Research by Allenby receives national attention
An editorial in the current issue of Marketing Science (Winter 2008), by Prof. Eric Bradlow of Wharton, singles out a paper by Fisher College of Business Professor Greg Allenby as an example of an academic "home run." Editor Bradlow wants to publish more high-impact research, and his editorial spells out changes he plans to implement to attract more of them. By home run, Bradlow is referring to work that changes how the field thinks about problems. He writes:
"... a common thread is that all home run papers trigger the reaction, “I wish I had thought of that”. Or sometimes paradoxically, “I could have thought of that idea, it is so simple”. Well, not surprisingly home run papers have that aura of simplicity – but usually that is due to hard work on the part of the authors."
Two papers are used to illustrate what he means by a home run. Interestingly, both papers are based in a new Bayesian paradigm that is attracting increased attention within and outside of the marketing discipline. The Fisher College of Business at Ohio State has been instrumental in developing and applying Bayesian methods to problems that are impossible to solve with conventional methods of analysis. Bradlow writes:
"... A second paper I would like to briefly mention, and it is also directly in alignment with a very valuable research area for Marketing Science – mathematical modeling of behavioral processes – is Gilbride and Allenby (2004). In this paper, the authors implement, within a Markov Chain Monte Carlo framework an approach to infer (derive a posterior probability) for the type of decision process (conjunctive, disjunctive, or compensatory) that a given respondent is utilizing. The beauty of this paper is that this is done by augmenting the parameter space with latent indicator variables, a well- established framework but never applied to this important problem. It is a very clever way to “walk around” customer decision space and is a nice blend between mathematics, statistics, and theories of consumer processing."
Research at the Fisher College on Bayesian methods first took national attention with a series of conferences, "Bayesian Methods and Applications in Marketing," held 1999-2002. These conferences offered a 3-day tutorial to practitioners and doctoral students. Travel stipends were awarded to more than 100 students from universities around the world to come to Ohio State and study. Many of these individuals are now contributing to the academic literature, and practitioners attending the conferences have been instrumental in diffusing Bayesian methods into practice. To date, they have had greatest impact in the areas of product policy and pricing, although Bayesian methods have contributed to all areas in marketing.
Ohio State continues to play a vibrant role in the development of these methods. The textbook Bayesian Statistics and Marketing (Wiley, 2005) by Allenby and colleagues Peter Rossi and Rob McCulloch at the University of Chicago, is a best-seller within the statistics community. It is used in doctoral programs at universities across the United States.
Greg Allenby is the Helen C. Kurtz Chair in Marketing at the Fisher College of Business. He is a Fellow of the American Statistical Association, and an Associate Editor for Marketing Science, the Journal of Marketing Research, the Journal of Business and Economic Statistics, and Quantitative Marketing and Economics. He has authored numerous publications that have appeared in leading marketing and statistics journals.