The next Pandora will be a mashup

Hardly a week goes by in which we don’t get a pitch for a site that aggregates and filters music videos to serve up streams of personalized entertainment. The latest entrant is’s Music Video Genome, which launched last weekend. was cobbled together within 24 hours as a project for the Music Hack Day in New York, but it might just be the next Pandora. co-founder Casey Pugh told me Tuesday during a Skype (s msft) conversation that was primarily meant to showcase the power of VHX’s new API, which was also released this weekend. The site allows users to enter the name of an artist and then simply press play. immediately queries to get recommendations for similar artists and then builds a playlist of YouTube (s GOOG) videos based on those recommendations. The result is a personalized music video stream that works in a similar fashion to Pandora’s radio streams. doesn’t offer users a way to vote on each track in the way that Pandora does, but it has some possibilities for interaction that take into account any holes in YouTube’s catalog. “It’s surprising how almost every song is on YouTube,” Pugh told me. However, some of these videos simply display a song’s cover art — which isn’ really a great visual experience if you’re trying to serve up a stream of engaging music videos. Users have the chance to flag these videos and suggest links to better versions, which can be both official and unofficial versions. “If there is no official music video, maybe a fan can make one,” suggested Pugh.

There are plenty of other sites that offer their own take on this kind of mashup between music videos and recommendation data. Ryan Lawler wrote about a more editorially guided approach at on Monday. We previously covered ListandPlay and Tubeify, and just last week, received a pitch for a site called that bills itself as an “easy way to listen (to) music from YouTube” and that makes use of the YouTube data API to filter out any non-music content.

I wrote back in 2007 that mashups are becoming a sort of litmus test for startups. Back then, I asked: “Is there really a business model for a startup if someone else could achieve the same thing with a quick mashup?” That’s still a valid question, but in the case of Pandora, there’s more to it: The company has built its streaming service based on the DMCA’s definition of non-interactive music services, which means it has to adhere to some strict rules to avoid direct licensing negotiations with record labels.

Pandora’s users can, for example, only skip songs six times per hour, and the company can’t play more than three songs of the same artist within a three-hour period. All of this allows Pandora to pay rights holders through collective licensing, but even those fees add up to millions per year, and fee changes brought Pandora close to bankruptcy in 2008.

Mashup makers, on the other hand, don’t bother with any of these rules, but simply take publicly available videos and leave the licensing details up to YouTube and others. That kind of behavior may scare rights holders, but Vevo’s success and Google’s monetization of fan-made content have both demonstrated there’s money to be made in music videos — and mashups like could play a big role in this type of monetization in years to come.