lastfm signs deal with warner music

Last.fm signs deal with Warner Music Group

Congrats to last.fm who just announced its first content agreement

with a major label.

The deal will let Last.fm's rapidly growing community have access to

Warner's amazing roster of artists through their free streaming radio

service and tee up a premium subscription radio service.

Last.fm is growing like crazy, with over 15m monthly active users now

experiencing the joys of discovering new music through the

recommendations of other community members. Making it easy for people

to access major label content will only make the service more valuable

to its users and easier for labels to promote both new and catalogue

product to consumers - so great news for everyone.

This service is particularly close to my heart (and not just because I

am a small investor) as I've been dabbling in the entertainment

recommendations space for over ten years now.

First with Firefly, where we launched the web's first music

recommendation service (later sold to Launch, now Yahoo Music) and

then again with Video Island (now Lovefilm) where we applied similar

principles to movie recommendations to build Europe's largest online

DVD rental service.

There is no doubt that at Firefly we were way too early and the

Internet was too immature a medium for the real power of

recommendations to take off, but in the last ten years both Amazon and

Netflix have done an amazing job popularizing collaborative filtering

and making recommendations a central plank of consumer discovery.

Amazon has elevated the application of datamining to help consumers

decide what products to buy to an art form. Netflix values the

approach so much it's offered a $1m prize to anyone who can improve

their approach to movie discovery.

More often than not the weapons of datamining have been wielded by

companies on behalf of companies - they mined to see how to sell more.

Companies in the future will put these weapons to work for consumers.

Last.fm is at one of the new businesses at the vanguard of taking

discovery to the next level.

One thing which is particularly nice about their approach is the fact

that through their scrobbling software (which attaches to your media

player), they make it so easy for you to contribute to and benefit

from the rest of the community.

We wrestled with the earliest applications of social information

filtering at Firefly and one of the big challenges was always about

how to best motivate explicit (ie. a rating) and weight implicit (ie.

an observed behaviour) data. There is no doubt implicit data

collection is much easier on the user, it takes no time, all you have

to do is agree to be observed and feel comfortable that in return your

information will get great value and your privacy will be protected.

Nic Brisbourne has an interesting post on this subject.

Last.fm works because people trust the service to watch what they are

listening to they get discover new things -- clearly this works

beautifully with music, but there is no doubt with the emergence of

other digitally based services we will see more and more of this

paradigm. My old friend Seth Goldtsein, as usual, is really pushing

the envelope here with his work on Attention Trust and Roots Market.

side note: another cool thing about last.fm is their integration

with Skype. All the communications service (Skype, Google, AIM,

Yahoo! and Messenger) are now opening up to 3rd parties here are

going to be amazing opportunities to build great application on top

of platforms which come with presence, contacts, voice, IM and

video built in. What an amazing platform for new applications to be

built around.

Like with Last.fm, obviously this can work with people's explicit

knowledge with video (think where Joost can go) and it is already

working to some degree without people's knowledge in online

advertising and the current vogue for behavioural targeting.

We are scratching the surface of how we will be able to leverage the

computational power now available to us for the analysis of vast

swathes of comples data to benefit consumers and help them improve

their decisions. One of the best examples of a service helping people

make better decisions through the application of hard core math and

computation is Farecast -- delivering an amazingly valuable view of

airfare information to help people decide when to buy tickets.

Congratulations to Last.fm for showing the tip-of-the-iceberg to 15m

people. I'm looking forward to seeing what's next.

Labels: amazon, collaborative filtering, datamining, farecast,

firefly, joost, last.fm, music, portfolio, recommendations