"What's really amazing about the Long Tail is the sheer size of it. Combine enough nonhits on the Long Tail and you've got a market bigger than the hits. Take books: The average Barnes & Noble carries 130,000 titles. Yet more than half of Amazon's book sales come from outside its top 130,000 titles. Consider the implication: If the Amazon statistics are any guide, the market for books that are not even sold in the average bookstore is larger than the market for those that are.
...The average Blockbuster carries fewer than 3,000 DVDs. Yet a fifth of Netflix rentals are outside its top 3,000 titles. Rhapsody streams more songs each month beyond its top 10,000 than it does its top 10,000...
When you think about it, most successful businesses on the Internet are about aggregating the Long Tail in one way or another. Google, for instance, makes most of its money off small advertisers (the long tail of advertising), and eBay is mostly tail as well - niche and one-off products."
-- Chris Anderson, Wired
Just over ten years ago, I wrote an essay called The Economic Consequences of the World Wide Web. The predictions I made were rather bold, considering that at the time the typical personal computer running Windows 3.1 or DOS could not access the Internet, and fewer than half a million people in the entire world had access to a graphical web browser.
One of my predictions was that narrow-interest movies and music would be strengthened by the Web. Wired magazine editor Chris Anderson's article on the "long tail" suggests to me that this has occurred.
Looking back on my decade-old essay, I think that it remains fundamentally sound. If Anderson's article tells us how far we have come, I think that relative to my original predictions we still have a long way to go. But the predictions that have not yet come to pass still strike me as reasonable.
Lower Concentration
The distinguished computer engineer Jacob Nielsen
pointed out that the page views on the Web tend to follow a Zipf distribution. Compared with the most visited site on the Web, the second most-visited site gets about half the traffic, the third most-visited site gets about one-third the traffic, and so on.
One characteristic of the Zipf distribution is that it has a long tail (hence Anderson's use of that term). On the one hand, it exhibits extreme inequality. As NYU's Clay Shirky pointed out in Power Laws and Web Logs, it means that the gap between the traffic on a top web log and the median web log will be large.
However, another characteristic of the Zipf distribution is that the tail is more important than the head. That is because as a fraction of the leading site, the fraction of the nth site declines slowly. See the table below.
|
|
Rank |
Fraction |
|
1 |
1.00 |
|
2 |
|
0.50 |
|
3 |
0.33 |
|
4 |
0.25 |
|
5 |
0.20 |
|
6 |
0.16 |
|
8 |
|
0.12 |
|
9 |
0.11 |
|
10 |
0.10 |
|
11 |
0.09 |
|
12 |
0.08 |
|
13 |
|
0.08 |
The total of value of sites 3 through 13 is 1.6, compared with a value of the top two sites of 1.5. This property holds in general for the Zipf distribution. The long tail accounts for more value than the top. As in the examples cited by Anderson, the chart-toppers do not account for 80 percent, or even 50 percent, of all the business.
This relative significance of the long tail represents a subtle but significant economic change. The industrial economy was characterized by concentrated industries, in media as well as heavy manufacturing. There is no "long tail" in the steel industry. The combined output of all of the steel mills below the top 1,000 is tiny. That industry does not fit the Zipf distribution.
Why the Long Tail?
Why do we see less concentration on the Internet than in the industrial economy? The most fundamental fact about the Internet is that it facilitates capitalism without capital. As I put it ten years ago,
"in building the interstate highway system the challenge was constructing the physical roads and bridges, not in coming up with the rules of the road. With the Information Superhighway, one could argue that the situation is reversed, and indeed that this reversal is one of the major problems with the highway metaphor itself."
The reduced significance of capital means that the cost of entry is lowered in many industries. Today, we see this in the shops that people have set up on eBay or in the blogs that compete with traditional pundits.
Another characteristic of capitalism without capital is less bureaucracy. As observers from John Kenneth Galbraith to Amar Bhide
have pointed out, corporate bureaucracy emerges to regulate risk-taking in an environment in which new projects are very expensive. Think of a new airplane or a new fabrication plant for computer chips.
When a new project can be hatched in a basement on a small budget, fast failure is more efficient than organized planning. Galbraith, writing in an era when the economy was dominated by heavy industry and oligopolies, saw entrepreneurship as little more than a quaint myth. Bhide, writing more recently, sees the entrepreneur as thriving in circumstances of high ambiguity and low capital intensity -- situations that have become increasingly prevalent in the computer age, particularly with the advent of the Internet.
