I log on to Napster and discover that there are 708,372 songs currently available for download. Simultaneously, using the Bearshare client for Gnutella, I have access to another 80,000 files. At MP3.com, 750,000 songs by more than 100,000 bands await my perusal. Meanwhile, Listen.com lists 900,000 songs and at EMusic I can purchase any of 165,000 different tunes.
So much to choose from -- isn't the Net great? But there's a problem: I don't know what I want, or rather, I want something that I don't know. Something new, something fresh, something that hasn't already been pummeled into my ears by Sony or AOL Time Warner or MTV. I want someone to tell me, in effect, "Try this, you'll like it." I want expert, personal treatment that will cut through all the chaff and guide me directly to the gems.
Most of all, I want the Net to deliver on its promise to break the chains that hold us in thrall to the major record labels -- I want it to expose me to new, unheard-of bands. Ever since a little outfit called IUMA got started back even before the Web broke big, I've been hearing about how the Net will cut out the middleman and usher in an era where the little guy has a chance. But until now, I've been waiting in vain. And so have all those obscure garage bands hoping that the Web will give them a leg up.
Big stars still rule the roost, with only a few exceptions. Despite the advice of a staff of informed critics, Listen.com's top downloads are Madonna, Eminem and Britney Spears. And although a few savvily marketed indie bands have broken through, MP3.com's top 40 downloads are still riddled with pop acts like Madonna, Eric Clapton and Faith Hill. The Net may have made available more music than ever before, but fans still aren't listening to a lot of new tunes.
Back in 1996, an MIT Media Lab-backed company called Firefly thought it had figured out a solution to the problem, and launched the personalized music recommendation system "BigNote." Using a technology known as "collaborative filtering," BigNote was supposed to suggest new bands you might enjoy based on the ratings of other users with similar tastes. It was a neat concept and it spawned a frenzy of competitors. For years, "personalization" was a buzz word to rival all others. But unfortunately for new music fans, Firefly just didn't work very well. Like so many other buzz words of the late '90s, personalization tantalized but never arrived.
Fast forward to 2001. Another company has emerged from Boston: Media Unbound. Like Firefly, Media Unbound is offering a personalized recommendation system that will suggest bands you might enjoy, based on ones that you already like. Unlike Firefly, Media Unbound does what it promises to do: introduce new, obscure bands you'll actually like.
And Media Unbound isn't alone -- there's also Mubu.com, which offers a similar service; and MoodLogic, which takes a more search-engine approach. Similarly, the music-discovery search engine Gigabeat was purchased by Napster two weeks ago.
And that's where it gets interesting. If personalization that works truly has broken through, it's possible to imagine a future in which obscure bands do get more time in the sun. Because personalized music recommendation technology in combination with file-trading services like Napster or Gnutella could be an amazingly potent brew. Get the recommendation, listen to the tune via Napster, then click a button and buy the CD. Finally, we may be at the verge of escaping the industry-imposed domination of pop pablum, a world in which the only albums you know to buy are the bland Top 40 hits churned out by your local radio station. Speaking optimistically, personalization may turn out to be not just a cheap buzzword that helps Web sites lure that V.C. cash, but the best thing to happen to indie bands and music fans since, well, the Net.
"In the future, helping people find, discover and navigate music content is going to be really important," says Michael Papish, the youthfully exuberant founder of Media Unbound, a Harvard student who left school to start the company. "When you have a subscription service you have access to all this music, 3 million songs; what do you listen to? How do you find it? If you don't have something to help you, it's no better than FM radio or listening to CDs; there's just too much out there."
There is nothing as inscrutable or obsessive as the vagaries of a music fan's tastes. I pity the poor soul who tries to make sense of my music collection: I love R.E.M., but only their first few albums; Nick Drake sends chills up and down my spine, but I hate Van Morrison. Massive Attack is terrific but Tricky's albums do nothing for me. And while I could dance all night to MJ Cole, I could not care less about the prospect of never hearing another Chemical Brothers track ever again. Then there's my secret stealthy (well, not so stealthy anymore) love of Destiny's Child's overproduced pop. From the outside my collection would seem to be all over the map -- and yet, to me, it makes perfect sense.
The music industry forces fans into neatly-packaged "genres" -- you're a "pop fan" or a "country fan" or a follower of alternative rock. This, in turn, is how our music is packaged: Turn on the radio and you'll find a "rock station" that regurgitates the same tunes you'll find on any other "rock station" anywhere in America. And if you're a "rock fan," this station is supposed to satisfy all your musical needs, sending you running to stores to purchase, lemming-like, the albums fed to you by some program director paid off by record labels to promote their albums.
