"The Long Tail Works!", reports Tail

“75 percent of tracks sold at least once on eMusic, as opposed to 80 percent of tracks not selling at all in MCPS-PRS’s sample. Why would eMusic’s numbers be a mirror image of MCPS-PRS’s numbers? Clearly, one reason is that there’s no major label music on eMusic.

“eMusic is the long tail,” reiterated Madeleine Milne, managing director of eMusic Europe. “Our customers buy music beyond the mainstream top 40 because we provide them with more context than any other major music retailer through Web 2.0 features, insightful editorial content, a passionate subscriber community and an easy-to-use and effective recommendation engine.”

In other words, for sites that define themselves by selling only non-hits, the Long Tail is applicable, and it “works”.

But! If you look at a different data set, say Bit Torrent networks (where everything, of course, is available, and with only the most cursory filters driving people toward one thing or another), then according to Will Page (that guy who recently depantsed The Long Tail) “popular music is popular wherever it is popular.”

I’m inclined to agree with Page more than Anderson, because, in his own words (from the above-linked interview), “I drew upon a great expression that I learned whilst in the Government Economic Service, which is to always strive for ‘evidence-based policy making’ and resist the temptation of ‘policy-based evidence making’. Increasingly, when I hear those words “here’s another great example of the long tail at work”, I’m inclined to expect that claim to lean towards the latter of the two.”

In other words, Anderson’s Long Tail was so popular because, like Malcolm Gladwell, he took incredibly complex concepts like log-normal distributions and rephrased them as very understandable/accessible anecdotes and easy-to-follow “rules.” And to his credit, Anderson gave a lot of people a completely new framework to think about how merchandisers sell culture online. But it also opened itself up to challengers (Gladwell’s got his own) who put it to the test.

The main problem with Anderson’s theory is summed up in the quote that inspired me to type up this quickie post, coupled with Page’s “policy-based evidence-making” explanation. If you approach a problem with a framework in mind–say, one that’s essentially a techno-democratic self-fulfilling prophecy–then yes, you will find the data you need. And if you’re a compelling writer with a built-in audience, then your theory will gain a ton of traction, especially at a time when arts criticism is most often about the structures that bring the art to audiences, not about the art itself.

(Anderson’s own form of publicity, for himself and his book, is, in a way, an argument against his theory. Consider the fact that the only bands able to really maximize their profits using independent, on-the-cheap pricing structures [Radiohead, NIN, etc.] are those who’ve built an audience using the old-school, mass-mediated M.O. Like having Wired’s base, which has been building since 1993, or right about when Pablo Honey hit the shelves.)

But what Anderson’s really talking about is the move in all sectors of culture-selling from curation toward aggregation. It’s the reason that most people vastly prefer hitting up the Hype Machine instead of subscribing to 300 individual mp3 blogs. It’s also the reason that tiny independent labels have to sign up with monster aggregating services in order to sit at the kids’ table with Myspace Music, and why, if you skew the data set appropriately, something like eMusic can appear to, in a press release, look like a solid defense for a theory instead of a mutually-beneficial bit of promo.