars funded misinfo.
The Stanford researchers also offered survey respondents the opportunity to participate in a 15-minute course on how companies can avoid having their ads appear on fake news sites. Only 18% of ad execs signed up for this training.
Trust, but verify
What’s unclear from the study results is to what degree brand safety solutions provided by media ratings and verification companies, including NewsGuard, DoubleVerify and Integral Ad Science, prevented advertisers from buying bad inventory.
The data did not identify whether brands used brand safety or ad verification solutions, let alone which tech vendors they partnered with.
However, it’s safe to assume the brands reflected in Stanford’s study were using at least one brand safety solution, because they’ve become a de facto requirement in the industry, Skibinski said. And since commonly used brand safety products haven’t changed much since 2021, the issues raised by the study likely haven’t been solved, he added.
“It’s possible there’s been some improvement,” he said. But NewsGuard has observed plenty of brands placing ads alongside misinformation since 2021.
Misinfo publishers are a moving target, Skibinski said. And brands are stuck with reactive solutions that identify misinformation after it’s published and ads have already run on those pages.
The researchers suggested potential remedies for the digital ad industry, such as industrywide disclosures of whose ads appear on shady sites or public rankings based on how effectively brands avoid misinfo sites.
“Improving transparency about where advertisements appear could by itself reduce advertising on misinformation websites,” Ahmad said.
