Some tools to predict failure probability on products recently launched on the market: A mobile phone case study

Abstract

Prediction about the failure probability of a product or system is essential in reliability engineering, where a decision‐making process is carried out on a regular basis. This paper proposes a novel index, that allows to predict the failure rate on products recently launched into the market; its calculation is based on sales information and failure reports. Maximum likelihood method is used to derive a mathematical expression for this index, and prediction intervals are obtained throughout a bootstrap approach. A good performance regarding the mean square error and coverage probability is observed in the simulation study included here. Monthly sales data and mobile phone failure reports, provided by the company in charge of the warranty policy process, were used to illustrate our proposal that emerged from an academic‐industry collaboration.