Apparently physics can predict which books will sell best based on early sales data, according to an article in Physics Review Letters by UCLA's Professor Didier Sornette.
As with my last blog post, this is an academic examination of print books - so, take it with the usual pinch of salt.Sornette claims that, although individuals are impossible to predict, networks do have a high degree of probability.
He started with the premise that there are two ways a book can become a bestseller:
1. Exogenous shock - Something external, i.e. publicity, gives a short sharp kick that sends sales into the stratosphere.2. Endogenous shock - Interest builds slowly. This is basically word of mouth advertising.
I will point out that he looked at 138 books, and himself admits 14 outliers were in that data set, so this is only a general rule of thumb. It isn't absolute.
Exogenous shock titles peaked typically after specific exposure - A television appearance by the author, a news article, a good book review or an advertising campaign. Clearly, unless you've got a huge budget, or something very controversial so as to be newsworthy, this isn't going to be your eBook.
Endogenous build slowly. These are the books we all talk about, share and love. If a book finds the right fan base then it can spread like an epidemic. The viral nature of social media inc Goodreads, Twitter, Shelfari etc means that we can reach huge audiences over time. It does take time however. You can't force endogenous shock. What you can do is focus on advertising to increase visibility to give the book a chance at building interest. Each exogenous wave is interest you didn't have before - which might then contribute to your word of mouth sales.
The average time from entry to market to explosion in sales, according to the Kindleboards data I examined in earlier blog was around 8 months - that's how long term you need to think. Having another book or two can accelerate things as can publicity, but it's a slow burner. Best sellers, unless of the exogenous kind, do not happen overnight.
NB: Bestseller has no set definition. There are lots of lists. This data was based on Amazon's bestseller lists, which employ a decaying average - allowing for explosions in ranking more easily than the 'total cumulative' model used by, for example, the New York Times.