If you’ve already plowed through every episode of Serial and are hungry for more, OG music streaming platform Pandora has promising news for you. The radio streaming service wants to help you find The One, if The One was not a human being, but rather, your next marathon-worthy podcast. Today, Pandora debuted its Podcast Genome Project in beta, a service engineered to help users find their perfect podcast fit.
In 2018, podcasts have become so ubiquitous that it feels like you're missing out on a cultural touchstone by not listening to at least one. Last year, the number of Americans who had listened to a podcast at least once grew to 112 million people (or 40 percent of the population), up four percent from 2016, according to a report by Edison Research and Triton Digital.
The Podcast Genome Project’s goal is to help users sift through and discover new podcasts, similar to how Netflix’s algorithms make show and movie recs based on users' viewing histories, or how YouTube suggests Fergie interviews after a night of blasting "Glamorous" one too many times (or is that just me?). The point is, the Podcast Genome Project is a potentially useful tool in a saturated market that’s debuting new content and shows faster than even the most avid listener can consume them. The name is an homage to the streaming site’s original Music Genome Project, the exhaustive database that helped establish Pandora's name through linking dozens of musical attributes to every song in the company’s music library, helping compile stations and aid user music discovery.
According to The Verge, the podcast version of the project echoes similar methods, relying on up to 1,500 attributes to help recommend individual podcast episodes. The company is using “natural language processing” to decipher podcast episode content, attaching episodes with descriptors related to the type of content and production style, among other factors like a user’s podcast ratings and which episodes they have skipped. That means that if you’re a true crime buff who’s listened to Dirty John seven times and want more, Pandora could help you find your next favorite in the genre based on your habits.
“In the case of podcasts, we can rely on machines more heavily than we can with music,” Pandora CEO Roger Lynch told Wired. “Machines can actually determine content, determine intent, there’s many more things they can determine about a podcast than they can about a song.”
Extending the DNA analogy, Pandora’s chief product officer Chris Phillips told Wired, “You can imagine that almost anything that’s discussed in the podcast is a candidate to become a genomic trait.”
Of course, if the algorithms can help users find podcasts, that also means they can help advertisers target listeners, offering a potentially lucrative future for companies hoping to better tailor podcast advertising to listeners’ interests.
But that future isn’t quite here yet, because the Podcast Genome Project is still in beta and will likely see tweaks to its formula in the coming months. For now, though, whether you’re a self-identified Podcast Person™, or are simply interested in becoming one, the service might help you find your next match.