Knowledge acquisition for marketing expert systems based upon marketing problem domain characteristics

Purpose - The purpose of this paper is to present various knowledge-acquisition methods and to show how existing empirical research can be used for mapping between marketing problem domains and knowledge acquisition techniques. The key to doing this is to create a taxonomy of marketing problem domains. Design/methodology/approach - This paper combines a thorough literature review with prima facie conceptualization to map a generic problem domain, and thereby provide guidance in the choice of knowledge-acquisition technique for developers of expert systems in the field of marketing. Findings - Recent empirical research in the field of expert systems shows that certain knowledge-acquisition techniques are significantly more efficient than others for the extraction of certain types of knowledge within specific problem domains. It is found that protocol analysis, while fairly commonly used, is relatively inefficient for analytic problems. In the synthetic problem domain, interviewing proves to perform better for simple problems and worse for more difficult-to-model synthetic domains. Research limitations/implications - The findings suggest that it may be worth exploring some of the non-traditional knowledge-acquisition techniques when working on some types of applications. Further research could offer guidance in choosing the appropriate technique, with the aim of improving the quality, efficiency and development of the resulting system. Practical implications - Designers of expert systems for marketing should consider interviewing and card sorting as the main means of knowledge acquisition for analytic problem domains, rather than protocol analysis as the main knowledge-acquisition technique for analytic problem domains. Originality/value - This paper is the first to suggest mapping between knowledge-acquisition research and marketing problem domains.