The Echo Nest raised $17 million to build out its international business and perhaps expand beyond powering music recommendations. The company uses natural language processing, acoustic analysis and machine learning to build music recommendation engines based on songs, social media data and other writing about music. Customers include Spotify, Clear Channel’s IheartRadio, Vevo, MTV, AudioVroom and others. TechCrunch spends an awful long time saying it’s a better business than Pandora, because it doesn’t have to actually license music. But there are only so many music services to sell to. And so far, it’s been very hard to prove that music tastes and tastes in other entertainment categories (or clothing or merchandise) translate across formats. That’s where the non-acoustic analysis will help, but I wonder if collaborative filtering will be of any use at all.