Warranty Spend Forecasting for Subsystem Failures Influenced by Calendar Month Seasonality

Although the need for collecting warranty data originated from financial reasons, it is also extensively used for modeling and analysis to support managerial decision-making in industries. Strategic, tactical, and operational level decisions involving warranty cost very often use warranty spending forecasts that are developed using statistical methods. Existing literature provides warranty forecasting approaches involving variables such as mileage accumulation rate, failure rate, repeat repair rate, and cost per repair. However, there are several key failure modes that are known to be influenced by seasonality. For example, `engine slow to start´ conditions drive a higher claim rate in colder months than in warmer months. Accommodation of such failure modes influenced by seasonality has not been considered in the warranty cost modeling literature. This paper presents a flexible approach for developing a monthly warranty spend forecasting model that incorporates calendar month seasonality, business days per month for authorized service centers, and sales ramp-up in addition to the earlier mentioned variables. On one hand, the model allows development of warranty spend forecasts for entire warranty coverage to support strategic level decisions; on the other hand, forecasts for monthly warranty spend help support tactical and operational level decisions. The workability of the proposed methodology is illustrated using an application example.