From warranty data analyzed after six months or less of field service, one can predict product reliability and lifetime warranty cost and repair rates… as much as two and one-half years ahead-of the-fact. In addition to reliability prediction, warranty data analyses can be used to identify differences in problem rates between assembly plants, months of production, vendor sources, and vehicle options, and can describe the type of complaint for each part. This close tracking of field performance quickly spots any new problems, helps to identify causes, and guides future plans. This paper describes how Ford Motor Company organizes warranty input data, analyzes it using Weibull distribution programmed on a computer, and uses the results.
