Big Data and the Music Paradox

It is recent news that Spotify has settled a complex licensing dispute involving mechanical licensing rights, allegedly in order to clean up its affairs in view of a prospective IPO.

Mechanical rights under US law must be obtained (in addition to performance rights) when reproducing a piece of music onto a physical or digital support.  It is not clear if Spotify willingly avoided to pay mechanical rights or – as it claimed – it had no availability of the data necessary to sort out which publishers had legitimate claims over songs (there isn’t a central and reliable database covering all music rights to all songs).

Music licensing is a complex and disarticulated. Different countries have different nuances of copyright applicable to music (and different ways and means of collecting royalties). In some countries, particularly the US, the music publishing sector has traditionally licensed the performing rights and mechanical rights separately through different entities. This means that music distributors need to have license covering both the song and the recording, and both performing and mechanical rights. In the US this issue is partially addressed through a compulsory licence covering mechanical rights with a pre-set statutory rate to be paid, so streaming services are not required to negotiate terms and price with each right holder. However, often the owner can’t be identified, as while there are collecting societies that licence performing rights, mechanical rights are not represented by a single society nor is there a single publicly accessible database providing this information, therefore – according to Spotify – making it impossible (or too burdensome) for a streaming music provider to comply with mechanical rights obligations for all songs.

Whatever the reasons for Spotify’s legal lapse, certainly, it is a fact that digital distribution of music and particularly streaming needs to take a further leap forward in its ongoing legal catch-me-if-you-can race which has been going on for the past 20 years – since the time of 1999 Napster.

I have been a late adopter of Spotify but their theme-based compilations, and especially their running compilations which select songs matching your personal running beat were recently a revolutionary discovery for me.  As a lawyer and a mom of two boys, time for listening to music – or especially time for discovering new music and updating my playlists has been one of the first to disappear on my schedule, with the effect that music slowly started disappearing from my life. Yet, discovering new music and enjoying music had always been one the most fundamental and joyous artistic experiences for me.

Then I discovered Spotify. Just this morning, while I was running to one of Spotify’s compilation which offers songs matching the user’s running beat, I listened to about 15 songs that I had never heard of, of artists I have never heard of.  This certainly was not possible in pre-digital ages, where buying a tape or a CD was so expensive that you would listen to the same music or playlist over and over again for months on end until a boyfriend/girlfriend would introduce you to some new playlist of his/hers by copying it on tape or CD. And you would listen to that for months on end.  But it wasn’t so even in the iTunes years – iTunes made buying music affordable, but in order to listen to a song a user still had to know it, select it, download it. The only way to discover new music was the radio, with advertising and limited availability of choice and customisation as songs were selected by a human mind – the radio host.

With services like Spotify, music enters the realm of big data and a seemingly infinite number of music pieces are available and playlists for all tastes, moods, desires and functional needs are created by algorithmic configurations.  This is a new radical change in the dynamic of the music industry – particularly the relationship between listeners and music is revolutionized.  The magic of data-driven approach applied to streaming is that music is available in such quantity and variety, that paradoxically the relationship between user and music is disinter-mediated and direct because the algorithmic data stream allows for more nuanced experiences and choice.

The consequence of this is that an average user can access music and artists that s/he would perhaps have never considered before – and at the same time artists that would have had no clout are heard of by a greater audience. This certainly is a boost for the music industry, and for any single musician who wishes to expand the reach of her/his music.  At the same time, disintermediation necessarily bypasses those structures that had been put in place in previous eras to protect legal interests.

It is certainly necessary to create new structures where music can be discovered without getting caught up in legal tangles, while at the same time compensating an artist for beautiful music.

As data driven services emerge in the music industry, a data driven approach needs to be adopted by music societies as well. I am imagining a universal music society with a database to which artists and music labels sign up and songs and recordings are matched by a Shazam-type service (with conflicts resolved through an online dispute resolution service).  All streaming and downloading services would link to this database and payment to the relevant right-holder would be automatic and immediate. I go further by imagining different levels of payments, where for example new songs by unknown artists are remunerated based on a rating level by listeners, so that copyright compliance can be a boost to music discovery rather than a gateway to its distribution.


© Stefania Lucchetti 2017.  For further information Contact the Author

Articles may be shared and/or reproduced only in their entirety and with full credit/citation.