Warranty Cost Prediction Using the Mahalanobis-Taguchi System
In any competitive market, businesses need to differentiate their offerings. Many industries use a warranty as an offering that not only differentiates the product or service, but also extends the customer’s expected durability of the product or service. By offering to share the risk of a customer’s purchase in the form of a warranty, a business assumes the cost of poor quality and therefore should have a vested interest to reduce this cost thus maximizing its profits. Warranty cost prediction methods are used by many leading companies, but these methods typically only focus on failure reduction as cost avoidance. Additional dimensions should be considered when predicting warranty cost. This paper presents the application of the Mahalanobis-Taguchi System to forecast warranty cost based on multivariate data. Using Mahalanobis-Taguchi System, a product or service’s warranty cost can be predicted using multiple characteristics thus providing a more complete estimate. The Mahalanobis-Taguchi System also enables businesses to estimate their warranty cost as early as conceptual design, due to the method’s ability to forecast based on small, correlated data sets. A case study utilizing warranty data is also provided to illustrate the technique.