We’ve looked at the stricken economics of the streaming sector and the challenges of the artist services sector, now it’s time to take a look at the evolving industry from a new perspective. A lot of effort is going into trying to recreate the old music industry online when perhaps we should look at how other sectors have been transformed forever by the Internet and look at it from a Web, rather than a music perspective.
‘The Music Industry’ is no longer limited to record labels and other related corporate music institutions. They still (and will continue to) exist but they have been joined by millions of artists – effectively sole traders empowered (although not quite yet enabled) by the Web to build their own career, and distribute and profit from their music, without the necessity to partner with a label.
There are 7m odd tracks available from the 4 majors and this is the catalogue that, broadly speaking, every consumer facing music site licences and seeks to monetise. But in addition to these there are well over 10 million artists and 100m other tracks lying around on sites such as MySpace. So why do retail/search/recommendation services almost exclusively revolve around the 7m catalogue? The whole point of the Web is that consumers can almost instantaneously find exactly what they are looking for. It works for books (Amazon), retail (eBay) and travel (Expedia) – these sites don’t just list products from 4 major suppliers, they list everything from everybody and you can search for it. So why not music? Have you ever sat in front of a Spotify screen and thought ‘7 m tracks, there must be amazing things in here, but how do I find them?’ Better still, in front of 100m tracks on MySpace and wondered how great it would be if you could be taken to your perfect track every time? At the moment, it’s like going onto eBay and only being able to select a category but not being able to enter a specific search.
It’s all about search. On Amazon you type what you are looking for and, after you have found it, there are recommendations on what else you might like. In music you just get the recommendations (and pretty average ones at that) – the fact is that other than artist name and track name, there is no search in music.
Fast Forward to Nirvana (the place, not the band), how should music search work? – well either you should be able to describe the type of music you are looking for (in free text) or, alternatively (and I prefer this one) use a reference track to search for all other tracks that were very similar, regardless of artist. For many people, the emotional connection with the artist is an important ingredient, but for the vast majority (over 95%) it’s just about finding great music on a consistent basis, regardless of who wrote/performed/released it. Emotion follows great music, not the other way round.
So why is there no music search? The nub of the problem is that the thing you are looking for must have data attached to it that is of relevance to the search being carried out. On eBay we know what the item is, its age, description, the quality of the seller and a good idea of what others think of its worth. It’s a magic blend of factual description, quality control and crowd wisdom. In music the name of the artist and the track name tell you next to nothing about the listening experience of the track itself (or its quality), but for most of those 100m+ tracks this is all we have. And for most of us it’s not about the artist it’s about the track, which is why most recommendation engines fall short, because they stick on artist and cannot go more granular.
Pandora are perhaps the current leaders in music search – and the way they have managed this – by manually wrapping data (up to 400 manually applied attributes per track + recommendation) around millions of tracks and then using the data around one track to map it to other similar tracks. Does it work? 50m users (that’s 1 in 6 people in the US) and the only profitable personalised streaming business in the world is the result, so I think that’s a yes. Pandora realised almost a decade ago that the secret behind music recommendation was data, unless a track is wrapped in data it is almost impossible to clump it with other similar tracks.
So where is this going…if you are an artist in the 7m catalogue you have some visibility because a small percentage of these tracks are played often enough to get you high enough in the recommendation engine to acquire new fans. But if you are outside the 7m, however good you are as an artist, the consumer will not be able to find you, even if your song is the one they have been waiting for all their life. Accept it. And the consumer is mind-bendingly lazy, they do not have the time, inclination or energy to look for you, and even if they did there is almost no chance of them finding you. It’s like a monkey on a typewriter, there is a tiny chance something good will emerge…but the odds are against you.
So, as an artist, what do you do? Most still strive for a record deal, and although a record deal is potentially a ticket to exposure, nowadays the financial returns for a relatively successful artist are pitiful. Alternatively there are numerous digital silo services you can use to try and promote yourself. You can pay to get on iTunes, you can pay to have your tracks fired at listeners on internet radio who like vaguely similar stuff and you yearn for radio play in the hope that someone randomly listening will connect with your music and track you down. It’s a disjointed digital ecosystem because its new and there is no integration, no logical progression. Its like a production line where all the machines are working but they are all out of order. All you want (and need) is fans (and their email addresses), fans that will love your music, a sales channel that you can feed with great new music. Fans that you can build a career around.
There is lots of data around consumers, what tracks they like, how much they like them. Pandora enables you to put in an artist/track and stream similar music – this is only possible because all the tracks are wrapped in data. If this could be turned on its head there would be a service for artists that enabled them wrap their track with data and then map their track to the wider market. This would tell them where they fit in the music universe and precisely target the consumers most likely to love their music, a sort of permission marketing. In the world of the long tail, this is the tail wagging the dog.
If this were possible then all digital services fall into place, relevance would direct promotion and opportunity and the world becomes one of mass personalisation. Artists can find their fans, publishers their customers, consumers music they love and record labels the next signing.
So how would this work in practice? An artist lays down a track on a Monday, gets it datawrapped on Tuesday, uses this data to map themselves to the wider established market, (and by extension who and where the most passionate audience are) on a Wednesday and fire their track at that precise audience on a one to one basis on a Thursday. By Friday they have enough new fans to start packaging music/merch/live bundles and offering this to the fans most likely to buy. Weekend off/writing…and repeat…
It’s a little simplistic, but you get the idea.
Think its fantasy – well it’s on its way, consumers have neither the time or patience to find the artists, but the Internet enables the artist to find their audience, on a one to one basis. This is the new music industry, one to many is yesterday, one to one is the new paradigm and the ecosystem to make this happen is already coalescing.
Music search is a universal solution for industry, publishers, consumers and artists – indeed anyone who wants to find or wants to be found. In the same way as a niche supplier can be the top of a list in a Google search, the smallest artist will be able to do the same in a music search. It’s all about rewarding relevance with prominence.