Music Streaming, Festivals, and the Eventization of Music

Abstract

A key aspect of music-streaming services is the user’s access to their vast libraries and abundant choices anytime and anywhere. This article explores how artists performing at a large music festival in Norway were streamed before, during, and after the festival over the course of four different years. The data shows that festival streams grew by more than 40% compared to control weeks and were particularly pronounced among users who lived near the venue. The article then argues that changes in listening patterns may reflect a more general shift towards the “eventization” of streaming media.

Acknowledgments

This article was produced under the auspices of the Clouds and Concerts research project at the University of Oslo. Many people were involved in the quantitative data analysis that provided the basis for the research. Research assistants Ola Løvholm and Xin Jian did the main work, in collaboration with the author. In 2010 Johannes Bjelland and Pål Roe Sundsøy at the Telenor Group set up the database, along with Kenth Engø-Monsen; they also mapped streaming to postal codes, performed the initial analysis, and helped with troubleshooting in subsequent years. Beathe Due (project leader at Telenor) was instrumental in initiating the 2010 study, as well as conducting the focus-group interviews together with the author, Anne Danielsen, and later Yngvar Kjus. Finally, research assistants Hanne Tråsdal and Marc Casanovas took part in some of the initial quantitative analyses. Gratitude goes to all of them.

Notes

1. Services like Pandora or iTunes radio do not let users search, play, or skip individual songs. So-called “Internet radio” or non-interactive streaming must thus be distinguished from interactive streaming or on-demand streaming, as is discussed here.

2. A browser version of the service is available at WiMP Music or TIDAL. See also screenshots of the service on various platforms at WiMP (WIMP Mediaroom).

3. It must be noted that placement in playlists cannot be bought, and the editorial team was very determined to cultivate an ethos of independence and authenticity regarding the music it recommends, but the service also acknowledged its shared interests with labels, artists, and festivals. Interview by author with Sveinung Rindal, Head of Editorial Campaigns at WiMP Music, 20 March 2011. See also Kjus.

4. Interviews by author with Birgitte Mandelid, Marketing Manager of the Øya festival, 16 June and 14 September 2010.

5. Official user data reported by WiMP Music in public stock exchange announcements listed 100,000 users by the end of 2010, 300,000 by the end of 2011, 350,000 by May 2012, and 520,000 by the end of 2013.

6. Kjus and Danielsen, however, draw upon unpublished streaming data from the present study in their work.

7. Further information may be found at Wikipedia or the official festival site for the Lowlands festival (Lowlands).

8. Sample scripts and further details are available on GitHub (GitHub). For most of the analyses presented here, only users with more than 100 streams over nine weeks were included, in order to avoid users who were new to streaming and to privilege those who streamed music on a daily or weekly basis.

9. Listening patterns of what one might call “festival listeners”—that is, the 20% of users in 2012 who had the highest percentage of their streams involving festival artists during the four festival days—were also compared with a random sample of the 20% of users who had no streams by festival artists (“non-festival listeners”). This analysis showed that festival listeners, on average, had 25% more streams than non-festival listeners. Of “super fans” within certain genres (that is, the top 10% of listeners with the highest ratio of streams of a particular genre), there was a particularly high correlation between festival listeners and fans of alternative/indie (25% of festival listeners were super fans of alternative), whereas only 4% of festival listeners were super fans of pop. On average, WiMP users were streaming 48% pop tracks and only 5% alternative/indie tracks at this time (Maasø, “Clouds and Concerts”).

10. An analysis of search queries during the festival week in 2011 shows similar patterns, with 24 of the top 100 search queries being for one of the 80 artists playing at the festival.

11. In 2006, Google added Google trends, with which one can track a search term’s use relative to total search volume across various regions of the world, and in various languages, over time (Google Trends). The annual top ten searches released by Google are based on year-versus-year search trends—in other words, how popular a search was this year versus last—rather than raw frequency. See also Yahoo! annual top trends (Chan).

12. Lamere did not report over what period of time these users listened to those 800 songs, however. Likewise, the study’s data came from a site where users could upload their own iTunes usage information, and it was hence likely skewed (for example, by age and gender).

13. See Maasø and Toldnes for a discussion of network effects and the way in which certain “songs of solace” after the 22/7 terror attacks in Norway spread through a combination of streaming and broadcasting.