Output details

Predicting the long tail of book sales: Unearthing the power-law exponent

<16> Long tail sales have become a phenomenon in e-commerce, where the accumulated volume of sales in the tail becomes significant. A novel method for the estimation of the tail sales is proposed based on a generative model, which can be used for sales analytics in relation to predicting and validating tail sales volumes. The problem of fitting power laws is inherently difficult, and the generative model provides an alternative based on parameters related to the rates at which new products are introduced and existing ones discontinued. The paper builds on previous research by the authors on stochastic evolutionary models.