Product vs. System Quality
Nonconformance doesn’t always mean the system has failed
By Dale K. Gordon
WE LIVE in an era in which complex and technologically advanced products are produced on a regular basis. Little concern is given to the engineering and advanced process capabilities that are required to produce them.
The initial awe of the developments forged during the dawn of the industrial age, such as locomotives, machine tools, steam engines and architectural wonders—made possible by advancements in materials and production capabilities—has faded to indifferent acknowledgement.
With advancements in precision manufacturing, material properties, software and process controls, there is an expectation that process variation can be precisely controlled and anomalies prevented from reaching customers.
From the early times of manufacturing, when the shift was made from artisan manufacturing to interchangeable parts and mass production, inspection of characteristics, components and product were sufficient controls to meet customer needs and demands.
Not to oversimplify, but as the complexity of the processes and products has increased, we have found that the process of inspection is insufficient to reduce the risk of product failure leading to injury and harm, let alone that the customer or end user will reject the product as not being fit for use.
The ability to produce complex products ranging from aircraft and spacecraft to advanced lifesaving medical devices, from computers and microprocessors to the host of industrial and consumer products requires complex internal organizations. These organizations include a product design and development process that must work with production and support functions while also being aligned with the customer’s expectations—meeting functional requirements and consistent on-time delivery at a competitive price.
This complexity includes advanced inventory management systems, complex hiring and training of workers with specific skills and environmental management of the workplace. Also required are financial and accounting systems to provide the needed capital for investment in precision equipment to produce the product, program and project management specialists to assure on-time and on-cost completion, and significant planning and development of manufacturing capabilities.
Add to that the intricate problem of transferring knowledge and requirements to a vast, global supply chain that is coupled with internal checks and procedures to make sure all the aforementioned activities work together harmoniously and seamlessly.
The complexity of the modern manufacturing organization and the need to ensure that all aspects of business processes are focused on meeting organizational objectives, including the primary mission of customer satisfaction, has led to the development of quality management system (QMS) standards such as ISO 9001.
The purpose of a QMS similar to or based on ISO 9001 is to help reduce the variation not only in the product, but also in the complex and integrated business processes on which we have become so dependent.
We know that the late W. Edwards Deming defined a system as “a network of functions or activities (subprocesses or stages) within an organization that work together for the aim of the organization.”1
In the ISO 9001:2000 Handbook, Jeff Hooper wrote, “The system approach to management is a quality management principle that states: identifying, understanding and managing interrelated processes as a system contributes to an organization’s effectiveness and efficiency in achieving its objectives.”2
Nowhere in this statement does it say anything about an organization’s accuracy in meeting the customer requirements for the product being delivered. This brings us to a long-standing question from some in the quality profession about answering mail from customers.
Often, after being on the receiving end of nonconformance or poor product quality, customers ask what, where or how the failure occurred in the QMS to allow this condition to reach their facilities.
Some are perplexed about how to answer because the truth is that the system might not have failed at all, but the nonconformance was a result of common cause variation in the business processes.
This allows for a certain amount of variability to exist and ultimately results in an overall process that is outside the capability of meeting 100% of the customer requirements 100% of the time.
Concept of risk
The understanding that no system is 100% error proof is one of the underlying reasons we have the terms “producer’s risk” and “consumer’s risk.” It doesn’t mean the organization shouldn’t improve its processes to reduce these types of errors, but instead that the existence of a nonconformance does not necessarily mean the QMS has failed.
This topic was briefly debated years ago when the process of recalling Bridgestone/Firestone tires installed on certain model Ford Explorers vehicles called into question the capabilities of a QMS and its relationships to the prevention of a nonconformance from reaching the customer.3
Indeed, some nonconformance can be directly linked to a departure or failure to follow the business processes, or a lack of a necessary process to prevent the nonconformance. But a nonconformance might also occur due to system capability or process averages found to be naturally occurring or just from the inherent intricacy of moving necessary information within complex organizations and the business processes being used.
One example of when the system is stable and functioning and yet failure occurs is the inability to always meet customer delivery requirements. In one situation, which I personally believe to be common, customer demand for product rose 20% due to increased sales. The organization gladly accepted the additional orders, which increased production output needs even though the company was already struggling to meet current orders.
During the planning of product realization (clause 7.1 and 7.2 of ISO 9001), management looked at stated capacity based on optimal conditions and decided there was enough excess capacity to handle the increase. Yet when time came to deliver on the customer due dates, shipments were missed and the customer asked, “What failed in your system to allow you to miss a customer requirement?”
