About the project, material and methods
The research project has been ongoing since autumn 2010, when we received funding from Telenor Group to do a pilot study. In 2011 the project received full funding from the Research Council of Norway for four years. In addition, Telenor contributed research-directed resources to assist with the analysis of streaming data and also provided support to hire research assistants.
The empirical material of the project is both sprawling and varied. WiMP Music (now Tidal) granted us access to streaming and search logs from all anonymised users of the service in Norway over a total of 72 weeks from 2010 to 2013. Another key empirical source was the focus-group interview; we conducted 23 of them with a total of 124 informants, ages 18 to 59, between 2010 and 2013. In-depth interviews were also conducted over several months with 12 other informants (via diaries, interviews and analysis of activity on last.fm and social media). We interviewed people who work in various areas of the music industry, ranging from the artists to players in the streaming services. We gathered material from social media platforms in relation to festivals and concerts. Lastly, master theses written within this project have applied methods such as cultural probes in studies of the interfaces and designs of the mobile services offered by Spotify and WiMP, and observation studies at local festivals.
We are very grateful to WiMP Music for sharing streaming data with the project, to the Research Council of Norway for their generous grant, and to Telenor for making the pilot study possible and providing key support in the early stages of the project. We would also like to extend our gratitude to all of the informants, who offered their time to provide us with insight into what streaming and live music mean to them.
Clouds: streaming services and online music
Everyday life
Streaming facilitates a variety of applications and has become an integral part of everyday life for the streaming users whom we interviewed. Over the course of the project, the term ‘streaming service’ became one of the most commonly used neologisms, and, in fact, music streaming is currently the most widely used music medium in Norway, along with radio.
Most informants’ main motivation for using streaming services was their access to a vast library of music wherever and whenever they wanted it. Among the WiMP subscribers, we found distinct patterns in the way music was used at various times of the day and from day to day. Friday and Saturday evenings, for example, represented prime times for weekly streaming. Weekdays between 1 and 3 PM were dominated by playlist listening, and offline listening via mobile phone was typical during commuting hours. Discovery of new or unfamiliar music typically took place in the evenings, when searches and playlist creation were more frequently undertaken.
On average, playlists occupied one third of the total time users spent listening to music on streaming services, and they represented the dominant mode of usage by people under age twenty-five.
Creating and maintaining playlists were important activities for many users. Playlists were structured according to the individual user’s logic in terms of both content and structure, and they were experienced as either collectibles that people possessed and displayed or ephemera that were created, deleted or continuously updated.
In a sample of 250,000 user-generated playlist names, up to 82 percent of them were unique. Of the 100 most commonly used playlist names, across all user groups, 35 percent were artist names, 29 percent were editorial lists, and 26 percent were based on listening context or event (e.g., workout, party, relaxing, homework, sleep, birthday). Genres only accounted for 6 percent of the most common playlist names. Diary studies and interviews indicated that the content in user-generated lists was often sorted according to personal categories and contexts rather than traditional genres and labels.
Over the course of the project, the mobile phone clearly emerged as the most important platform for streaming. Its usage differed from other platforms in crucial ways: it was used more often when people were on the move; most mobile-enabled streams were locally stored (offline); more mobile listening involved previously streamed music; there were far more searches; listening sessions were shorter; and there was much more skipping and playlists tended to be edited less. It is apparent that the mobile phone has enabled the integration of music into everyday life. Mobile music experiences varied in intensity, function, character and degree of attention demanded, but they were nevertheless acknowledged as very important.
WiMP users were nearly as likely to listen to songs in album mode as to a succession of individual songs. However, there were clear differences in listening patterns according to age and gender: young women primarily listened to single tracks in succession, for example, whereas men over age forty-five mainly listened to songs in album mode. The sequence of songs in an album or on a playlist significantly influenced a song’s streaming statistics as well. Overall, track #1 songs were streamed three times more than track #2, and nine times more than track #10, by WiMP users.
Sharing songs and playlists occurred far less than anticipated. Most of our informants experience streaming services as a personal rather than social music medium. More than nine out of ten playlists in the WiMP database were not shared with others and therefore evidently created for personal use alone.
Events and eventisation
Events and happenings greatly impacted the daily use of streaming services, including, for example, the passing of notable artists, the sharing of symbolic music following the Oslo terror attacks, important concerts such as the Øya music festival and significant album releases. At these moments, an ensemble of traditional media, social media, streaming services and face-to-face-communication contributed to large groups of people streaming the same songs and artists at the same time.
