A Weak‐Form Expert Evaluation of Customer Profiling Models

In this paper, we evaluate two models that were used for customer profiling: 1) a market basket analysis (MBA) model and 2) a customer segmentation model. The models were based on actual customer purchasing data from a large department store for the period 2007-09. A weak-form evaluation method, consisting of qualitative interviews of experts from the department store, was used. The questions focused on information quality according to the DeLone and McLean framework and were derived from the Doll and Torkzadeh model, i.e., they covered content, accuracy, format, and ease of use aspects. Seven experts from the case company took part in the evaluation process. Before the actual interview, the experts were asked to fill out a questionnaire regarding their background and current access to timely sales information, i.e., how often and how useful the information was that they were receiving at the moment. Later, during an interview, the evaluator discussed the questionnaire with each respondent. Then, the experts were presented with information on their customers’ buying behavior based on the results from the two models, i.e., the MBA- and the segmentation model. After each presentation, the experts were asked to respond to fifteen statements and four open-ended questions. All in all, the information gained through the MBA- and segmentation analyses was rated highly (4-5/max 5) by the experts. The experts considered the information gained with help of these models to be valuable and useful for decision making and for making strategic planning for the future. This implies that the models could be of valuable use for managers working within CRM, e.g., planning marketing campaigns, product range planning, service development, planning of store layouts, operative and strategic planning, and for further developing loyalty programs.