Here's the business model behind the media bias chart that's gone viral

For once, Twitter agreed on something: It didn’t like this media-bias chart.

Released twice a year, the most recent chart—a scatterplot of US media outlets organized, according to media ratings company Ad Fontes, by how partisan, accurate, and biased their coverage is—was recently nitpicked by journalists and commentators for attempting to, well, categorize the news, which can be like trying to compare apples to conspiracy-crazed oranges.

Case in point: Though shown as having less bias—i.e., being more credible—news organizations like The New York Times are thrown in with sites such as Infowars, a website owned by conspiracy theorist Alex Jones. The chart draws comparisons between political entertainment podcast Pod Save America and syndicated news and opinion site Breitbart, though neither does much reporting in the traditional sense. The Daily Beast is categorized in the same “news value and reliability” bracket as The Daily Wire, even though one has been nominated for and won awards for journalism, while the other...has not.

But Ad Fontes isn’t for the Twitterati. Founded in 2018, it’s aiming to sell its offerings to advertisers clamoring for a third-party to rate and review The News. It joins players like NewsGuard—a company that also rates news and courts advertisers—by essentially offering inclusion and exclusion lists, bespoke rundowns of publishers that media buyers can or cannot run ads on. The goal? To avoid violating the industry’s nebulous definitions of brand safety.

“People were looking for independent third-party ratings of new sources, and that’s ultimately what we provide,” Vanessa Otero, founder and CEO of Ad Fontes, told Marketing Brew. “Our mission is to make news consumers smarter, and make news media better.”

Otero, a former patent attorney, admits it’s a “lofty” goal. So far, the company lists media agency UM as a client alongside other smaller businesses like Zignal Labs, a B2B data analysis company, as well as Sightly, a media agency that focuses on YouTube, TikTok, and connected TV. Twitter recently used Ad Fonte’s bias ratings to determine if the platform’s algorithm amplified political content.

For $18,000 a year, Ad Fontes sells advertisers a tool that’s effectively an interactive version of its media-bias chart, giving them access to even more news sources than the ones listed—down to specific articles and their scores—allowing individual marketers the ability to decide the threshold of content they can stomach. (The company also has a sales pitch for educators and sells merch, like a mug featuring the chart that says “media literate af”).

So, how exactly does Ad Fontes rate the news?

Here’s the methodology:

To review individual articles, podcasts, and news shows, Ad Fontes relies on roughly 35 part-time analysts, each paid $20 an hour. Each piece of content is given three analysts who self-report as politically left-leaning, right-leaning, or center (uh, leaning?). They individually rate content—whether it’s an article or podcast episode—based on a scale, then combine their scores.

Up to 50 new sources are added every week. According to Otero, big news outlets—think Fox and The Washington Post—have 50 to 100 articles rated before they’re added to the chart. Otherwise, the minimum is 15. For TV shows and podcasts, at least three episodes must be rated until they can be added. Otero said Ad Fontes has rated more than 1,300 news sources to date.

Metrics ranked include “veracity” and “political position,” as well as headlines and graphics used.

• Topics can tilt a score; for instance, outlets that cover climate change tend to tilt left, while ones covering capital-gains taxes usually nudge to the right.
• The more opinion pieces an outlet runs, the faster its reliability score will fall to the floor of the chart. (Otero believes most readers can’t tell the difference between opinion and hard news.)
• By Ad Fontes’s evaluations, the Associated Press and Reuters, traditional wire services that have often served as the backbone of dry, breaking news, are slightly left of center.

Advertisers can then take these ratings and customize them, telling their media buyers to only run ads on publications in the upper-left quadrant of the chart, for example. Otero hopes that, one day, ad buys bought by rating will be done in real time—like contextual advertising, but fine-tuned for editorial bias.

“​​We’re able to create a much more nuanced and specific evaluation based on what’s important to us and to our clients,” William Bock, VP and group partner of global brand safety at UM, told Marketing Brew. Bock, whose clients use both Ad Fontes and NewsGuard, said there is a bit of an overlap, but not completely. Ad Fontes rates podcasts, NewsGuard doesn’t. Yet.

“If you look at those outlying publications that exhibit extreme political bias, but also low reliability or high inaccuracy, those are publications with which we wouldn’t want to associate with,” Bock said.

Though they may both rate the news, NewsGuard doesn’t measure bias—an important distinction, Gordon Crovitz, co-CEO of NewsGuard, told us. Instead, it rates outlets on “journalistic criteria” related to credibility and transparency, like to what extent an outlet clearly identifies the differences between opinion pieces and news.

“This is the problem with trying to rate bias; everybody’s got an opinion,” Crovitz said. “The core problem on the internet is that it is chock-full of misinformation and hoaxes, a lot of which has no ideological component at all.”

(Audience) size matters

Though a striking visual, a media-bias chart isn’t entirely reflective of how most advertisers find audiences; they aren’t always picking and choosing specific publishers to work with. They often rely on algorithms to jockey for targeted audiences and low prices—86.5% of digital display advertising in the US is done programmatically, according to eMarketer.

Though inclusion and exclusion lists can be shown to ad-tech vendors, they can lower the supply of sites, driving up the price of a campaign. Deva Bronson, EVP of digital investment at Dentsu, told Marketing Brew she’s seen campaigns with inclusion lists ranging from 10,000 to 25,000 and higher.

NewsGuard has only rated 7,227 publishers, with only about 65% considered reputable. An inclusion list of only 4,700 sites might guarantee a Nazi-free campaign, but it won’t get the reach advertisers are accustomed to.

However, inclusion lists may provide a reprieve for advertisers that take a heavy hand toward advertising on news outlets specifically. It’s an approach that differs from keyword blocking, the practice of dodging articles featuring words like “Covid” or anything relating to climate change, which can ultimately hurt news organizations trying to cover important and necessary topics.

Dentsu isn’t currently using NewsGuard or Ad Fontes, but Bronson has met with NewsGuard and said Dentsu is always evaluating its partnerships.

“A middle ground is exactly what we need,” said Bronson. “I would argue that a domain-based inclusion and/or exclusion list could be as heavy-handed as a keyword list in that entire domains would be excluded, regardless of content, context, sentiment.”

Bronson thinks better contextual advertising tools could be the answer. “We as an industry have the opportunity to evolve past it.”