AT&T, Xerox, and Ford Motor Company pioneered Statistical Process Control and Taguchi Methods in the United States with such success, that today, technology development industries, food services, medical facilities and other non-manufacturing environments are all moving toward process control and improvement. What it means is a new and different way to approach the whole idea of quality and success in the market place.

Traditionally, American business has defined success or failure narrowly, in terms of immediate return on investment. Taguchi Methods moves the focus from short-term profits to increased market share. Increased market share drives quality, and increased quality insures long-term profitability.

An important distinction must be drawn between Quality Engineering and Statistics. Taguchi Methods has been confused with Statistics and at times unfairly criticized by statisticians for deviating from classical statistical methods. In much the same way that Physics uses and applies the abstractions of mathematics to the tangible manipulation of electricity and other physical phenomenon, Quality Engineering utilizes some principles of statistics as the underpinnings for an engineering discipline.

A very obvious divergence is clear by contrasting the aims of the two methodologies:
Statistics tells you what has happened
Quality Engineering tells you how to make it happen

Mainstream statistical analysis involves comparison of means, interactions between variables, and results in prediction of means or responses, based on the assumption that the process is stable. The error variance of the responses is always assumed to be constant. This is a laboratory approach. Out on the factory floor, or in the field, noise (variability) is an inescapable fact of manufacturing and production existence.

Taguchi Methods represent a new approach. Quality is measured by the deviation of a functional characteristic from its target value. Noise causes deviation/loss of quality. Noise may be impossible, extremely expensive, or just plain too time consuming and difficult to eliminate. Dr. Taguchi proposes that a more intelligent and profitable approach is to admit that noise is a systemic problem, and then to compensate, or remove, or reduce, the effects of that noise.

Dr. Taguchi proposes that the reduction of variability should be one of the paramount goals in experimentation, and that "Robustness", the resilience of a product/process to:
Outer Noise (Environmental/Systematic conditions)
Inner Noise (Wearing out of materials, components, etc.)
Between Product Noise (Piece-to-piece variation, should be designed into a product or process from the very beginning)

His guidelines are fairly straightforward:
Create products and processes that are robustly designed against all the noises you can think of, using the least expensive components and materials that are adequate to your needs
Selectively upgrade only those components and materials that can cost-effectively reduce variability further, if necessary
Maintain on-line quality controls: measure quality characteristics on the production line, and feed the analyzed results back upstream for process adjustment if deviation from target value is occurring
De-emphasize the use of statistical assumptions in favor of the actualities of engineering. Design experiments with cost savings, technological improvements, and productivity enhancements, as well as quality improvement in mind

When all is said and done, Quality Engineering/Taguchi Methods has to be about increasing profits. However, increased profits do not have to translate into increased cost for the consumer. This is fundamental to Taguchi Methods, and is reflected in virtually all aspects of Quality Engineering.

Dr. Taguchi has introduced the Quality Loss Function , a means of computing the real-world costs of poor quality, as well as establishing how much expenditure on quality is financially justifiable. His concept of Parameter Design is a methodology that maximizes the quality characteristics of a product, without generating any cost increase to be passed along to the customer. Tolerance Design forces the most scrupulous scrutiny of dollar-for-dollar return on any quality-improving investment.

Quality Engineering encourages cost-reducing technological advancement by constantly refining the target values of our process/product.

This emphasis on cost versus quality improvement allows for the presentation of a high quality product at a competitive price. The point is not to price you right out of the marketplace by instituting quality improvements that no one can afford. The point is to offer a high level of quality at a standard quality level price. That builds market-share. And that builds profits.