Predicting a Bestseller

This weekend, we went to a conference done by New York Times bestselling author of The Five Love Languages, Gary Chapman. Since his first book was first published more than 20 years ago, he has written more than two dozen books. Some have done well, but none have done as well as his first book. Why? Certainly everyone has a different theory, and maybe there are as many reasons as there are theories. But, after all the guesses and conclusions, it boils down to one thing: Gary Chapman figured out what people cared about (i.e., marriage), and he did studies to figure out how to help them. Then he wrote a book that condensed everything he'd learned into simple language that was easy to understand and apply.

On Digital Book World today, Daniel Berkowitz wrote an intriguing article about predicting a bestseller. He analyzes data and gives examples, drawing from his review on the upcoming book, The Bestseller Code, which seems to claim that an algorithm may be able to predict if a book will be a bestseller. Berkowitz marks that this process may be perfected someday, but will never be %100 accurate, simply because of the unpredictably of human taste and fashion.

This is true—what people want to buy and read will always be changing, based on fad, trend, even season of the year. It's impossible to predict what people will buy, but based on past and present best-sellers, people usually want one of two things. Come back tomorrow to see what they are.