The above link discusses how the “long-tailed theory” could be wrong. This theory was created by Chris Anderson, editor of Wired Magazine. He believes that the internet is a medium that enables the consumer with more choice which is fueled by content on demand. With more choice, means that niche products will have an increasing influence on total sales. Niche products/services are not normally sold through traditional channels (e.g. brick and mortar operations) due to physical contstraints such as inventory, which leads to a lack in economies of scale to justify the availabilities of those products/services. However, when all those other niche products/services are added together, this amounts to a large number. So significant in fact, that he theorizes that the summation of the number of niche or non-mainstream products will exceed the total amount of mainstream products.
A Harvard professor came out with a study last week to refute this theory. Its quite an interesting read.