The Valuation of Price Experimentation on Profits

• Activity Number: 203
• Type: Invited
• Date/Time: Tuesday, August 10, 2004 : 10:30 AM to 12:20 PM
• Sponsor: Section on Statistics and Marketing
• Author(s): Alan Montgomery*+ and Ting Zhu
• Companies: Carnegie Mellon University and Carnegie Mellon University
• Keywords: experimentation ; pricing ; profit maximization ; functional uncertainty

Abstract

Price experimentation is often advocated as an appropriate methodology to learn about the relationship between quantity and price. The basic premise underlying experimentation is that by introducing artificial variation into prices, the relationship between price and quantity can be estimated more precisely. Unfortunately, this learning may incur direct cost to changing prices, and even more importantly, there is an opportunity cost from experimenting since short-term profits will be less than they could have been. Economic research on experimentation considers this problem under very strict modeling assumptions, which looks quite different from classical statistical approaches to experimentation. This research attempts to bridge this gap and answer some of many unresolved practical questions for guiding managers to determine how much price experimentation is beneficial, when is an experiment warranted, and how should it be implemented optimally.