Using an example from Understanding Variation: The Key to Managing Chaos by Don Wheeler,4 if we had plotted the output of the organization’s system for many months prior to the increase in demand, we would have seen an operation at a steady state of output regardless of the customer demand.
The steady state of this process could be characterized as a process mean bounded by some upper control limit. As quality professionals, we know that to produce a change in the process mean, some change to the process inputs has to occur or variation has to be reduced.
In this case, the variables could constitute every aspect of what goes on in the organization’s system. Product delivery is the culmination of all of the business processes, and output is a measure of system capability.
Variation is inherent
With respect to the quality system, where was the failure? The failure, if there was any, was management’s not recognizing the differential between stated and real capacity. That is the portion of capacity that was consumed within the organization by common cause variation such as scrap, rework, supply chain delays or engineering changes.
Again, Deming and Henry Neave have told us variation is inherent in our systems:
Most losses are unknown, often unrecognized, not even suspected. We must learn to look out for two kinds of mistakes, both of which cause huge losses beyond calculation:
Mistake one: To react to any fault, complaint, mistake, breakdown, accident or shortage as if it came from a special cause when, in fact, there was nothing special at all. In other words, when it came from the system: from random variation due to common causes.
Mistake two: To attribute to common cause any fault, complaint, mistake breakdown, accident or shortage when it actually came from a special cause:
There is no way of always choosing correctly between the two types of causes, and there never will be. So we need knowledge of procedures aimed at minimum economic loss from these mistakes. We need knowledge of the capability of a system and the knowledge about losses from demands that lie beyond the capability of a system, demands often made through the mechanism of management by objective.
We need knowledge about the interaction of forces, including the effect of the system on the performance of people. Interaction of forces may work for good or ill.Interaction of forces may reinforce efforts, or it may nullify them.5
The concept is that within all the processes a business has to execute, there is inherent variation. If the variation in each process is accumulated like a reliability function multiplying each successive amount to get a total system variation, then the total variation could exceed an acceptable system limit for preventing any nonconformance from getting to the customer.
Even in a reliability function, in which we put items in parallel and not in a series to improve the ability of the system, such as contract reviews, engineering design reviews, inspections, calibrations, audits and configuration management, there is still an accumulation of variation.
Even devotees of process control and advanced quality methods are not immune. Toyota, a recognized leader in manufacturing quality, in its bid to grow and become the world’s largest automaker by volume, also had more than double the number of recalls industrywide in 2005 than 2004, even though the United States registered a slight decline overall.6
The reality is that as organizations and systems get bigger and the products being produced become more complex, there is greater opportunity for variances in the business processes to have an effect on the product being delivered to the customer.
We can ask how the increased production objective affected Toyota’s QMS. Did the additional stress on the system allow more variation to creep into all of Toyota’s processes, driving the overall distributions further outside its control limits (the QMS) and, therefore, actually predicting the increased recalls?
This is why the ISO 9001 QMS structure is built on Deming’s plan, do, check, act cycle as an interactive function, not a singular one. The QMS is all about understanding where the process variation exists and continually improving to meet customer needs.
Should customers lower their expectations and not expect 100% conformance to requirements? They absolutely should not. When a nonconformance does occur that reaches the customer, is that a failure of the organization’s QMS? Not necessarily, but it’s definitely another learning opportunity to reduce any system variables that might allow a nonconformance to occur and influence customer satisfaction.
For that situation, we have a corrective action process built into the QMS. The best protection of quality and customer satisfaction is to make sure the corrective functions of the QMS are effective and functioning well.
Reference and Note
- Henry R. Neave, The Deming Dimension, SPC Press, 1990.
- Charles Cianfrani, Joseph Tsiakals and John E. “Jack” West, eds., The ASQ ISO 9000:2000 Handbook, chapter two, Jeff Hooper, “The Process Approach to Quality Management System,” ASQ Quality Press, 2002, p. 16.
- Susan E. Daniels,“Tire Failures, SUV Rollovers Put Quality on Trial,” Quality Progress, December 2000, pp. 30-46.
- Donald J. Wheeler, Understanding Variation: The Key to Managing Chaos, SPC Press, 1993.
- Neave, The Deming Dimension, see reference 1.
- Joe Benton, “Toyota Recalls Near 800,000 for July,” consumeraffairs.com, www.consumeraffairs.com/news04/ 2006/07/toyota_recalls.html.
Dale K. Gordon is vice president of quality for MPC Products in Skokie, IL. He is an ASQ fellow, past chair of the American Aerospace Quality Group and one of the writers of the current AS9100 aerospace standard. Gordon earned a bachelor’s degree in industrial engineering from General Motors Institute (now Kettering University) in Flint, MI, and an MBA from Butler University in Indianapolis.