This morning I attended a SXSW panel filled with leaders from Shazam, Sirius XM, and other digital music platforms that hoped to tackle the tough subject of music discovery and where it may be headed in the future. It was a continuation of a panel from 2014, but the content shared was wholly unique to this session.
The first topic was the rise in demand for professional curators. Apple is hiring some right now, but there is an argument to be made for whether or not machines can do they job just as well. The team from Sirius XM commented that this trend is good, but they have always made it a point to use both man and machine for their curation needs. Shazam seconded this notion, and added that their path to discovery begins with user generated questions. Users hear a song they like, then engage in discovery by opening Shazam to figure out who is responsible for the track. To help make a connection between personal discovery and more large scale music discovery the company recently added charts, which shows every user the songs most often being sought on the Shazam platform.
Other panelists had variations on the comments above, but almost everyone agreed the most difficult part of the curation and discovery process right now is how to expose people to talent they may enjoy without doing a hard sell. For Rdio, one of Spotify’s biggest competitors, the changes being made to better facilitate discovery began with the launch of their own radio tool. Users can choose the song or artist they like, turn on radio, and the service will attempt to serve users with artists and songs they believe to be similar to the user’s initial choice. Who makes this cut is based on cumulative data from all users, which is fed through an algorithm to find what artists/songs are most often played together.
When asked how they plan to take this idea and develop it further, Rdio mentioned personalized radio stations for users that take into consideration everything they listen to, as well as how often each song/artist/genre is enjoyed. Other panelists agreed that this idea was the way of the future, though each has their own take on how it should work. Again, everything comes back to data, and it seems that will be a constant theme throughout SXSW this week.
If we accept that data is the key to curation and digital discovery, which the panel in this session did, then we next need to discuss what parameters need to be placed on the data being used. People who use applications like Shazam or Rdio have possibly compiled years worth of data about their music interest, so how far back should these companies dig through that data when trying to determine new talent to show that user? No one seems to have an answer quite yet, but almost everyone agrees there needs to be some kind of cut off point. People’s tastes change over time, so it doesn’t make sense to use the entirety of your time with a service to generate new recommendations.
8Tracks founder David Porter was quick to reply to this conversation by adding that having too little data about a user is just as problematic as too much. He used Twitter’s ‘People You May Know’ service as an example, citing how the company did a poor job of recommending people you may actually know early on, but over time they refined their tools through the use of more and more data.
At this point the panel was asked whether or not users would feel comfortable with constant data mining being made in the name of content curation and future music discovery. The panelists were quick to comment that the answer will always be dependent on the way the discovery tools are presented to users. General recommendations like Prince or Madonna are only going to push people away, as most users will already know of these artists. The key is to find a way to show users lesser known artists who relate to the talent they already know without coming across and pushing a hard sell or otherwise being intrusive.
Building on this idea, Shazam commented that they have learned bad and/or poor recommendations tend to drive away users. If people get push notifications or emails that attempt to connect them with music they already are aware of they are far more likely to turn off notifications and unsubscribe from email chains than if no recommendations were serviced at all. Other panelists commented that they have seem similar trends with their various attempts to change or innovate discovery.
Before closing the panelists were asked to summarize their thoughts. There were some variations, as well as plenty of pitches for their various companies, but as a whole everyone seemed to agree that the future of music discovery and content curation is not about man versus machine, but rather man and machine finding a way to work together while leveraging the vast pools of data each business has collected. No one knows exactly what the next phase of discovery will be or what it will look like, but everyone is interested in seeing it brought to the public sooner than later. In order for these companies to thrive they need users who are constantly consuming content, and the only way to guarantee that happens is by consistently providing them with good recommendations, or at least giving them the tools to find artists they may enjoy. Once someone cracks that code, everything about content curation will change.
