The long tail theory of digital distribution implies that given almost unlimited choice, the total demand for less popular items will exceed that of the most popular items. Some studies have suggested that while the traditional demand curve may be growing a longer tail, it is getting flatter rather than fatter. The long and the short of it is that people still seem to gravitate towards towards the best-selling hits. This sting in the long tail could have important implications for online video distribution.

Chris Anderson popularised the concept of ‘The long tail’ in an article in Wired in 2004, and ironically it became the title of his best-selling business book. The incidence of Pareto or power law distributions was already well known, but the attraction of the long tail rapidly became a marketing mantra. The idea was simple. Given virtually unlimited shelf space, an online retailer can offer access to a vast catalogue. As a result, the total volume of sales of obscure niche items may exceed that of the most popular 20%.

Some recent studies have suggested that actual consumer behaviour may still result in a strong tendency towards popular hits. Some examples are summarised in an article in New Scientist.

A recent study from the MCPS-PRS Alliance, which is responsible for collecting music royalties in the United Kingdom, found that of the 13 million tracks available on a popular online music store, just 0.4 per cent, or 52,000 tracks accounted for 80 per cent of downloads.

“The inequality in revenue between hits and niches was jaw-droppingly stark,” observed Will Page, the chief economist of the MCPS-PRS Alliance and co-author of the study, which was presented at a recent Telco 2.0 conference in London.

Only 20% of tracks in the study sold once or more, meaning that 80% of them did not sell at all. Moreover, around 80% of the revenue came from around 3% of those tracks that sold. Only 40 tracks in this example sold more than 100,000 copies, accounting for around 8% of revenues.

Chris Anderson has responded by saying that much more evidence would be needed to undermine the long tail theory. Without identifying the source of the data set used by the copyright collection society it is difficult to draw broader conclusions.

Other researchers suggest that in a fashion-led market, where consumer preferences are fickle and fast-changing, the most profit can actually be delivered by a very small number of items, equivalent to the 10 or 20 titles often stocked in an airport bookstore.

It may be a good marketing strategy to offer a huge range of choice, but the greatest attraction for shoppers, it is suggested, is probably the aggressive discounts offered on best-selling products.

The recommendation engines based on collaborative filtering, suggesting products that people like you like, may paradoxically promote the sale of more popular items, tending to direct sales towards more mainstream products, contributing to a homogenisation of choices.

Social networking also tends to promote the popular rather than the obscure. In a highly connected community, a small preference for a particular item can rapidly amplify and spread its appeal. Increased social connection may result in bigger blockbusters, but makes predicting them all the more difficult.

In fact, the proliferation of choice may make people more reliant on the recommendations of others to determine their own preferences.

In practice, people are not entirely rational, making independent decisions based on optimising from all the available information. Mass media reflect a social context which is part of a culture of shared ideas which gives us a sense of identity and belonging.

To which one might add that blockbusters do not simply emerge organically out of the combined choices of a population of consumers. These choices are manipulated by many marketing influences that aim to produce popular, profitable products.

It is therefore important to take into account the sociology of media, based on an anthropological study of actual behaviour, rather than rely simply on economic ideals of perfect information and friction free distribution.