Demystifying Streaming Music Data

What happens when a user streams your music on Spotify or Apple Music? What happens when somebody asks to Amazon Echo “play jazz music”? What happens when your song is added to someone’s playlist?

Music is played, royalties get paid and data are created. Every click is tracked and most can be pinpointed to a location. That “someone” becomes a male user located in Paris, aged 39 years old, using MacOs who listened 23 times the same song for 160 minutes of play, a song he added to his favorite playlist shared with hundreds of people.

These data are sent from Digital Service Provides (aka DSPs, Spotify or Deezer) to intermediaries such as aggregators (CD Baby or Tunecore), publishers (TheOrchard or AWAL) but also directly to artists (Spotify Fan Insights). Most of these data are free and accessible through Application Programs Interfaces (APIs), others are not. Websites such as Soundcharts or Chartmetric gather, store and display these APIs-based data on a daily basis. Some data from Apple Music are only available to aggregators or publishers and not to Artists. Spotify and YouTube also have similar insights which are publicly visible. 

APIs create a digital-synchronized-transparent world. We need transparency in the music industry. But transparency can create an avalanche of data, a situation of information overload. What information consumes is the attention of the recipient of this information wrote 78′ Economics Nobel Prize, Herbert Simon.

Because Artists (and 99% of human beings) are not Excel-fans, more data would mean more headaches and less time to make music. Especially because some of these data can be purely quantitative (a number, a stream) or - more difficult - it can be qualitative (a comment, a word). Artists willing to track their own data would spend a large amount of time pulling it from many sources and understand trends, causes and effects. Here comes the layer added by some aggregators and some publishers: the analysis, the statistics and the display of data - what we call the analytics.

Myth 1: Analyzing data = too complicated and time consuming

Analytics are available for most music streaming platforms and it’s important to analyze these stats and use them to focus marketing campaigns. It continues to evolve and reveal very interesting and important data about best-performing songs and locations of highest-engaged listeners.

SoundCloud has been gradually putting more emphasis on playlists to match the strategies of its rivals: something that’s clear from the latest update to its analytics for musicians on the service. Spotify has been offering this for a while in its fan-insights dashboard, but that service’s heart is its own in-house playlists rather than those from external playlisters.

Ari Herstand has done an extensive analysis of the differences between music aggregators here. From my perspective, the more you pay for using intermediaries, the better are the analytics. Platforms are bringing new features such as insights or recommendations to help Artist making the right decisions. The AWAL App has done a pretty good job on this matter, displaying key data in short sentences and I quote (from the website) "easy-to-understand graphs and charts".

Myth 2: More streams = more fans and more gigs

Artists can now understand where and who the listeners are. But most of these people are not necessarily ready to come and see these Artists on stage. A difficulty that this company based in South Korea is trying to solve: MyMusicTaste. This website is a crowdsourcing platform that gives fans the power to organize concerts in their city. Fans willing to bring Artists on stage have to undertake some missions that will help the promoter and the band to take less risk in organizing a gig there.

A field study done in my data analytics class proved that each social media or DSPs is working independently from each other (silos effect): a spike in Spotify streams would not generated additional likes on Facebook or pictures on Instagram. With my students, we came up with the conclusions summarized in the table below:

This result is pretty consistent with Ari′s observation in his book:

Some artists have hundreds of thousands of followers online, but can’t get 15 people out to a show in their hometown. Go where your fans are. Understanding who your fans are is crucial to figure out where you need to be. How do you do this? Well, be everywhere at first and analyze the analytics and insights the platform provides. It’s too tough to master every platform. So pick one that you go all in on. This doesn’t mean ignore the others; it just means master one.

Myth 3: Analytics = Marketing

This is not only about knowing who and where the fans are but to understand how to collect the money you claim.

The process I described in the beginning of the article is far from being perfect due to metadata and databases flaws. The digitalization of music has created holes in the music value chain. In 2015, my institution, Berklee College of Music, published a study named ¨Fair Music: Transparency and Money Flows in the Music Industry¨ to understand the black box of revenue distribution between artists, distributors and intermediaries. The report's claim that 20-50 percent of royalties do not reach their rightful owner.

One of my colleague, Melissa Ferrick, has been pretty active in filing a legal class action that includes hundreds of songwriters – including Black Keys' Dan Auerbach, Pixies founder Kim Deal against Spotify. The company complains that in the absence of a global repertoire database, paying the writers is very difficult. Spotify used the Harry Fox Agency (HFA) to pay writers, but the HFA database was incomplete, the lawsuit alleges – short around 100,000 songwriters and 8 million compositions. Now in a fresh lawsuit filed in september, six publishers argue that Spotify’s failure to notify or negotiate amounts to a compulsory license:

In the USA, Spotify did not build proper infrastructure or require sound recording rights holders to provide data as to what specific composition the sound recording embodies.

Spotify could have avoided litigation by asking for the composition metadata from the record companies it needed to pay the writers – and making sure they got it. Or use a service such as Paperchain. According to the co-founder of that company - a former student of mine at Berklee - around 50 million of dollars of unclaimed royalties are filled by DSPs unable to reach the right owners last year in the USA. This process is called a Notice of Intent and is publicly available on the website of the American office for intellectual property. The result of this study has been the creation of Paperchain’s Compo Record ID (CRID), a universal musical rights identifier that will align compositional and recording identification systems.

I sincerely hope that the future will bring more solutions to Artists willing to compose, perform and share their music in the digital age...if not, next time will ask Siri to play "jazz music" nobody is going to get paid...

Note: this article has been written with the perspective of the panel I moderated at the music biz conference organized by BBC Worldwide in London (UK) the 8th of October 2017 (see picture)