These days, you can discover music you like, or might like, in a batch of different ways. You can discover in the ‘old school’ way by finding a radio station that hovers around your style. You can build a Pandora radio station or two that finds artists ‘like’ the ones you like. You can click on the ‘Genius’ feature in the iTunes Store to get recommendations based on what you already own. You can visit Amazon to find recommendations, as well.
But all of these approaches demand some concrete content input from you. They also assume that you can identify or describe what you like, or might like, working from whatever bundle of music you’ve already experienced.
Musicovery, on the other hand, will build a playlist based on your current mood. Drawing insights from the likes/dislikes of its many users, and from the analysis of the service’s music experts (drawing on 40 ‘acoustic parameters’ provided by their experts, their algorithm arranges music on the ‘mood pad’…of course). Then, as you like/dislike the items it plays for you, it can learn more about what you prefer to hear.
There’s lots of interesting stuff to learn from this experiment. One is how many ways we can organize and arrange content to connect to an audience. The old printed season brochure required a small number of clusters for a performing arts season, for example (chamber, vocal, jazz, Broadway, whatever). The web and automated systems can add a thousand different clusters, many of which may not be defined or informed by the presenter.
Beyond that, Musicovery suggests what we might learn from the aggregated choices of our entire audience. They’re beginning to play with the data to find emergent patterns about what people like, and how those selections are similar (many emphasize wisdom, compassion, love, and peace, as it turns out). Just take a deep dive into their interactive scatter chart of ‘liked’ music through the decades.
For the information and interface design wonks among you, Musicovery seems to combine two approaches to complex data or data services: progressive disclosure and faceted search (thanks Drew). The first approach (which I’ve blogged about before) limits the user’s choices to those he or she might need in that moment. The second approach offers data or discoveries through every possible facet of their content (author, title, genre, instrumentation, acoustical qualities, mood, and such). How would your organization’s ‘interface’ look if you also embraced those two approaches?
I’m reminded once more of the motto of the once-ubiquitous and -profitable corporation MUZAK: “Experts in the physiological and psychological applications of music.”