The ultimate goal of listening to music is pleasure. But first listeners need to have a reason to discover a new song, want to listen to it again. After several repetitions they start liking the song, and getting pleasure. Once they have listened to it too often, they reach a saturation point, and want to listen to it much less often, but those few times they enjoy it a lot. There is hence an appropriation cycle that goes through several steps: discovery, repetition, pleasure, saturation.
People need to between 1 and 7 times to a song to like the song. Repetition plays an important part, as this study explain. FM radio have played with this appropriation cycle for a long time. “Even involuntary repetition, quite against our own musical preferences, is powerful”.
The most challenging step in the appropriation cycle is the first one. Why would somebody listen to an artist he doesn’t know ? just out of curiosity ? If somebody does not know more about a new artist, he has no desire to listen to him. It requires a context, which can be several: a recommendation by influencers he trusts, including friends, the recognition that the artist belongs to genre/sub-genres he likes, knowing more about the artist (a biography describing the background of the artist)/the (new) release (reviews), whether it is new and trending...
When music is played for more functional reasons (meditation, music on background for cocktail,...), it is less difficult to introduce new artists.
The way people interact with music can be very different. First comes the level of interest they have for music. A study carried out in the UK in 2006 has divided music listeners in 4 groups: 7% are savants ("for whom everything in life seems to be tied up with music"), 26% are enthusiasts ("music is a key part of life but is balanced by other interests"), 32% are casuals ("music plays a welcome role, but other things are far more important"), 40% are indifferent ("would not lose much sleep if music ceased to exist").
The more people are interested by music, the more engaged and active they are, and the higher are the number of their interactions during a listening experience.
Why would somebody listen to an artist he doesn’t know ? just out of curiosity ?
A key factor also is the age of listeners. Combined with the level of engagement, it can provide a powerful segmentation tool: this study has identified 29 segments of listeners, according to 4 levels of engagement, and 11 ranges of age. This classification relies also on how people discover music, which media (TV, radio, press, internet sites, blogs, pure players) is instrumental for discovering a song.
Education plays also an important role. This study carried out in Spain shows how age, gender and education explain preference of music genres and the number of genres listened, and identifies 12 groups of listeners. Education is the main instrument for increasing diversity of the music listened.
People can listen to music in solo, in background, at parties, with friends, and may rely for recommendations on specialized or mainstream media, online/broadcast media, or friends, and develop their music taste on their own or like the community.
This study, done in France on young people aged 15-25, shows how various socio-demographics and behaviours towards music interplay. It identifies 6 groups, and we give here a selection of the most salient features for each group:
behaviour: music listened in solo, friends as influencers, genre = rock, k-pop; artist = One Direction, Green Day
demographic: female, high education of parents, practise of an instrument
behaviour: follow artist news, genre = French rap, artist = Maitre Gims, Booba
demographic: male, parents with low education
behaviour: genre=pop, electro; artist = Stromae, Maitre Gims; influence from main traditional media (TV, radio, press)
demographic: female, 23-25
behaviour: music listened in background, genre = pop, electro; artist = Pharrell Williams, Daft Punk; friends and traditional media as influencers
demographic: parents with high education
behaviour: genre= rap,electro; artist = Eminem, Rihanna; influenced from social networks
demographic: male, 15-18
behaviour: listened at parties, genre= electro, dance; artist = Avicii; influenced by main traditional media
demographic: parents are farmers
There are several limits for listeners to discover new artists and songs: the time they can dedicate to the listening activity (more limited for “indifferents” and “casuals”), the need to contextualize the discovery (understand and decide whether it is compatible with their music universe), curiosity limit (outside music listened by their own tribe), immediate pleasure to listen to known songs versus wandering outside their confort zone.
While getting older, listeners are more subjected to these limits: diversity of genres diminishes and people discover less new artists, and tend to listen less to new popular music.
This study shows that people start to listen less to new popular music (as measured in terms of "hotttnesss", a metric designed by The Echonest, measuring "how much buzz the artist is getting right now") in their mid 30s, with differences between men and women, and people becoming parents.
At the time Musicovery was operating its smart radio we did several analysis to categorize listeners. We use multiple metrics like listeners curiosity (the more curious, the less popular are the favorite songs), the genres diversity of their favorite songs, their skip rate, their preference between artist radios versus mood radios, the popularity of artists of their search and artist radios, the need to repeat plays for new songs, their preference for new releases...
We used classification models to decide the appropriate number of groups and their characteristics (see below the dendrogram of a Hierarchical Ascending Classification done on 100K listeners) for whom to offer some kind of personalization. It helped us decide for our benchmark of recommendation systems the number and types of personas to use to test the music platforms.
"we did several analysis to categorize listeners...it helped us decide for our benchmark of recommendation systems the number and types of personas to use to test the music platforms."
image source: Musicovery - clustering listeners profiles
Music plateforms have been providing recommendation systems to help listeners discovery new artists and songs. But to which extent are they used ? According to this research paper by Beuscart/Coavoux/Maillard, people using on demand streaming services explore the catalogue for 41% of songs played and listen to their personal library (artists added as favorite, songs as favorite or added to personal playlists,...referred to as "stock" in the table below) for 59%. While exploring, they are less active than passive: autonomous exploration (songs and artists search) accounts for 40% of exploration, while guided exploration (recommended playlists) accounts for 60% of exploration. Algorithmic recommendations represents 8% of total plays. It ought to be investigated whether autonomous or guided explorations lead to songs/artists/albums/playlists being added to personal librairies at the same proportion.
Recommendations tend to expose to less popular artists, especially if recommendation is algorithmic. People resorting to active listening approach tend to play again more new artists/songs that with a passive approach. Recommendation systems are more used by people listening to blues, soul, dance and jazz than musical, soundtrack and rap.
To sum up, the route to pleasure takes diverse paths depending on listeners. We will see in the next post how recommendation systems respond to this diversity.
We just sent you an email. Please click the link in the email to confirm your subscription!