Foundation for Defect-Based Metrics
Two questions form the foundation for a defect-based metric system.
During a recent quality training session, the class discussion centered on quality metrics and the frustration they can cause. With the plethora of metrics being used in many quality systems, people can become confused because of the subtle differences among the metrics. While this column cannot take the space to properly define the metrics, we can discuss the foundation for interpreting defect-based metrics in a meaningful way.
To use a defect-based metric such as Percent Defective (PD), Defects Per Unit (DPU), Defects Per Million Opportunities (DPMO), Parts Per Million (PPM), Rolled Throughput Yield (RTY), etc., effectively, it is important that the organization answer two simple questions. First, what is a defect? And second, how can a defect occur?
What is a defect?
It is my contention that anything not done right the first time is a defect—it’s just a matter of severity. Certainly, this means that we need an understanding of what it takes to do it right the first time. A defect happens when the outcome of any process, whether to manufacture a product or provide a service, does not produce the expected outcome. Of course, this requires that specific conditions of the expected outcome be defined and the conditions of the defect determined in advance.
A defect happens when the outcome of any process, whether to manufacture a product or provide a service, does not produce the expected outcome.
Once determined it is important to remain consistent with these definitions. Consider the following. An incoming unit was declared defective because it didn’t meet the expected outcome definition. After material review board (MRB) consideration the unit was deemed useable; however, the unit’s initial classification of defective should be retained.
Many quality professionals find their organizations are reluctant to adopt a similar position because it adversely affects their internal quality numbers or that of their supply base. Because a unit is useable, some organizations incorrectly conclude it must not be defective. A defect may or may not affect its usability; however, determining the unit useable doesn’t make it conforming and can cloud improvement action.
As an example, a friend purchased a new automobile with a factory-caused surface blemish (which was happily corrected to his satisfaction by the dealership under their warranty plan). By definition the automobile was defective because it had one or more surface defects, but the car was still useable. Reclassifying a defective unit as nondefective does nothing to help resolve the root cause of the defect.
How does a defect occur?
Quality professionals have helped their organizations compile lists of defect families and defect types within those families. Such lists should be as complete as possible in identifying all possible defect types. Additionally, each defect type should be independent and mutually exclusive of others. This allows for the recognition of the occurrence of multiple defects on any given unit.
If a defect family or type is known to occur, include it on the list. Why do we stress this seemingly basic point? Some organizations exclude known defect types because they happen infrequently—avoid this temptation. Also, it is useful to have a defect family or defect type called “other” because there may be a lack of foresight or knowledge to define everything in advance.
As the list of defect types is being developed, it can be helpful to define them in pairs, particularly when examining physical or mechanical characteristics. Consider, for example, the characteristic pair “too long” and “too short.” The temptation is to have just one defect type titled “incorrect length.”
Quality professionals need to be prepared for their management to argue that this level of classification is non-value-added. The quality expert, however, needs to stress that the classification sensitivity is needed because it is influenced by looking beyond the defect to the action the defect creates.
If the “too long” defect, for example, results in a unit requiring rework and the “too short” results in the unit being scrapped, the consequences of the defect occurrence are much different. Different consequences may require identifying and monitoring different defect types. Classifying the defect occurrence as two different defect types can allow for future root cause analysis and result in lower non-quality costs.
Establishing a foundation for your organization’s metric system is critical to managing internal and external quality, thereby improving the bottom-line profitability of the organization.