The Longtail is thoroughly debunked by empirical research

I posted back in July reminding those of us who take current Internet theories such as The Wisdom of Crowds at face value that many of these ideas are primarily marketing tools, rather than tested, research-based approaches.

As a fascinating follow-up to this, Alan Patrick from Broadsight has posted a fascinating analysis of Internet uber-theory, The Longtail, titled: The end of The Longtail?

Alan posts about a recent presentation given by an MCPS-PRS Alliance economist, Will Page, which argued that The Longtail is "fairly completely incorrect".

Page apparently helped Chris Anderson write The Longtail thesis, but has since carried out empirical research on a huge volume of global online music sales. The research found:

"while there was a long tail, it was extremely poverty stricken and much of it is moribund […] even Free doesn’t work – when Radiohead gave away their music for free, there were still 400,000 illegal downloads in the UK. Not only that, they have found that illegal services focus on the “hit head” even more than the average."

Hypothesising further, Alan reckons that most demand curves are Log Normal rather than Pareto Power Law Curves, an opinion strongly supported by one of the researchers.

A full and thorough debunking of The Longtail based on the research can also be found by Andrew Orlowski over at The Register.

As a footnote to this, it is maybe worth adding that the researchers work for an organization that enforces commercial copyright on behalf of composers, songwriters and music publishers.

Technorati tags: The Longtail, Internet Theories, Power Law, Log Normal


  1. This is interesting considering that I oppose both the idea of the long tail and also the idea of empirical research – which is severely debunked in ‘the black swan’. By the time a book has been written (six months minimum) and the publishers have agreed technicalities and business stuff (a technical term) the theory is already out of date…

  2. Hi Jed. You’re right in identifying a major problem with books in a digital age. However, I do agree with Alan that the basic premise of the long-tail – i.e. the Internet opens up a much wider range of options usually restricted by physicality – still holds.
    And while I haven’t read the Black Swan I would say that all research is ultimately flawed which is why frameworks or methodologies are devised to fix our interpretations of the real world.
    This doesn’t make all empirical research wrong or theoretical research right, but it does make some types of research (methodologies)better than others.
    See: for a right headache 😉

  3. Eric Schmidt makes the point that the head isn’t possible with the (zero-profit) long tail in this killer McKinsey interview from last week:
    (i.e. there’s no point to aggregating the tail if you’re not aggregating the head too; value comes from having it all)
    I never read TLT, but does CA come out and say “there’s a lot of profit in the long tail itself” or is he just saying that “there is a lot of activity in the long tail”?
    Definitely a big difference.
    I do know that on his blog Chris has made the point on a few occasions that the interesting thing about the tail is not what proportion of overall sales it is but instead the absolute value of activity down there.
    Alan P is the man either way 😉

Speak Your Mind