Reports from the warranty cases constitute the warranty data. Warranty data permits the analysis of the warranty cases, and the creation of more useful and informative warranty information from the entered warranty reports. It can be used to improve product quality and reliability through having knowledge on breakdowns, and for making a root cause analysis, and planning corrective actions for the main quality problems. Usually warranty data is quite fuzzy, and it is not easy to give a profound analysis on the current state of warranty claims, which may affect the quality of warranty analysis. (Blischke &
Prabhakar Murthy, 2006, pp. 132-133)
The warranty data collection and warranty analysis process is described in the following figure:
Figure 7 - Warranty data flow and use (Blischke, et al., 2011, p. 11)
Warranty data collection consists of a combination of warranty claims data and supple-mentary data. Warranty claims data alone does not allow the creation of a proper analysis, but supplementary data, for example sales volumes, production volumes and design mod-ifications, has to be included. This is predicated on the product portfolio.
Warranty data is divided into the three following classes (Blischke, et al., 2011, p. 10):
Product related: which component is broken, the failure type, age and usage as brake-down occurred
Customer related: operating mode and usage, environment, maintenance, user
Service agent related: fixing costs including the repair work, replacements, parts, possible refunds, transportation and freight, serviceman travelling costs
Besides these three classes, data can be, for example, compared to a competitor’s products and their break-downs, warranty cases and warranty costs. Unfortunately, comparison can be quite difficult, because warranty data is usually considered a major trade secret, and the selections behind the numbers given are often unclear for the comparator. (Blischke, et al., 2011, p. 10) (Blischke & Murthy, 2000, p. 41) For example, a company may include sales margin in the spare part prices used in warranty repair reports, while another com-pany doesn’t.
Warranty claim data represents a collection of some measurable features from the cases, such as component usage rate, failure mode or failure times, or the failed part’s name and pictures. (Blischke, et al., 2011, pp. 61-63) This data may enable making a proper analy-sis, but it’s not always assured. Usually the data is unstructured, which requires actions for making conclusions, and corrective actions for the current state. This problem is pre-sent also in this thesis.
Warranty Claims data analysis is based on the warranty claims data collected during the warranty process after product breakdown. The quantity of this data and its quality de-pends on the person who has performed the input, as well as on the company’s warranty process and the warranty report form. All of these factors should be on a necessary level to extract the wanted information out of the input data. (Blischke & Prabhakar Murthy, 2006, p. 132) (Blischke, et al., 2011, pp. 10-11)
1.8.1 Failure causes
Classifying the different failure causes may aid in understanding failures, and in produc-ing better information on the warranty data. Accordproduc-ing to (Blischke, et al., 2011, p. 37), failures can be categorized into six classes based on the failure causes:
Design Failure
Weakness failure
Manufacturing failure
Aging failure
Misuse failure
Mishandling failure
The classification may vary from the above list according to the product and failure char-acteristics.
Classification is not always well-defined, and several failure causes can be sorted into multiple classes, for example, misuse failure can occur when a design failure compels the user to misuse the product.
1.8.2 Warranty Data Analysis objectives, possibilities and threats
Warranty data analysis carries several objectives: it may help obtain useful information on product quality and reliability through the warranty cases. This helps not only in im-proving the current products during the product life cycle, but also in developing better new product models in the future. (Blischke, et al., 2011, p. 10). Every well-filled war-ranty report should be seen as a positive document, which gives information on occurred errors, granting a chance to fix the current problem and compensate the loss for the cus-tomer. Warranty data also enables applying corrective actions to the production process to avoid similar mistakes in the future. Warranty data gives an opportunity to develop better products, and to manage the product quality level and costs of non-quality, which allows a company to outperform in future. (Blischke, et al., 2011, p. 9) These possibilities should always be used in the required way, and ignoring the warranty cases and not uti-lizing the information in product improvement actions, may debilitate a company’s com-petitiveness in the long run. (Blischke & Prabhakar Murthy, 2006, p. 133)
However, data quality can be a big problem while analyzing the warranty data. Warranty reports are mainly produced by non-independent authors, who usually have economical or other incentives, which may affect the data quality. Wrongly produced data may guide corrective actions into the wrong direction, which may cause loss of money and other resources, or cause burying the relevant focus areas under the unwanted and secondary ones. Data may represent the inputter’s intentions, which may turn the actions unfair.
These matters should be evaluated while constructing the warranty reporting process and warranty system, and if possible, the neutral and trustworthy input of the warranty data should be ensured. (Blischke & Prabhakar Murthy, 2006, p. 132)