With brand safety issues among the biggest concerns for marketers, the audio advertising agency Oxford Road is teaming up with the artificial intelligence company Barometer to launch a new AI-powered brand suitability tool that they say will allow ad buyers to assess the “risk profile” of a podcast and whether it aligns with their brand values. The answer would guide whether a show is part of the media plan or not.
The tool is powered by natural language processing and would include risk profiles for shows and specific episodes for such things as adult and sexually explicit content, profanity and obscenity use, and discussions about controversial social issues.
The solution also monitors positive media behaviors, including values from Oxford Road’s Media Roundtable initiative, such as Tolerance, where advertisers can see ratings for a program’s efforts to support collaboration between parties with conflicting objectives and views.
“The solution we developed with Barometer was driven by the needs of buyers and is not funded by publishers or driven by political ideology. For the industry to achieve a sustainable solution, impartiality is critical,” said Oxford Road founder and CEO Dan Granger. He said their AI solution considers context and sentiment and will benefit not only ad buyers but also podcast publishers. “Finally, networks and platforms can have greater visibility into the content they are distributing or plan to distribute and offer a layer of protection from trusted third-party standards to their customers,” Granger said. “For platforms tasked with monetizing the long tail of smaller programs, this now provides a viable path forward to bring sponsors to the thousands of shows that wouldn’t get the attention needed to adequately vet for safety and suitability using human-driven or keyword-based solutions.”
Unlike some keyword-based solutions, Barometer analyzes entire transcripts in context, interprets each utterance, and assigns a concrete risk score based on components laid out in the Global Advertisers for Responsible Media’s (GARM) Brand Safety Floor and Suitability Framework. The data-driven content rating capability is then used to measure content-in-context across 11 content topics that may be harmful to brands. This new level of topic-modeling-based analysis references knowledge graphs the company has been building for years. The solution melds data from external sources into one “intelligent contextualization” algorithm for fair and meaningful podcast content rating.
"The agencies and brands we serve require a solution to support diverse needs and brand values because what's ideal for one brand is a risk for another,” said Tamara Zubatiy, founder and CEO of Barometer. “Ultimately, the content insights our solution delivers help brands inform their ad choices in much the same way that consumers count on nutrition labels to guide their food choices.”
No Need To ‘Bluntly De-Platform Controversial Voices’
The Barometer dashboard will make it easy for agencies, brands, and networks to see exactly what they need to quickly evaluate podcast content at a very granular level and then instantly filter podcast shows and episodes that do not align with their brand values and safety requirements. Networks also benefit from the Barometer, as it can programmatically plug into any internal platforms and systems via API.
“As media fragmentation drove a renaissance in new forms of content that consumers love, serious risks emerged as well — making it harder than ever for advertisers seeking to balance their brand values with their business objectives. With so little oversight of content, particularly in the podcast ecosystem, this led many brands to avoid particular genres and, in some cases, podcasts altogether, a classic case of throwing the baby out with the bathwater,” said Granger. “The right way of tackling these major challenges is nuanced and technology-based. It does not require publishers to bluntly de-platform controversial voices without the insights needed to guide their choices. It instead provides the unbiased information everyone deserves to have about the content they are consuming, sponsoring, or distributing. In short, it gives brands the information needed to reduce risks and drive the results audio can offer.”
In addition to her role leading Barometer, Zubatiy is also a member of the Interactive Advertising Bureau’s (IAB) Tech Laboratory, working to advance technology standards to support responsible scaling across the industry. “Our expert team brings a data-critical approach to AI bias through model transparency and ongoing model revalidation. Our goal is to become an IAB standard by year’s end,” she said.
Barometer and Oxford Road’s new brand safety and suitability solution for podcasts will be unveiled at SXSW on March 14.