Development, analysis and applications of a quantitative methodology for assessing customer satisfaction using evolutionary optimization

In this paper, we develop a method for the quantitative modeling of a Customer Satisfaction Index (CSI) function for consumer vehicles. The mathematical model is evolved using an evolutionary computation technique such that the satisfied and dissatisfied customers are equally distributed on either side of the mean satisfaction level. Instead of relying on a conventional survey based assessment, we extract various important features from the service (field failure) data of five different vehicle models and build optimized CSI functions for each. Next, the approach is extended so that a single CSI function can predict the satisfaction for all customers of all five models, thus providing a measure of the market’s perceived quality of one vehicle model relative to another. Different combinations of vehicle models are used and the corresponding CSI functions are validated against the ratings published by Consumer Reports. A sensitivity analysis reveals interesting information about the features extracted from the service data. Thereafter, the CSI function which best differentiates between all five vehicle models is chosen for further use. Finally, we present two applications which use the chosen CSI function to (i) identify high-priority problems for different vehicle models and (ii) data-mine the service data, both of which are very important from Customer Relationship Management (CRM) point of view.