|
Abstract:
|
Warranties are agreements between customer and manufacturer, but they are also product attributes to which customers assign value. Manufacturers naturally want to keep their costs low and customer perceptions of value high. Good basic engineering is obviously important in delivering first time quality. But for items that do come in for service under warranty, good diagnostics and highly effective repairs minimize the number of comebacks. Some automotive examples are presented in this talk to illustrate these issues. Some of the analysis methods include basic plotting and analysis of warranty claims data, examining variation in search of best practices, and primitive data mining to reveal patterns of repairs that might be useful in improving diagnosis. In some cases the methods supporting the examples are in routine use, while others are experimental.
|
