With all the hype swirling around Spotify, the hot European startup that finally launched in America today, it’s easy to forget that there’s a whole universe of music services out there. But there are> plenty of others to choose from: from stalwarts like Pandora and Last.fm to newer businesses such as Rdio and MOG.
Arguably the smartest of them all, however, is one you probably haven’t heard of before. Clio, which is based out of Philadelphia, is a music startup that uses technology to search tracks and provide music recommendations. Since relaunching earlier this year, it’s been making leaps and bounds — to the point where today it’s announcing a significant deal with one of America’s biggest music libraries, APM. The agreement means that Clio will be powering the search system for APM, a joint venture between EMI and Universal that provides music services for film and television shows like Bridesmaids, 30 Rock and True Blood.
The reason they’re using Clio is simple: it’s an incredibly powerful system. APM president Adam Taylor characterizes it as “a fast and objective music search and discovery tool” — but that doesn’t really do the service justice. Let me explain what I find really interesting about it.
Clio is different because it doesn’t categorize music in the same way as other services. While most music systems categorize tracks by adding metadata and keywords that describe it — somebody tagging it “rock”, or information that says it’s 120 beats per minute — Clio, instead, takes the track itself and pulls it apart. It understands what founder Greg Wilder calls “the grammar of music”; the groove of the drum beat, the level of aggression, the playing style, the riffs and hooks and melodies, and much more. The technology actually listens.
After dissecting a track, it can then compare on a number of different criteria and recommend music that actually sounds like what you want. Linked into APM’s catalog, for example, means that somebody scoring a TV show who wants a track that’s dark and moody but with a driving, throbbing beat, can plug in those requirements and it will recommend a string of tracks that have those elements in common.
In this respect it’s a little like the Music Genome Project, the technology that underpins Pandora — but whereas Pandora relies largely on human interpretation to listen to tracks and rank up to 400 attributes, Clio does it all programmatically and in milliseconds.
“Humans break things into little chunks,” says Wilder, a composer and academic who founded Clio’s parent company Orpheus Media Research, four years ago. “We can hear the whole thing.”
Here’s their promotional video.
And here’s a Soundcloud wave that shows you a sequence of songs that Clio’s matched starting with a single seed track — it really shows how the system picks up elements of the music that aren’t easily definable by keywords or simple metadata.
So why haven’t you heard of Clio?
The reason is simple: it’s not a consumer product. The company is largely aiming at the professional market, striking behind-the-scenes deals that will help it become the engine for services — rather than the service itself. This is a deliberate and careful strategy: the business is trying to scale up in a way that gives it the best chance of long-term success, by creating an enterprise-level system that outstrips anything else on the market. Given the constant financial trouble that music streaming services seem to be in, that may be a smart move: only last year Om spoke to Pandora’s Tim Westergren about how close they’d been to failing over the years.
But Clio says that being the engine doesn’t mean that it wants to remain in the background forever. They’ve got serious targets — and today’s deal is, they hope, just the beginning.
“We want to provide the backbone to streaming services and recommendation all around the world,” COO Alison Conard told me. “Google has a lofty ambition to organize all the world’s information: we want to index all of the world’s music.”