• by adam , Shelly Palmer , July 7, 2005

The Eskimos have 27 words for snow... The Great Wall of China is the only manmade object you can see from space... more than 20 percent of value is locked up in the Long Tail. These are all myths that have such romantic power, you just want to believe them, but they are still myths.

Before the Long Tail (a wonderful, well-written article by Chris Anderson that appeared in Wired magazine a while back), there was the Italian economist, Vilfredo Pareto, who created a mathematical formula in 1906 to describe the unequal distribution of wealth in his country. He observed that 20 percent of the people in Italy owned 80 percent of the wealth. In the 1940s, Dr. Joseph M. Juran attributed his own observation of "vital few and trivial many" to the 80/20 Rule which he called "Pareto's Law" or the "Pareto Principle."

That's enough history. The practical reality is that in almost all of the observable universe, the 80/20 rule applies. Twenty percent of the distribution embodies 80 percent of the substance. You fill in the discipline, you fill in the variables, it holds up time and time again. If you make a graph of Pareto's Principle, you get a bell-shaped curve, otherwise known as the Big Hump.

However, in all things there are exceptions to the rule. Take video on demand or video rentals -- the business of hits. And we're not talking about hits to your Web site. Hit records, hit movies, hit books, etc. follow a different curve known as Zipf's Distribution. Originally, the term Zipf's law meant the observation of Harvard Linguist George Kingsley Zipf that the frequency of use of the nth-most-frequently-used word in any natural language is approximately inversely proportional to n.

The classic case of Zipf's law is a "1/f function." What that means is, given a set of Zipfian distributed frequencies, sorted from the most common to the least common, the second most common frequency will occur half as often as the first. The third most common frequency will occur one-third as often as the first. The nth most common frequency will occur 1/n as often as the first.

Zipf's law is an experimental law, not a theoretical one. Zipfian distributions are commonly observed in many kinds of phenomena. The curve that illustrates Zipf's law looks like a long tail (hence the title of Anderson's article).

Since both of these distribution models are fairly common in everyday life, why would I have the temerity to call the Long Tail a myth? Let's review: With regard to content (audio and video on the Internet), there is a pretty good theoretical argument to be made for the value proposition of the long tail. After all, in theory, it goes on forever and it is full of valuable things like old Sidney Bechet and Bunk Johnson recordings. If you had the pleasure of hearing these guys play, you would immediately understand that the 78 RPM records from which the files were made are truly priceless.

These artists, by default, have a place on the long tail. But, how will you find them and if you do, who will get a paid? (Fodder for a different article, sorry for the digression.) You may have heard of these guys and even be willing to search for their stuff, but who has heard of the '70s disco band Demon Rum whose recordings are also on the tail?

There is a very real point where marginal cost exceeds marginal gain and it happens very close to the borderline of the top 20 percent of the selections available on the Long Tail. So, in essence, the Big Hump is still the appropriate model for investors and content owners to follow.

Take a movie that already exists. Retain counsel (internal or external) to get all of the appropriate clearances for reduction of the work to the final form factor - ones and zeros. Rent a film chain and color correction suite for the day. Make a 4K file of the entire movie. Down-convert it to every usable format for current distribution and then create files in every resolution required from 56k up.

Now the fun starts. Create meta-tags and descriptions that are meaningful to search engines so people can find the work. Add a DRM (digital rights management) wrapper, so you can get paid. Come on, take a guess... how much have we spent on this single title? Call it $15,000 sitting on a server ready to go. Now remember, we don't know who wants to watch this movie, we are just sure that in time, every movie will be watched according to the Zipfian distribution. Add the cost of money to this equation and multiply by every title in your library. Still sound like a good idea?

I didn't think so. You probably want to do an analysis to determine which movies are the best investments: Which will bring the best return in the shortest time frame? In other words, you will apply the Big Hump to the Long Tail to determine if your content is worth encoding at all.

But isn't the Long Tail an accurate way to look at content usage? Yes, it is. It's just not a profitable way to look at a back catalog that was not created with the Long Tail in mind. It is never less expensive than in the original edit suite. But after the fact, there are very real costs associated with encoding and storage. When you calculate those costs and the cost of capital, the value chain (or lack thereof) is clear.

So, as it has been for time immemorial, the Big Hump is the clear winner for modeling return on investment in the information age or any other. No disrespect intended to those who have been spinning the tale of the Long Tail.