Home > Advice & Tips, Digital & Mobile, Indie, Music Producers, Technology, The Future... > Finding New Fans With Recommendation Systems

Finding New Fans With Recommendation Systems

telstar logistics last memory record store Finding New Fans With Recommendation SystemsOn Saturday, we pointed out the possibility that a team has won the Netflix Prize, and why this is good news for musicians.

To recap: recommendation systems will increasingly become a way for music fans to discover emerging and niche musicians.

Also last week, we took a look at what the Big 4 record labels are doing to take advantage of the fundamental shifts in the music business. They focused a lot on “music discovery.” So, getting listeners to discover new music is one of the keys to success in the music commerce frontier, no matter what level you are at.

It may be helpful, then, to think of the basic business model like this:

1) Discovery

2) CwF + RtB = $$$

The second step is a concept coined by Michael Masnick, though I noted that even the Big 4 are following a plan very similar in spirit. For musicians and labels to be successful, they need to Connect with Fans and give them a Reason to Buy. But before they can do that, the fans have to find out about them.

Enter recommendation systems

Recommendation systems are improving every day, as the team that reached the goal in the Netflix Prize illustrate. But, as Paul Lamere put it, recommendation systems are broken. The most commonly-used recommendation engine (collaborative filtering) still keeps people stuck in the “head” of the distribution curve (as opposed to making their way toward the “long tail“), because it bases its recommendations on user behavior. A lot of people buy Harry Potter, it gets recommended more often, more people buy Harry Potter from those recommendations, and so on.

Recommendation systems that use the content of the music itself rather than user behavior offer a lot of potential for pulling in more relevant recommendations from all parts of the long tail. Pandora uses content-based recommendation to drive its playlists.

Social tagging provides another way for collaborative filtering systems to recommend more niche and lesser-known products to listeners. Last.fm has probably one of the most robust recommendation engines for a purely music-based site. It offers the ability to tag on all levels: single tracks, albums, artists, even users themselves.

Taking advantage of recommendation systems

For the independent musician, advice for taking advantage of the opportunities that recommendation systems offer is not too different from advice for marketing in general or search engine optimization.

1) Get your music onto sites that offer recommendations for listeners. Last.fm is just one example; it’s very easy and free to start a page for yourself or your band. Heck, you may already be on the site since Last.fm pulls a ton of data from other sites, desktop music players, and mobile devices (in which case, it’s just as easy to ‘claim’ your artist page).

2) Tag it, tag it, tag it. You want listeners with similar tastes to find your music – if you tag it with “Britney Spears” or “Eminem” thinking you can game the system, you won’t be pulling in people who may potentially like your music. Use relevant tags: tags that describe the music, instrumentation, genre, similar artists, etc.

3) Just as with general marketing and search engine optimization, make sure the text (bio, artist description, influences, etc) on any site you have your music is relevant, accurate, and descriptive. More and more, recommendation systems are crawling the web and pulling semantic information from band pages to incorporate with other data.

I’ve only covered the very basics for now. In the future, recommendation engine optimization may be just as important as search engine optimization for musicians. For those artists who haven’t thought about harnessing the power of recommendation systems to bring in new fans, hopefully this post has inspired you to get started.

(image courtesy of Telstar Logistics)
Share with others:
  • E-mail this story to a friend!
  • Digg
  • Facebook
  • Twitter
  • StumbleUpon
  • del.icio.us
  • MySpace
  • Sphinn
  • Mixx
  • Google Bookmarks
  • Yahoo! Bookmarks
  • Ping.fm
  • Global Grind
  1. June 29th, 2009 at 14:13 | #1

    First off, I think you’ve done a great job starting a list of elements that work both in independent and major business models. I agree that “Recommendation Systems” are part of the future. We will be each others “Taste Makers”, and I believe this is exactly what we should be looking at.

    I signed up on Last.fm and immediately after the account was verified I was given a multitude of different recommendation paths at my fingertips. Popular terms were MGMT, country, Linkin Park, and punk rock. I enjoy MGMT which allowed me to directly jump into the tagging and familiarization process of the system.

    Also, as I noticed they have a program called the Scrobbler:

    “Get music recommendations
    based on the music in your library.”

    “Connect your media player or iPod
    to fill your library automatically
    Get Last.fm on your iPhone or Android”

    Not a bad place to go. Good read!

  2. June 29th, 2009 at 21:07 | #2

    Thanks, glad you enjoyed it!
    Yeah, I didn’t go into any detail about Last.fm’s Scrobbler, but I think it’s pretty amazing. Recommendation systems work better the more data you have, and the Scrobbler collects mountains of data.

  3. June 30th, 2009 at 12:36 | #3

    I always thought the automatic recommendation engines worked better than when a website asks a band “who do you sound like”; this often leads to wishful thinking (We sound like the Beatles if James Brown was the singer).

  1. July 4th, 2009 at 14:09 | #1
  2. August 9th, 2009 at 22:05 | #2
  3. December 12th, 2009 at 00:21 | #3