But musical tastes often run wide instead of deep -- which makes it difficult to recommend music to a stranger. Unless you are intimately familiar with a person's musical tastes and quirks, suggesting new music is often like trying to shoot at a saucer with a rifle full of buckshot; sure, you may occasionally hit, but just as often you're going to miss your target altogether.
In 1996, MIT Media Lab professor Pattie Maes decided that the best way to tackle this problem was through collaborative filtering, and helped launched Firefly's BigNote music service. The concept was neatly logical: users would rate and review music, building a grand cross-referenced database of musical tastes. The more you told the system what you liked, the more Firefly would be able to make specific music recommendations based on what other similar Firefly users liked. If you rated Alanis Morissette highly, and other Firefly users who rated Alanis highly also showed a fondness for Fiona Apple, Firefly would turn around and recommend that you check out Fiona Apple. Elementary.
The collaborative filtering concept very quickly became the buzzword du jour. Companies like LikeMinds and NetPerceptions sprang up overnight, offering collaborative filtering services to Web sites that would help you find everything from good books (such as Amazon.com's Recommendations Explorer) to movies (like MovieCritic.com) to Web sites (such as SiteSeer.com). But most of these sites faded quickly as the collaborative filtering buzz gave way to the next hyped Net hysteria. (Push, anyone?) Firefly searched for a business model and floundered; the pieces of its technology were sold to Microsoft and Launch.com in 1997.
Firefly's music recommendation engine was more of a novelty than a truly useful tool: Sure, it could tell you that if you like Alanis you'd like Fiona, but wasn't that pretty obvious already? The system would rarely, if ever, break out of the mold of mainstream bands and recommend fringe music you'd never heard of before. And if your tastes strayed across numerous niches -- say, you liked country and pop and techno, but weren't particularly devoted to any one genre -- Firefly was equally problematic; the odds of finding a community of users with identically eclectic tastes were slim.
Maes was aware of the limitations of the system. "One big problem that Firefly didn't solve is that when all the recommendations are only based on what other people enter, then it's harder for new artists to get recommended," says Maes. "Enough other people need to have taken an initiative to put in a rating for an artist before it gets recommended to people with similar tastes."
Maes now serves on the advisory board of Media Unbound. And while Media Unbound uses collaborative filtering at its base, the system also relies heavily on the opinions of critics. Not only does the software use statistical models and mathematics, but Media Unbound's staff of musical experts pick out the best music and build extensive maps with "distance metrics" that correlate how close to each other various genres, bands, sounds and songs are. These critics would, for example, recognize that R.E.M. had several different "periods" and that songs from specific eras appeal to different tastes; if you don't like "Shiny Happy People," you might still like "Exhuming McCarthy."
Before offering any advice the Media Unbound software asks questions like "If you were talking with a friend, how would you describe your music preference? Sound, Artist, Mood, Era or Genre?"; "Name your favorite five bands"; and "What type of guitar instrumentation do you like?" Next, you'll rate clips of songs from three of your favorite musical genres, and explain what, exactly, you like about the music. The resulting recommendations are surprisingly precise. After fifteen minutes, Media Unbound will create a personalized radio station for you, using some music that you've heard of and a lot that you haven't; but almost all of which you'll like. The more you listen, the more the system will tailor your personal profile.
"In general, we try shying away from drawing distinct lines," says Papish. "Saying this song is alternative, and therefore you are an indie listener so you can't listen to that song. The music universe is very fluid, so we can draw parallels and lines around things and let you move around them, but we don't limit you to a particular bucket."
I, for example, am very picky when it comes to indie rock; I don't like much of it, but do have very specific bands that I love, which cross a number of genres. The radio station that Media Unbound put together for me was perfect: It had the Pixies, Liz Phair, Massive Attack, Sleater-Kinney, Stereolab and Elliott Smith, some old Cure and Smiths classics, certain Nick Drake songs, and a bit of Belle and Sebastian -- all of which perfectly epitomized my rock tastes. But it also spit out artists like Gomez, Helium, the Lucksmiths and Neutral Milk Hotel; obscure bands I had never heard of before, but which I discovered that I liked. (Media Unbound also suggested a few songs I hated, but heck, no technology is perfect.) In just half an hour, Media Unbound managed to convince me to investigate 10 new bands that I never would have heard on the radio.
