Interesting new entry in my published papers alerts: a paper on
Herd behavior in purchasing books online, which attempts to prove that information on sales and user ratings positively influence purchasing behavior. Which is pretty obvious, if you think of it; with so much informationa available, if you buy a book with low ratings you're on your own.
Although the general idea of the paper was interesting, I didn't like how it was carried out; instead of using real data from a real online bookseller, they used students who had to decide whether they would buy or not; thus, what you are measuring is not really purchasing behavior, but purchasing intention.
Besides, I think that the phenomenon is a bit more complex. Reviews influence not only sales (but not always, as shown by my -now defunct-
HP7 statistics site), but also other reviews; even with declining sales, average reviews were surprisingly stable, which points to a retro-feedback effect that is broken for external reasons.
Besides, another problem is that it really does not prove
herd behavior; there's no flocking, no follow the leader, or nothing pointing to a
madness of crowds behavior. All in all, a bit disappointing. But interesting as a first step.
