MSI grant awarded
Post date: Aug 10, 2018 12:46:35 PM
Hannes Datta, Bart Bronnenberg and I have been awarded an MSI research grant on the impact of music streaming on the similarity of listening behavior across consumers. Summary of the project is below:
Streaming platforms offer consumers massive variety, e.g., Spotify’s library contains over 30 million songs. Users are unlikely to know about all available titles and need to search for music that matches their tastes. A natural question is to what extent the platform lowers these consumer search frictions in a way that caters to individual, idiosyncratic tastes. Conversely, the streaming platform may have incentives to promote the same music to everyone, effectively concentrating listening on a handful of artists; for example, the most popular playlist on Spotify, “Today’s Top Hits”, has 18 million followers. Given the rapid growth of these platforms, and concerns about their increasing economic power over producers and consumers, the extent to which platforms encourage uniqueness or promote homogeneity in consumption is an important question. Platforms have formidable power to shape the search process, potentially directing consumption away from the most efficient consumer-product matches. Now that Spotify is bypassing labels and beginning to license music directly from artists, it may have a strong incentive to direct search to its own artists. Consumers likely have highly idiosyncratic tastes for music; hence, we propose to examine whether platforms make pairs of consumers more similar or dissimilar in terms of their music consumption. We investigate changes in listening similarity for pairs of consumers around the time of their adoption of Spotify. Because incomplete knowledge contributes to concentration, we also plan to ask whether the effect of adoption on similarity differs for titles with which consumers have some knowledge about (e.g., have consumed in the past) vs. titles with which consumers are likely to have no knowledge about (e.g., new-to-the-world titles from new-to-the-world artists). Lastly, we plan to investigate similarity over several dimensions (e.g., titles, artists, genres) and over time.