IBM Predictive Warranty
Predictive Warranty is a feature of Predictive Solutions Foundation on Cloud that looks for conditions that lead to accelerated wear and replacement of manufactured products that are under warranty. Such conditions might include variations in the manufacturing process of the product, variations in the quality of vendors' materials that are used in the product, or the ways in which the product is used.
A small delay in detecting the conditions that lead to accelerated wear can cause more warranty claims and related losses. By understanding the factors that lead to warranty claims, you can take corrective actions such as the following actions:
- Improve manufacturing processes to prevent warranty claims.
- Recall faulty products or product batches that have safety implications.
- Set pricing for warranties and extended warranties.
- Evaluate vendors of the materials that are used in the product.
- Replacement rate
- QEWS alerts you when the product's random failure rate exceeds a computed threshold. The threshold can reflect product reliability goals (for example, the product population in the field must not exceed a specified failure rate) or financial liability goals (for example, the cost of reimbursing product warranty claims must not exceed a specified total amount).
- Wear-out
- QEWS alerts you when it finds evidence that product failures are not random in time, but are indicative of wear-out. Wear-out means that products that are in customer use for a longer time fail more often than products that are in customer use for a shorter time. Because wear-out can have serious consequences, QEWS alerts you when it detects evidence of wear-out regardless of how many product units contributed to the detection.
- Sales model
The Sales model identifies variations in product wear and replacement rates according to the date of sale. The date of sale might correlate with in-service conditions, seasonal climatic conditions, a particular customer, or other important similarities.
For example, a product carries a one-year warranty. In cold conditions, the product becomes brittle and wears prematurely. In certain geographies, products that are sold and enter service in winter initially suffer rapid wear, followed by slower wear during the latter part of the warranty period. The opposite is true for products that are sold and enter service in summer. These seasonal variations affect the product wear rates and weighted replacement rates, which are detected early by QEWS.
- Production model
The Production model identifies variations in product wear and replacement rates according to the production date of the product, not the resource in which the product is used. The production date of the product might correlate with the manufacturing equipment operator, the manufacturing process, or other important similarities.
For example, a faulty batch of products is produced during a particular period. The products are installed in resources that have different manufacturing dates. Although the resource manufacturing dates and the product production dates are unrelated, QEWS makes it easier to identify and understand the real cause of the warranty claims.
- Manufacturing model
The Manufacturing model identifies variations in product wear and replacement rates according to the manufacturing date of the resource in which the product is used. The resource manufacturing date might correlate with assembly problems that occurred during a particular period.
For example, due to a short-term problem with the manufacturing process of a resource, some of the products that are used in the resource fail prematurely. Although the resource manufacturing dates and the product production dates are unrelated, QEWS makes it easier to identify and understand the real cause of the warranty claims.
You can adjust the frequency at which data is captured and input to QEWS, and the frequency at which QEWS analyses are run, according to the requirements of each situation. For example, monitoring data from a network of field service personnel might best be done on a monthly basis.