In the past few weeks a non-fiction book has taken the publishing world by storm: The Bestseller Code: Anatomy of the Blockbuster Novel. Written by a former literature lead researcher at Apple and an associate professor of English, the book has a compelling blurb:

What if an algorithm could predict which manuscripts would become mega-bestsellers?

Girl on the Train. Fifty Shades. The Goldfinch. Why do some books capture the whole world’s attention? What secret DNA do they share? In The Bestseller Code, Archer and Jockers boldly claim that blockbuster hits are highly predictable, and they have created the algorithm to prove it. Using cutting-edge text mining techniques, they have developed a model that analyses theme, plot, style and character to explain why some books resonate more than others with readers. Provocative, entertaining, and ground-breaking, The Bestseller Code explores the hidden patterns at work in the biggest hits and, more importantly, the real reasons we love to read.

The authors, Archer and Jockers, scanned nearly 5,000 novels into a computer, amongst them 500 New York Times bestsellers, and then programmed the computer to predict which would succeed. Their algorithm returned 80 per cent correct predictions.

Most authors don’t set out with the express aim of writing a blockbuster (unless they have fantastically large egos); they are like EL James, who has said, ‘I never set out to do this. Getting to number one in the New York Times bestseller list wasn’t even a pipe dream.’ As for publishers, they do their best to predict what will sell, but will openly admit that there is no exact science to publishing a blockbuster – remember that JK Rowling was rejected over and over again with Harry Potter.

So a formula that can predict a bestseller is surely very exciting for publishers (can it be applied to the ‘slush pile’?), and for authors (by studying this book’s analysis of theme, plot, style and character, can we write a guaranteed bestseller?). The idea of a computer telling us what to write and publish, however, doesn’t prove to be inspirational.

First, most of the findings of the analysis amount to common sense for writers – for example, that ‘human closeness’ is key in a popular book.

Second, there’s just no predicting the mood of a time. As Knopf editor Carole Baron said to The Atlantic, ‘Can you predict the future in literature and art when you can’t factor in the zeitgeist? We’re always surprised.’

Third, where is the art in analysing books, spotting ‘must haves’ and then inserting them into your fiction – to writing not from the heart but to a formula? What would the future of literature be if books were increasingly written to rule, and purely in order to be bestsellers (remember, some of the very best works of literature are not bestsellers)?

Finally, what about meaning – writing and publishing for the love of it and the fun of it? Where is the meaning and enjoyment in success no longer being at least in part random?

What do you think about analysing books and then writing to rule? I would be very interested to hear your thoughts.