Internet radio has the potential to be the most ubiquitous form of media ever. More commanding of your attention than film, television, or books. This is because listening to music can be enjoyed while doing other activities. Before I go further, let me make an important distinction: there are two types of listeners, lean back and lean forward. Lean back listeners hear music programming via a playlist or radio station (think “set it and forget it”), whereas lean forward listeners actively select individual songs. The majority of people prefer a lean back experience.
In 2014, one third of Americans used their phones to stream music. Young adults (18-24) listened to internet radio more than terrestrial. Two of the top five most popular apps in America (Pandora and Youtube) are used for streaming music. With Americans now spending more time on their phones than watching television, there has never been a more opportune time to maximize internet radio experiences.
Let’s consider the potential internet radio landscape. The average person sleeps approximately seven hours a day, meaning there are up to seventeen listening hours per day that one could listen to internet radio. Additionally, it’s expected that in two years 3.5 billion people will be online, bringing the total of possible listening hours worldwide to 59.5 billion per day. With the average revenue per thousand hours amounting to $42.77 (Pandora’s rate in 2014), there is a possible daily cap of approximately $2.5 billion in 2017. Granted, this assumes that the market rate is equal throughout the world, which currently is not the case.
For this potential to be realized, companies will need to provide highly personalized listening experiences that have yet to be fully optimized. Internet radio will need to match every part of your day. Imagine passively being pushed the right music that helps you wake up, motivates you to run faster, work more productively, and more. This type of personalization has already begun in advertising, but hasn’t been well implemented in internet radio.
However, the barriers to this hyperpersonal internet radio are slowly being eliminated. The swath of personal data that comprise our tastes is growing, which in turn means we are also able to better understand the tastes of similar people. With the decreasing costs of streaming, collecting and storing large sums of data, as well as growth of powerful tools to analyze it, our ability to explore and draw inferences from this wealth of information is seemingly endless. Lastly, the vanguard (major record labels) are now willing to make this digital shift. In the last three years, labels have signed deals that make their music libraries legally accessible on-demand, setting the stage for this transformation in digital discovery based on big data.
In particular, location, social, and biometric data streams will provide the most relevant inferences for personalized listening experiences. To date, location information has primarily served to provide music programming experiences reflective of the general tastes in those regions. Within a hyperlocal context, internet radio services could push music reflective of microcultures in neighborhoods, parks, rooms of a house, etc. Listeners that aren’t generally music consumers could still receive personalized experiences if internet radio providers could draw from their social data to find people with similar tastes in other areas such as films, foods, books, etc. and reference those listeners’ music preferences. Finally, leveraging biometric data from fitness bands could enable internet radio providers to push music that reflects the perceived activity or mood of that listener.
It will take time to gain listeners’ trust to connect these types of data streams to music services. In the meantime, music programming that relies on crowdsourced curation, similar to Twitter or Reddit, can surface highly contextual programming (e.g. mood, activity or genre). Within this paradigm, the best content will emerge because the crowd deems that the context provided around the programming meets their expectations. A different approach is accumulating data passively through smart audio devices. One example is the Cone, a speaker that learns a listener’s music preferences based on where that person listens to the speaker. If the listener prefers indie rock music when in their bedroom, the speaker will play that type of music once it realises it’s within that room.