Date of this Version
A primary tool that consumers have for comparative shopping is the shopbot, which is short for shopping robot. These shopbots automatically search a large number of vendors for price and availability. Typically a shopbot searches a predefined set of vendors and reports all results, which can result in time-consuming searches that provide redundant or dominated alternatives. Our research demonstrates analytically how shopbot designs can be improved by developing a utility model of consumer purchasing behavior. This utility model considers the intrinsic value of the product and its attributes, the disutility from waiting, and the cognitive costs associated with evaluating the offers retrieved. We focus on the operational decisions made by the shopbot: which stores to search, how long to wait, and which offers to present to the user. To illustrate our model we calibrate the model to price and response time data collected at online bookstores over a six-month period. Using prior expectations about price and response time, we show how shopbots can substantially increase consumer utility by searching more intelligently and then selectively presenting offers.
intelligent agents, utility theory, information retrieval, stochastic modeling
Montgomery, A. L., Hosanagar, k., Krishnan, R., & Clay, K. B. (2004). Designing a Better Shopbot. Management Science, 50 (2), 189-206. http://dx.doi.org/10.1287/mnsc.1030.0151
Date Posted: 27 November 2017
This document has been peer reviewed.