Predicting Warranty Expense using Reliability Models

Manufacturers set reliability targets and then design and develop products based on an expected product lifetime (i.e. Useful Life). During the Product Development process, progressive manufacturers conduct extensive reliability testing to minimize the risk that products will fail prematurely.  Key factors to consider are the usage environment as well as how customers use the product. Despite these efforts, unexpected failures may occur due to design flaws, manufacturing process changes, excessive variation, or a misunderstanding of the product use environment. Premature failures alienate customers and significantly impair brand and company reputations. It can also increase litigation risks. Field failure data should be tracked and modeled to forecast future failures and identify emerging issues that present financial risks to the organization.

Forecasting of future failures may be used to:

  • Predict warranty costs
  • Accrue for future liabilities
  • Set appropriate warranty periods
  • Forecast spare parts requirements
  • Identify emerging product issues
  • Manage customer expectations and customer relationships
  • Drive product robustness improvements

This webinar by industry expert Steven Wachs explains how models developed with life data analysis (e.g. Weibull) may be used to predict how many failures will be expected to occur in the future. The concept of conditional failure probability is key for this purpose. Steven will also review basic life data analysis and will also focus on using the developed model to develop a forecast of future failures, both during the warranty period and beyond.

Webinar Objectives

Many companies struggle to use available field failure data to proactively address risks to their profitability and success. Analyzing field failure data appropriately can help to understand and characterize the reliability of products in the field. The resulting models can then be used to forecast how many units are likely to fail in the future. This is essential for projecting warranty costs and potential failures that may impact customer satisfaction. 

During this live webinar Steven will provide specific knowledge required to develop a forecast of future expected failures both within the warranty period and beyond. Although software will be typically used for this task, the underlying concepts and methods are explained in some detail. After attending this webinar, you will be equipped to apply life data analyses to forecast future failures as well as handle common modeling issues such as the presence of non-homogeneous groups in the data.

Webinar Agenda
  • Making use of field failure data to develop models to predict future failures
  • Typical warranty data setup and formatting
  • Life Data Analysis overview (e.g. Weibull Analysis)
  • Developing a Warranty Forecast using Reliability Analysis Methods 
  • Accounting for Model Uncertainty
  • Identification of Non-Homogeneous groups in the data and accounting for the impact on the forecast
  • Reacting quickly to emerging field concerns
Webinar Highlights
  • How to summarize field failure data for use in common software programs
  • How to use a reliability model (distribution) to forecast the number of future failures (use of conditional failure probability)
  • The importance of handling a non-homogenous dataset correctly
  • How to ensure the uncertainty in the forecast is accounted for using appropriate confidence bounds
  • A tool for detecting emerging issues quickly
Who Should Attend?
  • Design Engineer
  • Engineering Manager
  • Quality Engineer
  • Quality Manager
  • Product Engineer
  • Reliability Engineer
  • Program / Product Manager