Last FM – the challenge of growth

Matt Ogle and Anil Bawa Cavia from Last FM, which tags itself “social music revolution” , spoke about how they built the business. Matthew starts with some stats: 15 million tracks being played and “scrobbled” (metadata sent to the servers), 10 million artists, 70 million tracks, 700,000 tracks which are streamable, 145k artist wikis.

The service works like this: user listens to track, metadata sent to server, other tracks with similar metadata recommended to user, user listens…. The more you listen the better it gets, they say.

The early growth lessons: don’t overextend, make sure revenue scales with usage, involve users in the web application’s story. They recommend putting the blog at the forefront so that there is an open conversation. People prefer to be told the bad news than no news at all. People will be more likely to tolerate your problems.

Anil takes up the story of growing up. “People trump process”, he says. Have simple process and good people. Take simple tools and customise them. Radiate information across different channels in the company. “We use IRC to communicate across our office which we customised.”

He recommends opening up the product as a platform so that users and companies can extend the power of the product. This is a key way to harness critical mas”.

“Openness is the key to web apps.”

Matt is back on stage. He talks about Scrobbling data – it’s “attention data” . The server knows when you listen, when you stop. It’s a kind of “spyware” he says, but people will accept that if there is a real benefit to it. Then it’s called “myware” – spying on yourself.

The social web – attention aggregators – is where the action is.

Some examples of what they do with attention data – a custom “Time Out” for your tastes and your area; blog recommendations based on your taste.

Anil back on. Monitising attention: microchunkit , free it, syndicate it, monetise it. In Last FM’s case they sell labels the chance to pay to have music played to a very targeted audience, and they provide really detailed feedback of what listeners liked, and what they didn’t.

He spoke about tag cloud moderation. The principle is censorship is not acceptable. But by weighting tags by attention then those people who listen to a particular artist count more than those that done.

Matt is back. The future for Last FM?

More:
growth, streamable music, ambient findability, personalisation (things you can do with your data)

Less:
interfaces (rationalise and streamline), barriers to entry (collaborative filters have a cold start).

Technorati Tags:

Leave a Reply

Your email address will not be published. Required fields are marked *