The publishing world is an industry that has been around for many years and has produced millions of books. From the books that are published every year, only a few make it to best-seller lists, like the one The New York Times publishes annually. But what makes certain books so successful?

An editor’s job in any publishing company is to acquire books that fit into their guidelines but that also show promise. Many would consider that finding a best seller in their slush pile is a question of luck, of gut instinct, and therefore, there is no easy answer to the question of what makes a book become a best seller.

The twenty-first century, however, has brought many attempts to find the key to a best selling book by both researching and building software programs. One of the most relevant studies in the field is the one conducted by Jodie Archer and Matthew L. Jockers, which was published in 2016 under the title The Bestseller Code. This book is based on an extensive database built with best-selling and non-best-selling books that had the aim to identify the characteristics of best-selling books and to build a machine that was able to identify if a manuscript was going to be successful or not.

The analysis of the different books considered for the study was carried out through Natural Language Processing software that investigated language aspects like frequency of words and sentence length.

As Jia Tolentino points out in her article for The New Yorker, the result of the study identifies the characteristics and topics of best-selling books. The main characteristics range from writing style to character development, and include traits like colloquial writing style, decisive and active characters, fast pacing, short and clean sentences, nouns and verbs over adjectives and adverbs, short and clear titles, and connection and closeness between the reader and the characters, among others. The main topics that best-selling books cover are those connected to technology, family, work, and relationships.

The publication of these findings after five years of investigation revolutionized the publishing world and had everybody wondering if artificial intelligence would in fact become a part of presses and change the way things were being done at that time by, for example, introducing machine learning into the acquisition process and, thus, making editors available to do other tasks and improving the rate of successful books per press.

In fact, a few years later, Jodie Archer brought this study to life and created Marlowe, an AI that analyzes manuscripts and gives suggestions for improvements. This technology focuses on some of the points that were highlighted in the study by giving feedback about plot, pacing, characters, dialogues and narrative, and language (adverbs, adjectives, verbs, nouns, and punctuation), among others.

All the efforts to decode best-selling books show that the industry is working toward understanding what is behind a successful book. Could this be the beginning of a new era in the publishing industry where the acquisition process changes?

I guess we will find out with time.