To Compliment a Complement? Modeling New Avenues for Podcast Content Discovery

Abstract

Although podcasts have emerged in recent years as the fastest growing media form in the US, consumption remains significantly skewed in favor of the top 1% most popular podcasts – with many new podcasts largely remaining undiscovered. Among the primary culprits for this friction are the still mostly primitive and generally limited content discovery tools and features offered by today’s leading platforms, from Spotify to Apply Podcasts. In this study, we approach this multifaceted challenge in three phases. First, we conduct an extensive qualitative analysis of field study data on CUNY University students, coding interviews surrounding their podcast consumption habits and experiences with the medium more broadly. Second, we provide a quantitative analysis of the students’ listening behaviors, particularly with a view to understanding the impact of listening frequency and platform selection upon content and consumption preferences. Ultimately, in observing that the content creator is perhaps the most neglected entity in the podcast value chain and is uniquely positioned to support any content discovery intervention strategy, we develop a model that identifies complementary podcasts based on listeners’ subscriptions, content categories, and podcast descriptions. With this proof-of-concept implementation, we propose a new avenue through which to unlock content discovery opportunities for podcast creators and listeners alike.