James Allen (Jim) Smith
President and CEO
Cambridge Management Sciences
St. Petersburg, Fla.
Airworthiness Reliability Consultant
Assume that we own a company that makes toasters. And the toasters are sold with the advice to customers that the function of the device is to toast breads (and, perhaps, other specified items such as muffins). Without our encouragement, a customer purchases one of our toasters in the expectation it will function as a radio receiver. Not surprising to anyone but the customer, the toaster’s ability to receive and amplify radio signals is nonexistent. The customer is not satisfied. Do we, as toaster makers, have a quality problem?
According to the currently fashionable definition of quality, which was first proposed by Armand Feigenbaum and more recently adopted by the ASQC — quality is what the customer says it is — our toaster has a quality problem because it fails to satisfy the customer. Clearly, however, such a Lewis Carroll world is irrelevant to business management. Consequently, we define quality as the degree to which a product provides all the features promised by the manufacturer for a specified period of time under specified conditions of use. As it happens, this is also the definition of dependability. Therefore, we define quality as dependability.
Design and Quality
Features promised in the dependability definition can also be termed the product’s promised functions. Consequently, the maximum extent to which design engineering departments permit products to provide promised functions can be defined as functionality, an overused word, but one to which we will attach special meaning here.
Functionality and dependability are closely related but not identical. One difference is that dependability contains three other attributes besides functionality — manufacturability, testability, and durability. (We will consider those attributes next month.)
A second significant distinction is that functionality is in large measure latent while dependability is achieved reality. That is, functionality represents potential quality while dependability is the product’s realized quality. Realized quality can never exceed potential quality. On the other hand, realized quality can, and often does, fall short of potential quality. In other words, dependability cannot be greater than functionality and will often be less.
Design alone determines functionality. But activities besides design engineering — including, but not limited to, production, quality assurance, and supplier operations — determine whether dependability lives up to functionality. This is why a perfect design (i.e., perfect functionality) can still end up with subpar dependability.
Features relate to quality only by their absence. So quality problems exist if a product fails to provide any specified features for the specified period of time according to the specified conditions of use. Conversely, our operational definition of quality says that any product which provides its specified features for the specified time period under specified conditions of use has perfect quality.
Selecting features is outside the mandate of design engineering. Marketing departments specify both the features to be included and the anticipated selling price. (Whether customers are attracted to that set of features at the offered price is a separate matter entirely and does not concern designers at all. Features are selected by marketing for sales purposes only.)
This does not mean that satisfactory design engineering entails merely following directives from marketing personnel. Design-engineering departments, for example, should assess and advise marketing on the feasibility of including requested features. Additionally, engineers keep abreast of emerging technologies and should draw attention to newly available features which marketing personnel may have overlooked. Therefore, advice from design engineering to marketing is both reactive (evaluating the practicality of features requested by marketing) and proactive (informing marketing of newly available technologies).
Ultimately, therefore, design engineering revolves around meeting specifications. This concept seems straightforward on paper but in practice has given rise to many controversies over the years. One of the most prominent controversies is the disagreement between Philip Crosby and Joseph Juran about the relative merits of their respective “conformance to requirements” and “fitness for use” concepts of quality. While Crosby and Juran each thought the other wrong, in practice there is no meaningful difference between the two concepts. Both “conformance to requirements” and “fitness for use” can be summarized as “ensure the product meets the specs.”
Note that no reference has been made to the customer’s needs. This is because needs enter the equation only when they correspond to the customer’s desires. Sales result from meeting customer desires and so are based on the marketing department’s interpretation of what customers want. Since customers typically have difficulty identifying or fully appreciating their own needs, needs generally have little impact on product demand. But identifying and meeting customer desires (to which the term “requirements” is normally applied) determines a company’s ultimate success or failure. Specifically, three possible relationships between specifications and requirements can be identified:
1. Specifications are set below customer requirements.
2. Specifications equal customer requirements.
3. Specifications exceed requirements.
Products are only optimized when specifications exactly equal customer requirements.
Whenever specifications exceed or fall short of requirements, the company suffers. As contemporary quality- management theory emphasizes, setting specifications below requirements or needs will certainly not help the company prosper. Many purchasers of underspecified products will ask for refunds while others will not buy the product again and may tell acquaintances of the product’s shortcomings. In some cases — where use of the product as provided endangers the customer’s health, for example — the manufacturer may be forced to retrofit improvements to sold units, recall all units and face possible legal actions from injured customers (or their estates). All told, insufficiently demanding specifications can easily drive companies out of business.
While instances of underspecification are certainly not unknown, few design engineers today are guilty of setting their sights too low. Far more commonly, designers set standards unreasonably high. Engineers occasionally believe the company’s interests are best served by including features that are not part of the specifications originally established by marketing. In this respect, they operate very much like quality inspectors who reject acceptable product for touchup and rework. Since potential customers will not be informed of the additional features and thus cannot be influenced by them in their purchase decisions, overzealous design does not translate into higher quality, or, for that matter, greater demand.