So-called eventisation in music has clearly become more widespread with the introduction of streaming services: individual users may now search vast collections of music in relation to almost any event.
On one hand, this contributes to a collective and shared music culture within which people talk about the same events and artists around the water cooler in an era otherwise characterised by fragmentation and abundance. On the other hand, so many large events make it difficult for artists to gain attention in such an information-rich musical culture.
Discovery and fascination with the new
Streaming services encourage users to find new music but thus far have not acted to facilitate the navigation of one’s personal listening history or archive. As a result, users have developed personal navigation and orientation practices to save and access their own listening history and favorites, through, for example, the use of playlists
WiMP subscribers listened to a wide variety of different music—on average, approximately one hundred different artists over just a nine-week period. This pace varied only slightly over the four years of the present project. Interestingly, we also found that the individual listener, year over year, primarily listened to music that s/he has not streamed before. Nearly one third of the users only listened to new artists during a nine-week period, when compared to the same period of time in the previous year.
Curated playlists were very important to WiMP users as well. More than 40 percent of the playlist streams that were new to the given user originated as curated playlists.
Whether people preferred curated, algorithm-based or social recommendation systems through which to discover new music largely depended on the context in which they played the music, their musical interests and various identity issues and traits.
Revenue share and concentration
The way in which subscription fees to the streaming services are shared and distributed among artists is complicated. Users with an extremely high activity level had a greater influence over how money was distributed than those who were less active. Even though two users might be paying the same subscription fees, the money paid by less-active user will go to the artists that the more active user listens to.
As a result, the so-called ‘pro-rata model’ of streaming services tends to favour certain genres, artists and formats (e.g., short songs over longer works), unlike the model for sharing revenues from physical CDs and LPs, where rights holders were awarded a percentage cut of each item sold.
When we calculated how the real-life usage of all users in Norway would influence revenue shares between the two models, the alternative distribution model—where money paid by the individual user would be distributed to the artists to whom the user listens—would benefit two out of three local artists. A larger share of the revenue would also fall to the most-streamed artists, and especially the most popular local acts.
Even though users listened to a lot of different music, there was a tendency towards heavy concentration in streaming-service usage. Over a nine-week period, five thousand artists accounted for roughly 90 percent of all streams, and 1 percent of those streamed artists received nearly 80 percent of the revenue. In addition, more than 90 percent of the streaming service’s catalogue was not streamed at all during this period.
The tendency towards concentration in music listening is also found in societies that are not dominated by streaming. This may be due to the enormous availability of music in general, with a tenfold increase in titles in the past two decades, without any corresponding increase in the number of users. If a song or artist happens to win public attention, the effect tends to be cumulative, while many alternatives, in turn, are left in the dark.
The editorial intermediaries in streaming services can contribute to this concentration by promoting or validating already trending artists, but they can also counter it by foregrounding small or local artists. Interestingly, while the share of Norwegian music streamed on Spotify in Norway is just above 10 percent, the local share is around 25 percent for WiMP, presumably largely due to that service’s dedication to promote local music.
Concerts: Live music in the online age
Artists: Studio technology to the stage
Music technology that has long been used primarily in the studio for recording, editing and manipulating sound has now been digitized and made mobile and can be used actively and creatively on the stage. This affords new creative opportunities but also challenges.
Interviews with artists in technology-heavy genres such as EDM and improvised ‘live electronics’ indicate that studio technology is increasingly being applied on stage, but that there are complications, artistically as well as practically. We encountered a great deal of reflection regarding the ways in which the connection between artists and their music is created. This is reflected in eternal discussions regarding what to play on instruments and what to play on the computer.
A distinction arises in this regard between artists who are working with a studio recording and artists who favour improvisation instead. On stage, artists will work on upscaling or downscaling elements from their recordings, sometimes evoking the creative process from the studio, other times prioritizing elements that align with the audience’s activity at the live concert (singing along, dancing, and so on).
In addition, those artists who rely on improvisation tended to apply studio technologies to the development of their live creative opportunities, to the extent that their technical set-ups function as traditional instruments and even equally natural extensions of themselves as musicians. Others like to be surprised by the ways in which technology can impact their performances.
Audiences: The live experience should be unique and surprising
The value attributed to live music appears to be increasing. What it takes for a live event to be a unique and intense experience varies greatly by genre. Whether or not musicians actually perform live is not necessarily a determining factor.