Personalization is making a clear comeback. Media Unbound has at least a half-dozen competitors pushing their own music recommendation engines -- technologies from Mubu.com, Muze, Gigabeat and MoodLogic, each of which has its own approach to music recommendation (Gigabeat uses "relationship maps" to suggest similar artists; MoodLogic allows you to find music according to search criteria like era, tempo, vocals or popularity; and so on.) But of the competitors, only Mubu.com remotely approaches the accuracy or utility of Media Unbound's personalized recommendations.
Unlike Media Unbound, Mubu.com forces you to pick a genre from the outset (electronic, country, rock, pop, jazz, urban), and then a subgenre that you'd like to explore. After you rate eight music clips, the system will recommend an equal number of songs that it thinks you'll like. If you like what it offers you, you can immediately buy the album or add it to your "favorites" list.
The system is relatively accurate, and is a good source for finding new bands within a genre that you already like. It isn't particularly deep, however: unlike Media Unbound, which tweaks and expands your profile with every click, Mubu.com's recommendations seemed to peter out after one or two rounds of ratings, and the engine will simply recommend the same dozen albums over and over. With Mubu.com, you're also starting your search from within a genre, which means that from the outset you have already limited your recommendations to a relatively small pool. (Under the "Rock," genre, for example, you have to choose between alternative metal, modern rock, rock, college radio, jam rock, detention, hardcore metal, goth/industrial, blues, stoned, anarchy, rock for girlz, pop rock, evil, acoustic, reggae, working class, world class rock, tattoos & poolcues, and indie. Call me clueless, but what exactly is "tattoos and poolcues rock"?)
Still, after spending a day exploring the "electronica" section of Mubu.com and expanding my indie rock tastes with Media Unbound, I had a list of several dozen lesser-known bands whose albums I was interested in purchasing. This, in turn, is where applications like Napster come in handy: I wanted to hear not just the Gomez single that Media Unbound had offered me, but other songs from the band's albums, before deciding if I really wanted to shell out $17 for a CD. After all, I could have easily spent hundreds of dollars on all those new bands' music -- and how many music fans can afford that much on a regular basis? A peer-to-peer file-sharing system would be the perfect place to test-drive the new bands that the recommendation system had told you about.
There's no question that this is where the personalization technology companies are headed. Both Mubu.com and Media Unbound are packaging themselves as technology solutions for Web sites, file-swapping services and radio stations that are hoping to better target consumers. Listen.com is using Muze.com; and Gigabeat was recently purchased by Napster, although it remains to be seen how the service will be incorporated into the system. Similarly, Microsoft's new MSN Music service, which launched on Wednesday, suggests radio stations based on your "mood" or your favorite artist, using a recommendation engine purchased last year from MongoMusic.com.
In an ideal world, the music delivery services of the future would not just let you download music but would expand your horizons, too. Patti Maes describes it like this: "For me, an ideal system is really one which gives me a mix of things that I like, that are related to my past interests, but that also introduces me to new things and tries to push the boundaries of my taste. These kinds of systems can do this, because they have a model of what a user is interested in and they can give you stuff that's just outside your boundaries. There's an element of serendipity; not just tracking the changing interests of the user, but changing the interests of the user by introducing them to new things."
Personalization isn't going to turn an indie band like Goldfrapp into the next Britney Spears. Adventurous types who want their tastes expanded may love these services, but there are also a whole lot of lowest-common-denominator music fans out there who will still demand their hits spoon-fed to them. Mubu.com CEO John Adams, for one, is more cynical than Maes. As he puts it, "Why is it that the top 20 Internet songs, downloads, streaming and sales, are all major-market hacks? They aren't independent unknowns; the Net is just another medium to communicate with, and people who like things off the Net are going to like the same things on the Net. You have to accept those social and fiscal realities if you're going to do business here; it's about building a profitable business that provides a valuable service to other businesses."
Nor are radio stations and rock critics likely to quake in fear of Media Unbound and similar technologies taking their place in the pantheon of music arbitration and recommendation -- Rolling Stone and KROQ are still going to be pumping out the mainstream pablum for the masses for a good long time.
But the future, finally, looks bright. For years, the Net's evolution has been accompanied by at least as much hype as reality. Everything was going to change, right this second! In the current climate of crashing dot-coms, lawsuits and general digital malaise, it has become apparent that the revolution as advertised is moving a good deal more slowly than the Net's once-breathless pundits predicted. But it is moving, inexorably and irresistibly. The marriage of personalization technology and file-sharing services is going to be difficult to break up. Better music, or at the very least, better music distribution, is on the way.