Pareto/NBD (negative binomial distribution) is another type of model used to predict the future activity of customers. The Pareto/NBD model uses order history as the primary input, and in particular takes into account the frequency and the recency of orders. While the basic Pareto/NBD model only uses order history, the results become more rich and more accurate when covariates are layered into the model.
The basic Pareto/NBD model simulates two events. It uses a “coin” to determine whether a customer churns and then it uses “dice” to determine how many items a customer will order. The coin is modeled using a Pareto distribution and the dice are modeled using a negative binomial distribution. The more information you have on a customer, the better the models can fit them to a specific distribution and the more accurate the predictions end up being.
Pareto/NBD and related models have seen much discussion in academic literature. The foundation was laid by Schmittlein et al in a 1987 paper and expanded upon in 2003 by Fader et al. For more practical information, Bruce Hardie provides a multitude of tutorials and Excel spreadsheets for using probabilistic models in a marketing context.