My first and still favorite music recommender system, Pandora has been struggling to survive and has had several near death experiences over the past 10 years. So I’ve been delighted to see a flurry of reports in the past few weeks about how things have recently changed for the better for Pandora. See the list at the end of this post for several of these recent reports. Perhaps what I found even more valuable was the larger lessons which emerge from the Pandora story such as how to survive by being agile and adaptive and how perseverance, both on the part of both the founder and CEO Tim Wetergren and all the many employees who stayed with him through the troubled times and in many cases going years without paychecks, really does win out.
Pandora was one of the first tangible examples Erik Duval and I seized upon and used when showing others what The Snowflake Effect looked like and proof that it was possible. Pandora truly lived up to its mythical name and let the Snowflake Effect genie out of the box for good for Erik and I.
While I continue to watch and experiment with many other music recommender systems such as Last.fm, Slacker, and Spotify, I’ve always found that Pandora does the best job of helping me find just the right songs at just the right time, which is at the heart of the premise and the promise of the mass customization and personalization of The Snowflake Effect. Others prefer to find music based on the tastes of others by adding and mining data from social networks to their tools, but for me I’m looking for music that matches my tastes, moods and context and Pandora, with its Music Genome Project database containing intricate details of each song, seems to do do this best.
* See “The Song Decoders” NYT article for an interesting insight into the people doing this work.
Music is but one of the almost infinite areas where we are seeing the transformation from a culture of mass production to one of mass customization and personalization. Experiencing the difference between finding songs and artists via a service such as Pandora compared to something like Top 40 radio is one of the most compelling ways of seeing this transformation which will help you see just how powerful and how real this transformation is and will be. We are already seeing similar examples of how we can have similar “decision support” in finding the best books to read, sites to visit, people to talk to.
If music recommender systems are new to you or it has been a while since you have tried them, I’d encourage you to check them out and see if they don’t help you find just the right songs, just for you, and in doing so give you a glimpse of the Snowflake Effect future before us.