Observations on Soft Computing in Marketing

Abstract

Marketing managers make use of a variety of computer-based systems to aid decision- making. Some of these models would be considered “hard” models in the sense that they are based on quantitative data, usually historical data related to some type of market response and some empirically derived functional form of the relationships among actions in the market and market response (Hanssens, Parsons and Schultz 2008). Such models have been widely employed to decisions involving pricing and promotion, advertising scheduling and response, product design, and sales call scheduling, among others (Lilien and Rangaswamy 2006). These models, while very useful, require very rich data, as well strong assumptions about the generalizability of historical data to future events. These are assumptions likely to be less and less valid in an increasingly volatile world that includes regular introduction of new means of communications and product/service distribution, as well as new product and service innovations.