A Data-Driven Pricing Model for Extended Warranty Under Different Product Failures and Customer Characteristics

Most commercial products are sold with warranties. The prevailing uniform pricing strategy of the product's base warranty and extended warranty often lead to higher-than-expected warranty cost. Information on variations in product field reliability resulting from different customer characteristic such as heterogeneous operating environment and uncertain customer behaviors is not available upon sale and thus overlooked in base warranty cost estimation and pricing. The extended warranty should then be properly priced with the availability of information from failure claims and warranty cost generated during the base warranty period. In this work, we propose a data-driven method that explicitly captures a customer's random effect and his report probability of different failure types and prices the extended warranty according to different degrees of knowledge on customers and products. The proposed method is applied to a real case warranty dataset, and the results reveal that the method can well describe the failure process and the customer characteristics and helps to obtain the optimal price that achieves the maximum profit.