Welcome to Niche Nation

Long Tail author and Wired editor Chris Anderson on infinite markets, the death of the Top 40, and the birth of personalized politics.

Nick Gillespie | September 22, 2006

(Page 2 of 7)

reason: You talk about how Rhapsody, the online music service, which gets about 40 percent of its sales from— >

Anderson: —from music not available in Wal-Mart. The typical Wal-Mart store carries about 4,200 albums, which nominally turns into 55,000 total tracks. And if you do the math right, and realize that not all tracks are equally popular, Wal-Mart effectively stocks about 25,000 popular tracks.

reason: Similar trends are easy to find. In the television market, for in-stance, people watch more TV than ever before, but there are more TV channels and more TV shows. The supply isn’t unlimited, but the typical household gets dozens more channels than it did even 20 years ago. How do the audiences for top-rated shows like Seinfeld or E.R. compare to early ratings champs?

Anderson: Take The Cosby Show. That was the 1980s, and it had an average share of about 35 [meaning 35 percent of households with televisions tuned in]. In the ’60s, The Beverly Hillbillies was a 40. Seinfeld was about a 23. I Love Lucy was almost a 70. Today, CSI is around a 10.

reason: What is driving the Long Tail?

Anderson: It’s an economic shift driven by a technological shift. The formula is really easy. When you lower the cost of distribution, you can carry more stuff. You lower the cost of shelf space, you can put more on it. The Internet effectively lowers the cost of distribution and frees up the shelf space.

If you only have room for 10 things on your shelf, you will find that the marketplace as you measure it will choose from among those 10 things. If you have room for 10,000 things, you’ll find that those 10 things may still be the most popular 10 things, but they’re not 100 percent of the market.

As with water being poured into a bathtub, audiences distribute when given the chance. We’ve been measuring our culture through the lens of sales and hits. That’s what we thought we wanted, but what we find is that limited choice was really what the distribution networks wanted, what the retail channel wanted. It was an artifact of scarcity, an artifact of bottlenecks in the marketplace. We basically had a limited menu, which resulted in limited expressed choice.

reason: Who knew people wanted computers in any color other than putty when that’s all they came in?

Anderson: Exactly.

reason: Will there be a case where the tail is actually worth more in terms of sales than the head? If so, when and where is that likely to happen?

Anderson: I predict music. You have to be very careful about definitions. If you define the tail as music not available in Wal-Mart, the tail will be larger than the head in the digital music services in less than a decade. For Rhapsody, the trend suggests that it will happen in the next five years.

I should say that the theory of the Long Tail is nowhere predicated on the tail being less than half the market. The point is that the tail is significant in size, that it’s a new market and a growing market. What its ultimate size is depends on the specific industry. From industry to industry it’ll be different sizes.

reason: You’'ve been in a bit of a pissing match with Lee Gomes of the Wall Street Journal over The Long Tail. He says that you’re actually overblowing a minor phenomenon. That in terms of retail, even at Amazon.com, retailing is still largely based on hits. He wrote that 25 percent of Amazon’s revenue comes from its tail—defined as everything past the 100,000 best-selling titles. He argues that’s just another way of saying 2.7 percent of its titles represent 75 percent of its sales. What’s wrong with his analysis?

Anderson: I struggled a little bit with some of those statistics in my first week of researching this, too, but fortunately, I had time to actually do the math and the problem is in percentages. Amazon has an inventory of about 3.7 million. One percent of that is 37,000, more titles than most book stores have at all—and about half the inventory, or nearly half the inventory, of even a book superstore, so it sounds like a tiny little bit. But in absolute numbers, it’s a huge amount. So the problem with the percent-age approach is that if the denominator can grow infinitely large, what you find is that actually in percentage terms, these markets look more and more hit centered. That’s just a statistical error. The right way to do it is in absolute numbers and that’s why I keep the definition of a hit consistent at the inventory available in a bricks and mortar marketplace.