The early aughts will be remembered for the number of songs we could fit in a pocket. At the beginning it was a thousand, Apple co-founder Steve Jobs holding a just-announced iPod on stage. Towards the end, the likes of Rdio and Spotify were boasting millions, songs stored not locally but beamed via wireless signals from the remote, limitless cloud. My, how our pockets have changed.
But in 2014, we don't care about pockets anymore. The digital song catalogue arms race of music streaming services is over. Everyone, more or less, has the same stuff – too much of it, arguably, an abundance of choice. And so the conversation has shifted now, as it always does, to value, and ways that we can get more of it out of the albums and songs and artists that everyone already has access to.
Value can mean a lot of things, but in this context it means recommendations and discovery and all of the things that aren't music, but relate to it. Algorithms and human curators are already sifting through these massive libraries so we don't have to, and finding other things that we might like. On basic level, that might mean serving up songs that sounds similar to the latest Beyonce single, or knowing to exclude artists that sound similar to that one band you hate. But context is also about more than just relationships between one recorded sound and the next. It's all of the other stuff – tour dates, merchandise, photos and biographies that live on other websites, within other apps, and have been divorced from digital files and streams – that matter now more than ever to companies and listeners too.
In other words, if you're a music streaming service, it's not so much the music that matters anymore. It's the metadata. And the next big thing will belong to whoever can make sense of it all.
In the early days of building corpuses of music metadata, there was arguably nothing better than Last.fm. It was a social network, recommendation service, analytics service, concert calendar and more, all in one. For music. If you listened to that new song by The Strokes in iTunes – hey, this was the early-mid-2000s, after all – the Last.fm app on your desktop would take a little note, add it to your play history, but also display user-created biographical information and photos, concert dates, purchase links, music videos and more. It was eye-opening at a time when iPods were still mostly greyscale, music was still ripped from CDs, and getting album art into iTunes was a dark, dungeon-learned magic.
Last.fm still exists today, but has since fallen into a state of neglect – its model of user-generated data unable to keep up with the explosion of up-and-coming Bandcamp punk, Soundcloud DJs and hip-hop mix tapes. Which is to say, there are lots of holes, and some of what's there is woefully out of date. Montreal-based software developer Greencopper built a mobile app for this year's NXNE music and arts festival that was supposed to pull in artist data from Last.fm, but found that entries for 50 per cent of the bands playing at this year's festival simply didn't exist.
Typically, festival organizers acquire photos and biographies from the artists directly. But this was a problem for a whole different reason, explained Greencopper CEO and founder Gwenaël Le Bodic during a talk at one of the festival's Interactive sessions last month, because metadata was fundamental to the app. If a band didn't exist in Last.fm, then it was harder for the app's recommendation feature – based on a concertgoer's past Facebook likes – to match that user up with similar bands.
Metadata can make or break an experience now, and no one wants to be caught without it. Take Google's recent purchase of the streaming service Songza for a reported price of $39-million: The service is a lot like Pandora with radio-style streaming, but relies heavily on human playlist curation, music discovery and recommendations. On the opposite end of the spectrum is The Echo Nest, easily one of the biggest names in music intelligence, due to its wealth of fine-grained metadata on everything from a song's beats per minute to its key, all of which Spotify acquired back in March to improve its algorithmic recommendations.
In both cases, no longer is the focus on the size of a service's music library, but how to make that library most relevant and useful to each user. Knowing what bands play the same genre of music, in the same key, the same instruments, whether a band sings happy songs, sad songs... it's all metadata, stuff that can be categorized and tagged. But we're arguably just scratching the surface.
Beats Music, recently acquired by Apple and not yet available in Canada, is perhaps the most forward-looking service of them all when it comes to metadata – an on-demand service that doesn't put your library front and centre, but instead focuses both on curation, recommendation as well as a full-blown social network with artist news and reviews. It literally cannot function without metadata, which is splashed across every part of the app – details that help Beats tell you not just what to listen to, but when you can buy tickets, and where you can buy an artist's merchandise. It's not perfect by any means, but Beats Music aspires to be a full-service experience that elevates streaming above the iTunes-style library model we've had for so long.
Le Bodic called out yet another startup called OpenAura, which isn't actually a music streaming service, but caters to them. OpenAura launched in beta this past spring, and describes itself as "the only platform where artists control and make money from their digital identity – amazing photos, artwork, info and more – giving fans a deeper, dynamic experience on music apps and services around the world." It's sort of like Last.fm, but geared more towards artists than listeners (though it's anyone's guess whether it will succeed).
In essence, all of these things relate to music, but they aren't music. Music is just the scaffolding on which the future of music consumption is being built. We don't usually think of things such as concert listings, interviews, merchandise and photos as metadata, but in this context they most certainly are. And they'll soon make all the difference.
In early June, The Wall Street Journal published an article titled "In a Single Tweet, as Many Pieces of Metadata as There Are Characters." The basis was that our pithy jokes and tweets about cats have more information attached to them than most of us realize – location metadata, the ID of the recipient, the number of favourites, and the number of people who follow you.
You might argue that Twitter's worth is no longer in the tweet's themselves these days, or the volume of those tweets, but all the value that these tweets contain – little tiny metrics identifiers that can be used to offer new experiences.
We're only just beginning to realize this, but the same holds true for music. It doesn't matter how many songs you can fit in your pocket if you don't know what they mean.