The gradual rise of music collaborative filtering

By Author

A piece just out in the New York Times covers the current array of music collaborative filtering services (though it doesn’t call them that), including Pandora and Last.FM. I’ve written about these numerous times in my books and blog, including my initial thoughts on discovering Last.FM in 2003 (still love it!) and a comparison between Last.FM and Pandora. The article refers to a report by Gartner that predicts that by 2010, 25% of online music sales will be driven by collaborative filtering engines. That’s a high figure, given that social networks and personal recommendations will always be at the heart of individual musical discovery, but I do agree that much of the way new music will become visible will be through these kinds of tools. To review, the concept of “collaborative filtering” is that we collaborate to filter the virtually infinite possibilities we face. This is largely done through tools that compare our tastes with those of others, so we can benefit from what people with similar taste to us have discovered. The magic of this is that it becomes far easier to find what we like (be it entertainment, information, or anything else) in a world of infinite choice. One of the most important impacts of technology has been to uncover creative talent, since access to either high-quality production or distribution is no longer a barrier. This means we have far more entertainment choices than ever before. The growth of the “long tail” is vastly enabled by software that provides recommendations for things that we love, that we would never find otherwise. These collaborative filtering services are fundamental to the way the future media landscape will unfold. Progress on these for the last decade has been slower than I would have hoped, but it is picking up, and the promise is there for all of us to find far more music that we love.