CUSTOMER SATISFACTION SURVEYS: A SIMPLIFIED METHOD TO CREATE A LEVERAGE INDEX USING QUALITATIVE DATA

摘要

When analyzing customer satisfaction data it is very often requested for the metrologist to provide a leverage index in order to identify which attribute of the product or service (predictor) needs to be improved to increase the satisfaction level (dependant variable). This is usually done by the means of a questionnaire with many items each covering an attribute and performing a quantitative analysis using partial least square, Theil's index or neural networks. In practice it is observed that long questionnaires give very few responses (5-10percent). Shorter questionnaires are poor in information but increase drastically the response rate (20 to 40percent) and more interestingly the customer comments are quite systematic (up to 90percent of opened questions are documented). Facing such behavior leads to the question of creating a leverage index out of qualitative data. This paper will present a method to categorize the verbatim in such a way that a numeric contribution of product or service attributes can be measured.