Features that go beyond specification most often relate to variation in performance such as higher signal-to-noise ratio in communication equipment. Less commonly, functions that weren’t in the product specification are added. If the additional features increase product cost, the design engineer has exceeded his mandate and, assuming marketing knows its business, harmed the company. This facet of design engineering’s duties — to ensure that features do not exceed requirements — is seldom mentioned and even more rarely obeyed. Still the issue is pivotal to every manufacturer’s success. Design engineering must discourage designs that increase costs by exceeding specifications as conscientiously as it shuns designs that fall below specification.
Meanwhile, designers must not accept unrealistic marketing demands. One such unrealistic demand would be features that exceed the company’s technical competence. Another example would be unrealistically short development cycles.
Of course, design departments are always under pressure — both selfinduced and externally imposed — to deliver cutting-edge features in record time. These pressures are inevitable in a modern marketplace which rewards only those companies that constantly add new features despite shorter and shorter development cycles. More and more, marketing’s search for new features and shorter development time causes engineers to rush into production before designs have been proven and include features based on technologies which neither engineering nor manufacturing have yet mastered.
The consequences of unsuccessfully pushing the technological envelope can range from disappointment to embarrassment to calamity. If the company hasn’t publicly announced plans for the product and they can be abandoned before investments in product-specific materials and equipment are made, the only losses are internal. Despite lost development funds and bruised egos, the net effect may even be positive if the company acquires additional expertise or just new-found humility that prevents future costly acts of hubris.
After contracts have been signed or marketing programs begun, the costs of failure increase substantially. Failure to deliver contracted features in a timely manner may lead to customer lawsuits that could jeopardize the contractor’s existence. And bringing the product to market even though the manufacturer knows performance will be less than specified undermines the company’s credibility and angers customers.
Committing to features that prove undeliverable is widely perceived as the fault of overzealous sales personnel. Unquestionably, sales and marketing departments do enter into unfulfillable contracts for blue-sky products. Indeed, some companies seem to specialize in promising much more than they can deliver. But not all the blame can be attributed to sales personnel. Naive or overconfident design engineers commonly commit to delivering technological breakthroughs without any experience to back up their selfassurance.
Unfortunately, conflicts between dependability and market share stemming from time-to-market pressures are not always easily resolved. While the relationship between time-to-market and flawed designs is almost universally inverse (though unpredictable), the quicker a product makes it to market, the greater the likelihood it will contain the most advanced features and thus command premium prices. Provided the rush to market does not unduly compromise dependability (a provision for which, unfortunately, there are no guarantees), the financial benefits of being first with attractive features are simply too significant to ignore. This is where freedom from the burden of rigid quality orthodoxy proves especially worthwhile.
Despite what mainstream quality advocates claim, higher quality is not always consistent with maximizing the company’s long-term profits — the company’s overriding goal. There will be times when reaching market first warrants some compromises in dependability. Enlightened design engineering departments cooperate with sales to determine the combination of time-to-market and dependability that moves the company closest to its goal. One important activity in finding that middle ground is joint preparation of cost-benefit analysis. Such interdepartmental collaboration constitutes a superb example of quality optimization at work.
Marketing (and, therefore, engineering) must take durability as well as features into account when setting specifications. Durability, in turn, incorporates two main issues:
1. Environmental extremes in which the product will operate dependably.
2. Stability of performance over time.
Greater durability usually coincides with higher costs, although some durability enhancements are available through cost-free process refinements. Deciding the proper level of durability, therefore, requires balancing customer requirements and willingness to pay against product cost.
Durability can be improved by the use of more rugged components or materials. In the electronics industry, for example, some elements that could be construed as ruggedness might include physical sturdiness, resistance to fungal growth in tropical climates, and the ability to withstand rapid thermal changes.
Most components and materials are available in various grades. Referring again to the electronics industry, printed- circuit boards can be fabricated from a wide range of materials ranging from resin-impregnated fiberboards to complex glass laminates. Components for printed-circuit boards are often available in through-hole or surface-mount and leaded or leadless configurations. Packages may be plastic or ceramic, and in some cases, circuitry is not packaged at all. Variation from rated values can be negligible or significant. The possible permutations and combinations of components used to perform what are nominally identical functions can be overwhelming. And the design engineer must take account of all these options.
In a relatively small number of applications, particularly those where lives are at risk, the decision is straightforward: use the most dependable (i.e., least variable and most durable) components regardless of price. But most of the manufacturing world is not free from cost constraints. Even if customers will pay enough that profits can be made despite using the most costly components, higher product cost that doesn’t translate into tangible market benefits cannot be tolerated.
Compromises must be made between lack of variability or other significant performance characteristic and cost. Trade-off decisions are traditionally made by the design engineer. But there are definite advantages to including representatives from other departments — primarily quality and production — in the decisions. Certain component attributes that might seem unimportant to the design engineer will be recognized as very significant by other departments that work with the actual product.
Engineering departments typically keep a list of specific makes and models of components and materials that have been proven acceptable for use in various grades of the company’s product. This approach normally works very well and, more importantly, is usually the only feasible way of specifying parts. Occasionally, however, a previously unused combination of approved components will operate outside tolerances. This design flaw will only be apparent when prototypes are built and tested.
The only rigid guideline for optimizing design according to variability and component cost is that the upper and lower performance extremes possible with stacked tolerances must fall within specifications. All other choices of components must be based on balancing costs and benefits for which no single set of rules applies.
The Design Engineer and Quality