Modelling customer choice

The purpose of this paper is to introduce a relatively obscure analysis technique called logistic regression, and examine some potential applications to hotel revenue management. When modelling customer choice, many ‘traditional’ statistical techniques fall short, and this paper will describe how, using specific examples. This paper uses a hypothetical example of customer choice — modelling the relationship between income and a customer's decision to purchase a sport utility vehicle (SUV). We know anecdotally that these two are related, and common sense would support this, since SUVs tend to be expensive. The data also support the claim of a relationship between income and purchase decision. However, to benefit from this knowledge, we need to understand the relationship — this is where logistic regression is quite useful. At the end of the paper, some other potential applications are explored.