Associate Professor of Marketing and Statistics
Ph.D., Cambridge University, 1998

Summary of research interests

Dr. Liechty’s research is focused on developing new statistical methods and models that arise from real world problems; typically these problems come from areas of marketing and finance. Following are a list of some of his areas of application.

Sequential decision making: Development of methods for incorporating sequential information, via observed behavior or through prior elicitation, into automated consumer decision support systems. The main objective of this research is to develop methods that will allow firms to deliver customized services and products in an automated manner via the Internet.

Data mining: Development of methods and theory that lead to optimal strategies for summarizing and extracting actionable information from extreme size data sets.

Dynamic marketing models: Development of statistical methods for models with parameters that are stochastic processes. Types of models included hidden Markov models for identifying different types of cognitive behavior while reading print ads and Poisson models where the purchase rate changes dynamically over time.

Bayesian statistical methods in marketing: Development of statistical models that entail novel Bayesian analyses and that can be applied to a wide range of marketing problems. Examples include multivariate Probit models, multivariate Poisson models, and Logit mixture models. Application settings include issues in market segmentation and evolution, cross-selling and bundling, product variety and mass customization, and new product design and pricing.

Representative publications

J. C. Liechty, V. Ramaswamy and S. Cohen. 2001. Choice-menus for mass customization: An empirical modeling approach for pricing and demand assessment. Journal of Marketing Research.

J. C. Liechty and G. O. Roberts .2001. MCMC methods for switching diffusion models. Biometrika 88: 299-315.

Last updated: May 21, 2003