Acast Debuts AI Tool Matching Podcast Conversations With Relevant Ads
Podcast platform Acast has introduced an AI-powered system for embedding ads in specific podcast episodes. The Conversational Targeting feature transcribes podcasts and uses natural language processing (NLP) to understand and categorize individual episodes to match ads with thematic links to the discussion, even if the overall topic of the podcast is unrelated.
Standard podcast advertising relies on category tags for a whole show, such as real estate, movies, or voice technology. Conversational Targeting analyzes and tags each episode separately, making ad insertion much more dynamic and relevant to what the host or hosts are discussing. For instance, a podcast all about analyzing cars might have an episode where the hosts talk about their favorite burgers for 20 minutes. Acast’s new tool would note and tag the episode as one where food and burgers are brought up and a listener might hear a restaurant ad at the end of the burger talk where they might previously only heard automobile-related marketing. Advertisers are suddenly reaching far more people without the ad coming off as a nonsequitur. The new system could boost podcaster revenue in tandem with the potential sales bump for the advertisers. As a bonus, the producer-side focus of the targeting also means listeners aren’t getting targeted based on their personal information. The ads can be more relevant without raising privacy concerns.
“Conversational Targeting is another major innovation from the Acast Marketplace, bringing advertisers the ability to reach listeners within the most relevant context ever — all while protecting the listener’s privacy,” Acast CEO Ross Adams said. “Our mission is to make podcast advertising the best it can possibly be, and this is a big step for us in bringing better targeting solutions for advertisers — as well as helping open up more podcasters’ catalogs for ad revenue, so they can keep on making the shows their listeners love.”
The advertiser-based approach taken by Acast is taking stands out, but applying conversational AI to better categorize podcasts is not new. Audio content search startup Audioburst offers tools like the Finder widget for spotting words and phrases in audio in order to create the short ‘bursts’ that Audioburst was namer for. The idea is to help podcasters share their shows and make it easier for people to discover audio they may want to hear from even just a small snippet of a show. And Google has directly linked its search engine to the podcast world with search engine results that include podcast episodes.
Acast has plans to enhance the Conversational Targeting over time with better context understanding and keyword avoidance. There are also plans to analyze the emotional content of a podcast for the best ad insertion strategy
“As well as knowing the exact words being spoken, we’ll be working to better understand the emotion and sentiment of those conversations,” Acast vice president of advertising product Chris Wistow explained. “We’ll know, for example, whether a topic is being discussed with a negative slant that could reflect badly on an advertiser. Conversely, we’ll be able to highlight positive conversations relevant to your brand as a great moment to reach listeners.”