SOME DATA BASED MODELS FOR ANALYZING SALES FLUCTUATIONS

Abstract

ABSTRACTThe models developed in this paper can be used to study period to period fluctuations in sales. They allow for a heterogenious population of consumers; that is, frequent and infrequent buyers and for one package and multiple package buyers. The data needed for the models are readily available from many sources, particularly consumer panel data. Methods for estimating parameters and detailed numerical examples are given. These models allow a manager to answer the following question: “Which type(s) of consumer is most responsible for the fluctuations in my total sales volume.” Some new results in Bayesian inference are also developed.