Strategies for feedback on reliability and robustness from automobile warranty data

Abstract

Yearly warranty cost of major automobile companies runs into billions of dollars. In addition, market- driven need for wider warranty coverage exposes these companies to the risk of higher warranty spending. Warranty cost reduction programmes have thus become a major objective and challenge in these companies. Towards this, warranty data that capture vehicle failures in field conditions contain rich source of information for useful feedback to designers and engineers working on current and forward model vehicles for reliability and robustness improvements. However, obtaining accurate feedback using warranty data becomes challenging due to issues involving data quality. This paper presents strategies for automobile warranty data analysis that helps obtain meaningful feedback towards reliability and robustness improvements leading to warranty cost reduction. Role of hazard plots in giving clues about nature of noise factors that influence failure is presented. Strategies for stratification of warranty data in terms of customer-reported concerns and/or parts reported to be the cause of failure to prioritise reliability and robustness improvement projects are discussed. Uses of geographic plots in gaining insight about seasonality and location effect are also provided.