The Long Tail, by Chris Anderson (the editor of Wired Magazine) seems to have confused a lot of people. Many discussions and reviews of the book are themselves confusing. That being the case, I'll take a shot here at explaining what the book and the idea are all about.
In a virtual nutshell, Anderson's thesis is that a huge cumulative market for poor-selling items exists on the Internet.
But there's more to it than that, beginning with the 80/20 rule in retailing, which says that 80 percent of profits are generated by 20 percent of items stocked by a retailer. This is generally accepted and has been demonstrated to be true. If you make a sales-by-units graph of all items offered by a retailer (Wal-Mart, Stero Barn, or Rhapsody.com) with the x- or vertical axis being units, and the y- or horizontal axis begin individual products, the bestselling products would make a tall, narrow hump at the left (the top 20 pecent), while the products that don't sell well make a "tail" to the right, dropping down as the number of units sold decreases with each product. Hence, "the Long Tail."
As a consequence, retailers stock as much of the top-selling 20 percent as they can, and ignore most of the other 80 percent of items they might stock--on occasion giving one or another a trial (due to to manufacturer or distributor pressure, or just to see whether a particular item will sell).
Thus, while you can find a copy of The Da Vinci Code in nearly any bookstore, you cannot find The Odysseus Solution because it doesn't sell well. Moving at most one or two copies per year, it sells so poorly that it's not worth keeping in stock. Why have The Odysseus Solution in shelf space that could be used to stock The Da Vinci Code or The Year of Magical Thinking, either of which might sell anywhere from 30 to 300 copies in that same year.
This situation seriously limits the consumer's choices to the most popular items--the "hits," if you will. It's why you can't get that brand of jelly you love at Winn-Dixie; they dropped it because it wasn't among the top-sellers. You have to satisfy yourself from the six other brands offered, do without, or go elsewhere (and you still may not find it).
For the same reason you don't see Blood, Sweat and Tears' first album, The Child is the Father to the Man, on store shelves, though you'll probably find a BS&T greatest hits album. Again, your choices are being limited.
According to Anderson, there are people who would buy Brand X jelly or The Child is the Father to the Man if jars or copies were in stock. But they aren't, and won't be, because the return on the investment is less than it costs the store to stock the item. According to Anderson's theories, this condition does not obtain on the Web, especially where downloadable media products are concerned, as there is no "stocking" charge. All that needs to be done is to let the customers know the item is there.
With no stocking cost, a retailer can stock a near-infinite number of products. This, in turn, means that the retailer can reach the market for items that a brick-and-mortar store could not serve. Potentially, there is a large cumulative market there, supplying thousands of buyers with things they can't get out in the real world.
What tipped Anderson off to all this was a study of the online music markets, where nearly everything can be stocked--including The Child is the Father to the Man and every tune ever put out by a garage band or Indie label. He learned that 98 percent of everything sells at least once in a year. Which means that the online music retailer was raking it in more than the brick-and-mortar music store, because she is able to sell same the big-profit 20 percent that the brick-and-mortar store sells, plus the other 80 percent.
Anderson follows this to its logical conclusion: there are large cumulative markets that conventional retailers are ignoring because it doesn't pay them to stock items that will sell once in a year or so. But on the Internet, where you have no stocking expense ... Which is only one way of presenting the idea.
Anderson takes several runs at explaining the concept, which got him skewered by several reviewers for being redundant. Anderson also talks about ways of capitalizing on the long tail, and the implications on business and culture. (For example, he foresees a market that is less-driven by "hits" than by personal recommendations, which I don't see as happening to a great extent, due to the power of advertising.)
While some reviewers have made a big deal out of the fact that Anderson's long tail concept exists under other names (like Pareto tail), this doesn't invalidate his message, which is, again, that huge niche markets exists among those who are willing to serve them.