A neural net-based model for decision making in marketing

The creation of models able to explain the relationship among sales and the different stimuli that reach the consumer, including publicity, is a very important topic in market research. Interrelations are loosely defined from a quantitative point of view, and studies are usually made at a macroeconomic or aggregate level. This paper presents a microscopic or individual-based model that represents explicitly every consumer, introduces ads as positive or negative reinforcers of consumer behaviour, and sales as results of individual decisions. Several experiments have been run on populations of these individuals or agents, and results checked against what is obvious or well known in reality. These experiments prove that the qualitative behaviour of the model closely resembles reality. Next step is to try to obtain quantitative results out of the model, and put it into an easy-to-use program.