Bayesian Price : Threshold Estimation

Abstract

A leading organization in the FMCG sector was interested in studying how the demand of one of their brands experienced changes as the price increased. In particular, they were interested in knowing whether there existed a price-point, called a "price threshold", beyond which there was a major down-slide in sales volume of their brand, and if such a price-point indeed existed, they were interested in knowing to know its value. That is, the problem was to first check for the existence and then estimation of this-price, if there was one. To begin with, the very nature of this so-called price-threshold, in the sense of the kind of behavior in the change in relationship between the two variables, namely price and demand, that was to be expected, was not very clear a priority. Further more, it was also not clear whether indeed there was a price-threshold that was lurking in the given data set or not.