Complexity modelling in marketing: A look "under the hood" of the NBD

Abstract

This paper demonstrates how agent based modelling can be utilised to model a complex marketing system. The Negative Binomial Distribution (NBD), a well validated, widely accepted marketing science model is used as the foundation to demonstrate the approach. Using the agent based approach individual shopper (agent) rules are defined that comply with the underlying assumptions of the NBD model. Using these rules and progressing from a simple to a more comprehensive picture achieves two objectives. Firstly the agent based approach’s contribution to marketing is demonstrated and secondly it is possible to ‘see under the hood’ of the NBD model, namely what drives it and how it works. This is designed to assist in general terms, through demonstrating ‘how’ the NBD actually works, as well as demonstrating an approach that might lend itself to further exploration. By using the NBD as a foundation, the opportunity exists to move beyond the constraints of stochastic market modelling and incorporate new agent rules based on other theories (or even conjectures), allowing a means to develop falsifiable hypotheses where previously these theories and conjectures may have been untestable.