Quantitative modeling of customer perception from service data using evolutionary optimization

This paper proposes a novel method for using the service (field failure) data of consumer vehicles to estimate customer perception. To achieve this, relevant variables are extracted from the vehicle service data and provided as input to the proposed algorithm which then comes up with an optimized mathematical model for predicting the Customer Satisfaction Index or CSI. The methodology is then extended in a way that allows comparison of the CSIs of two or more vehicle models, thus providing a measure of the market's perceived quality of a vehicle model relative to another. Validation against the Consumer Reports data shows that customer experiences and their consequent response in surveys are indeed a reflection of the numbers the service data provides. However, it is argued that the proposed model is more generic than the Consumer Reports because: (1) it doesn't rely on consumer surveys and (2) it can be used to assess individual consumer level satisfaction.