The industrial economy required planning and bureaucracy. The Internet economy instead is better described by Friedrich Hayek's terms spontaneous order and competition as a discovery procedure.
As I wrote ten years ago, "More of the value of a product will be created outside of the firm." The firm's executives and committees are now less well-positioned to make decisions than are its consumers and others on the periphery. The Hayekian consequences of the Internet include outsourcing and Free Agent Nation.
Economics of the World Wide Web, Revisited
A decade ago, there were only a few thousand Web sites, but I predicted that as the Web grew that "Consumers will be highly dependent on agents of various sorts, including catalogs, search tools, and 'best of' lists." Since then, sites have emerged, such as Google, which are now important not only to the Web but to our entire economy.
I foresaw a reduction in the need for sales clerks, as online ordering and self-service would increase. This is taking place, but at a much slower pace than I would have predicted. My guess is that sales clerks' main role these days is to talk people into buying more than they intended. For example, at fast food restaurants, why is the line that says "place your order here" staffed by a human? The only reason that I can come up with is that humans are better at suggesting to customers that they should buy a soft drink.
I predicted that consumers would "start to make 'end runs' around professional lawyers, doctors, real estate agents, stock brokers, insurance agents, etc., because the demand for these professions is a function of consumer ignorance." Internet sites with medical, legal, and real estate information have turned out to be extremely popular, and many consumers do seek their own information in these areas. However, the professionals have retained their gatekeeper roles to a greater extent than I would have expected. It is difficult to assess how much change has been hindered by consumer inertia and how much change has been hindered by regulators who in turn have been captured by the professional trade associations.
I also predicted that consumers ultimately would be able to find alternatives to obtain "effective and credible education." Today, nontraditional education still has a tremendous credibility problem. If and when that problem is solved, my prediction is that traditional institutions of higher education will take a very hard fall. I think that one of the sources of tension between professors and conservatives is due to the fact that colleges and universities have lost their near-monopoly on intellectual talent to a long tail outside of the academy, yet many faculty act as if this nothing has changed.
Information-age Government
Our government is much more at home in the industrial age. It is familiar with taxing and regulating concentrated industries. It is not at all used to dealing with the "long tail."
The philosophical and technical differences between regulating a concentrated industry and a long-tail industry are enormous. For example at the Federal Communications Commission, as even Chairman Michael Powell is aware, the expectations for regulation of telephony and broadcast indecency are increasingly anachronistic.
As more workers join the "long tail" of small business and self-employment, what will happen to our system of employer-provided benefits, especially health insurance? Economist Brad DeLong thinks that without corporate provision of a collective safety net, we are headed toward a larger welfare state. He writes
"The coming generation will be one of massive downward mobility for many Americans. The political struggles that this generates will determine whether America
will move more closely to the social-democratic norm, or find some way to accept and rationalize its existence as a country of high economic risk and deep divisions of income and wealth."
As the industrial age gives way to the information age, the S&P 500 are under more competitive pressure and will lose market share to the long tail. Therefore, I agree with DeLong that the ability of corporate America to provide welfare-state type benefits is eroding.
However, my reading of history is that American affluence has consistently grown in spite of the alleged perils of structural change. For example, after decades of warnings that the middle class depended on manufacturing jobs, we have reached the point where less than 10 percent of the work force is on the production line, and yet most indicators of economic well-being, such as the home ownership rate, suggest that prosperity remains broad-based.
It is always possible that this time the doomsayers are right, and that widespread downward mobility is just around the corner. However, I think that what we will see instead is widespread upward mobility among the self-educated and self-starting people of the world, due to the reduced requirement for huge concentrations of capital. It will be less important to be born into affluence or to please the boss at a well-known company, and more important to enjoy a measure of economic freedom with reasonably unfettered access to people, technology, and markets.
In my view, capitalism without capital could prove to be a golden age. It provides self-educated self-starters with opportunities that never existed in the industrial economy. It offers the potential for individuals in the poorest nations to plug into the affluent world.
Government in the information age could go in one of two directions. One direction, advocated by DeLong and others on the Left, is for a larger welfare state to try to fight the tendency toward inequality. The other direction would be toward more of what I've called a Bleeding-heart libertarian state, without concern for inequality between the top and the middle.
My belief is that this is the wrong time in history to adopt European-style democratic socialism. Under capitalism without capital, we should encourage people to be self-educating and self-starting, not state-parented. Our government should be compassionate and generous to those with mental illness and physical disabilities, but otherwise it should have a smaller footprint, not a larger one.
