AI

Choice Models for Product Optimization and Pricing

Abstract

Presented at Enterprise Applications of the R Language (EARL), London, 2017.

This presentation introduces the topic of discrete choice models, also known as conjoint analysis, and how to model them in R. Choice models estimate the importance of product (or project) features and how customers or other stakeholders value tradeoffs with respect to cost. They are used to determine optimal pricing, product demand, and differences among groups of consumers.

We describe the concept in the context of survey research and discuss why a choice model is preferable to other models. Next, we will see how to estimate such models in R, with an overview of available R packages. This includes a brief walkthrough of code (made publicly available) with an end-to-end demonstration.

Finally, we describe a real-world application at Google, applying the model to understand civic behavior. The technical content is approachable for any R user; the talk will be of interest to R users who are engaged in business strategy, product planning, customer insight, forecasting, or market research.