Abstract:

In this paper, traditional methods used by market researchers to predict consumer behaviour are criticised and the argument put forward that a formal modelling approach is the only way that reliable predictions can be made. Such an approach may involve several types and categories of models, and a typology of models arising out of discussion in the British Market Research Society Study Group on Modelling is described. It is argued that models most likely to be of value in consumer behaviour prediction problems will be Information Processing models and be concerned with the decision making process. They will be microanalytical behaviour models and work via a disaggregated data base rather than 'average' consumers. With the application of the technique of simulation it is argued that they represent a powerful research tool that can solve many of the problems market researchers are currently grappling with unsuccessfully, using traditional methods.