What makes an effective quality system in manufacturing?

We all know that quality has profound effects on company performance and customer satisfaction. We all know that the effect of poor quality is never good, to say the least. We also know that it can be very difficult to predict just how bad the effects of poor quality really are going to be. The outcome depends on a number of variables, ranging from the customer affected to the degree of failure to the actual cost of the failure.

So questions such as How much should we expend on quality assurance and control? and How should we organize the quality assurance and control activities? always seem to involve trade-offs between the cost of providing adequate quality and the cost of failing to provide it—even though we often don’t really know just how much the latter is.

We do know two things: In any economically justifiable quality system, the potential cost of failure caused by poor quality must exceed the actual cost of a quality system designed to prevent or minimize those failures. We also know that the cost of a system that ensures perfect quality at all times is very high—theoretically infinite. But beyond these two notions, everything else related to quality is open for discussion.

Some Definitions

Quality is used here to mean the percentage of a product that satisfies a predefined set of rules or specifications. This gives us the quality level, measured as the percentage that is acceptable or the parts per million (PPM) not acceptable.

Generally, the higher the percent (or lower the PPM), the higher the quality. Note that, in most applications, providing quality levels well beyond what is required or expected generates no real competitive benefit but can generate higher costs.

Quality assurance (QA) as used here means the set of defined regular tasks a company has designed to ensure a product meets or exceeds defined quality levels. Quality control (QC) means the execution of these tasks. From a practical standpoint, QA focuses on preventing defects and QC focuses on detecting them.

The cost of quality (COQ)—sometimes called the cost of nonquality—is the sum of the costs of prevention, detection, and failure. (Editor’s note: The “Financial Ratios & Operational Benchmarking Survey” from the Fabricators & Manufacturers Association International^® will include COQ starting in 2016.)

The COQ generally has a bathtub-shaped curve. At poor quality levels the COQ is high primarily due to excessive failure costs. COQ falls as quality improves, usually dramatically, but rises again as quality improves beyond the inherent quality level allowed by the capabilities of the production system. These costs come from the high costs of prevention and, especially, detection.

Design Considerations

So how do we design a quality system that actually works, one that ensures we meet our internal and external quality needs without either costing too much or becoming a huge impediment to flow or flexibility in production? This question often has more than one answer, and each can result in similar performance but differ in overall costs in dollars, flow, and flexibility.

The cost of the prevention and detection relates highly to product complexity; that is, the number of things that have to go right. COQ is also highly related to the capabilities of the people and machines building the product.

What is being built should be well within the overall and individual capabilities of production resources. This is the heart of statistical process control and Six Sigma. Operating at the edge of the capability envelope is asking for trouble, with rising inspection and failure costs. Operating beyond the capability envelope is not just asking for trouble, it’s guaranteeing it.

It’s also possible to have more than one quality system in a plant, but only if the dollars associated with each justify it. For example, a plant may build one- or two-operation products as well as complex assemblies, and each may have very different customers and requirements. One all-inclusive system might be too expensive in the first case or woefully inadequate in the second.

A quality system designed to accommodate both could work, but its design must be scrutinized and thoroughly examined for COQ. That said, it also can be pretty difficult to manage multiple quality systems, especially if the same shop floor personnel perform the QC tasks for both quality systems.

Try to avoid “gluing on” ad-hoc special inspection, special sorting, or “special this, special that” appendages to the mainstream quality system (usually quality control). You can do it, but it’s very risky. When you look at the actual COQ, you usually find that the costs of this appendage are much higher than those for the mainstream quality system.

One of the most common root causes of poor quality performance relates to product design. A flawed design that’s not conducive to economical production typically leads to much higher-than-necessary COQ. The problem usually relates to the designer being unaware of the true capabilities of the materials, people, or machines that make the product. This makes prebuild design reviews an important part of an effective prevention scheme.

Inspection stations in the product flow lanes—that is, inspections are performed before transferring products downstream—can clog or constrain flow if every piece must be inspected. In some applications, such as mission-critical (military, aerospace, life/death) products, it is required. But generally this structure is not amenable to rapid-cycle-time production and should be avoided.

For most commercial and industrial applications, the inspection part of QC can and should be done by the operator making the part. QA sets the rules, provides the calibrated measuring devices, and audits the inspection procedure. This forces production to be part of the quality system—and that’s a good thing.

However, QA should remain independent of production. Virtually all OEM and industry-standard quality systems require that independence. In any dispute between QA and QC, QA must have the upper hand.

Form Must Meet Function

Quality systems have numerous other practical design considerations, and most depend on the company, its situation, and available resources. But they have a common thread: They incorporate COQ, customer and industry quality demands, and a company’s specific set of value propositions. Together, these factors ultimately determine whether or not a quality system is effective.

If a company has a value proposition that states it produces products with excellent cosmetic fit and finish, then it must design a quality system that supports it. A quality system should never put that value proposition at significant risk. The QA portion of the quality system must regularly verify that processes behind those cosmetics are extraordinarily capable—in machines, skilled employees, and the procedures they follow. In essence, these capabilities allow that value proposition to exist. The same can be said for any other value proposition, including those related to service.

Most custom fabricators are not big enough to justify a separate QA department. Actually, most, if not all, standard quality systems do not require it. What they do require, however, is independence from the production management and a reporting structure in which QA reports directly to the organization’s head person.

It’s fine to have a “manager of engineering and quality assurance” but not a “manager of production and quality assurance.” And again it’s perfectly fine, even desirable, to have production people perform many of the QC functions, especially the physical measurements. But QA must set the rules and have the final say.

An effective quality system does not have to involve burdensome and expensive organizational structures. In fact, these are often the least effective. But the critical tasks of QA and QC must be done. (A number of good sites on the web detail the common tasks.) If you list the tasks, employ practical design considerations, and closely examine the skills on hand, you will uncover a number of options and needs. It then becomes a matter of choosing the one that best fits.

The metric that tells you if you’re effective is COQ. Measure it, track it, and improve it.