Focus-group interviews with people attending the Øya festival indicated that the live concert was generally expected to be a unique and surprising experience. In a situation when recorded music is available in abundance everywhere at any time, live music may be assuming a complementary role as a rare, unique and intense.
Concerts have always been experienced as something happening here and now, and this has not changed much with the introduction of digital technology. Whether pre-produced elements in the live concert compromise the perception of ‘liveness’ varies greatly by genre. An important aspect of this calculation is the relationship between the energy seen in performance and the sounds that are heard there, with regard to, for example, gesture, number of performers on stage, and the overall balance of the soundscape. Digital music technologies have introduced greater ruptures between the visible and audible elements of concerts, and the audience will generally accept them, as long as the most significant parts of the music are visually accounted for. Instruments and musicians, in turn, take on an almost symbolic function, guiding the audience’s attention towards key sonic elements rather than physically playing all of those elements.
Strong logical breaks between what is heard and what is seen can weaken the live experience, but what is perceived as acceptable with regard to the use of technology both on stage and in the studio keeps changing. Surreal effects tend to become naturalised and supply the new norm against which later effects are measured.
The interplay between cloud and concerts
The Øya music festival provided a central case for the exploration of how a large event might influence streaming patterns. We investigated streaming over a period of seven weeks surrounding the festival during four different years, in tandem with a control period in late spring of each year. Focus-group interviews and social media also provided us with insight into what the festival meant to users and why they streamed artists appearing there.
Festival effects
We observed a large increase in the searching and streaming of Øya artists during the festival period. Streaming of these artists in the weeks surrounding the festival was over 40 percent greater than during the control weeks. The increase in streaming of festival artists was particularly pronounced for users who lived near the venue. On average, 15 to 20 percent of all daily streams by Oslo residents during the festival involved artists performing at the festival.
In addition, the streaming of Øya artists impacted the listening patterns of many users beyond those attending the festival, indicating a general trend towards eventisation in relation to music-streaming preferences and inclinations.
Pre- and post-listening
Many people used streaming services to familiarise themselves with the music that they expected to hear live at the festival. This pre-listening practice was dominated by playlists, and especially those curated by the WiMP editorial team, in collaboration with Øya festival organizers. These editorial playlists account for almost 90 percent of all of the playlist streams involving festival artists.
Pre-listening, then, has increased with the spread of streaming, and it has become a unique mode of listening because it is directed at a future live music experience rather than a present life experience. It allows listeners to become acquainted with lots of festival artists in advance, or dive deep into the catalogue of particularly anticipated artists, with unprecedented ease and speed.
Streaming during the days and weeks following a festival saw the replacement of those editorial playlists with streams by single artists and personal playlists. Post-listening, then, functions as a personal extension and processing of prior musical experiences. This listening mode also impacts whether or not a live performance will produce new and committed fans.
The transition from the live experience to post-listening can be jarring, if only because listening to a recorded version of a song often pales in comparison to an emotionally strong live experience. Prior to the availability of music through streaming, however, post-listening was limited to whatever relevant CDs one could find and afford with artists one had had an especially strong live experience with.
Mobile phones and social media at concerts
One of the most notable changes in audience behavior over the last decade involves the use of mobile phones at concerts. Informants displayed a great deal of ambivalence towards this phenomenon, however. Many wanted to share their experiences with people who were not there and also liked to document the highlights. Others thought that mobile use wrecked the aesthetic experience, often when it was at its most intense, perceptually, psychologically and socially.
Streaming and social media have entirely different roles in relation to concerts. As discussed, streaming typically takes place in the days and weeks leading up to and following a concert, whereas social media has a distinct ‘here and now’ character that is most often exploited during or immediately following a show.
The significance of proximity
We found that live concerts impacted online activity much more for local artists than for international acts. The increase in local-artist streaming started earlier and lasted longer than it did for international artists. A succesful concert with a local artist resulted in a particularly long streaming tail.
Interestingly, among the one hundred artists in the WiMP Music catalog with the most dedicated fans (the highest so-called passion index), one third were local artists.
Music distributors and vendors in Norway have reacted to increased global competition online by strengthening their local profiles and capitalising upon their close relationships with Norwegian artists and proximity to audience groups. Organising and supporting local concerts and festivals, covering these events in various ways and creating exclusive local content are all examples of strategic moves being made by both music-streaming services and record-store chains.
Intermediaries hence attempt not only to offer access to music but also to provide a particular and strong experience with music. Record-store interest in renewing vinyl records as media for unique musical experiences is part of the same trend.