When I first joined Bell Labs in the early 1970s, one of the key concepts constantly discussed was the adequacy
of prior controls. My colleagues and I in Bell Labs' Quality Assurance Center spent many hours explaining to senior managers throughout what was then known as the Bell System what we meant by
quality assurance. These discussions were necessary because the executives and managers had limited understanding of the quality-assurance audit inspections. They often thought of the audit
inspections as equivalent to the acceptance sampling described in military standards. Nothing could have been further from the truth. The purpose of the audit check inspections, as they were
called, was to ascertain the adequacy of prior controls. In other words, was the quality control system working?
The reasons for the confusion, however, were understandable.
After Walter A. Shewhart's creation of the control chart in 1924, Bell System (AT&T) quality leaders had slowly implemented quality control throughout Western Electric, the system's
manufacturing arm. Brilliant engineers such as Joseph M. Juran (then in Western Electric's famous Hawthorne plant) had seen the value of moving the control function upstream to the operators and
placing the controls on the process rather than on final product, but others had grasped the concept more slowly. Juran created evening classes at Hawthorne to advance the understanding of
statistical quality control in the mid-1920s, but the implementation proceeded slowly in Western Electric and even more slowly outside the company until the advent of World War II. During the
war, the need for improved quality in the production of war materials, especially ammunition, was the driving force for extending these methods to thousands of sites. W. Edwards Deming is said to
have personally trained more than 10,000 people in these methods during the war years.
But many other methods were also developed during the war years. Most notable were the
series of acceptance sampling plans that later became the well-known military standards, then ANSI/
ASQC standards, and later international standards. These standards were especially
useful for ensuring that goods received from suppliers met minimum quality standards. When buyers had no knowledge of the production or quality systems of the suppliers, these methods provided a
reasonable means for assuring certain minimum standards of quality.
After the war, the creation of the American Society for Quality Control (now ASQ) helped promulgate these
standards throughout the United States. The development of quality assurance methods within the Bell System, however, proceeded along different lines. The role of the Bell Laboratories' Quality
Assurance Center was to assure to the primary users of manufactured goods from Western Electric--the Bell Operating Companies--of the products' quality. The key method for this assurance was the
audit inspection, used to ensure the adequacy of prior controls. Although no one could resist using audit results for identifying occasional lots of bad product, the main use of the audit was not
to make determinations of product quality but of the quality of the processes that produced the product.
The company also used other means for further investigations of process
quality. The Class I audit was similar to many quality system assessments used today. In fact, the methods used in these audits formed the basis for U.S. military standards that were later a key
building block for the ISO 9000 series of standards. The point, though, is that the focus was on process quality. In the former Bell System it made perfect sense to use an independent agency, the
Bell Labs Quality Assurance Center, to operate on behalf of the entire company in setting standards, reviewing quality results and determining where to put new efforts. One company, AT&T,
owned most of the parts of the U.S. telecommunications system. The consumer bought telephone services, not products, from the operating companies. The operating companies leased the phones and
other equipment to their customers and provided all maintenance and repair. Therefore, managing the total life-cycle costs was paramount to AT&T. When these costs were understood, it was
clear to everyone that moving the quality system as far upstream as possible made the most economic sense.
Most other companies operated in far different ways. They weren't
monopolies and usually didn't have good cost data concerning field failures or other customer costs. They often had design centers that were quite disconnected from their manufacturing operations
and purchasing organizations located far from the design or manufacturing centers. The service departments also often acted independently, and in some companies, these departments had actually
become profit centers. Field failures, repair and service costs, and customer dissatisfactions were far removed from the designers' and producers' vision. In any case, most of these failures'
costs were borne by the customers, not the supplier.
We are also rediscovering how critical the adequacy of prior controls throughout these complex chains are in ensuring that
the entire process works.
About the author
A. Blanton Godfrey is dean and Joseph D. Moore
Distinguished University Professor at North Carolina State University's College of Textiles. Prior to his current assignment, he was chairman and CEO of Juran Institute Inc. E-mail him at email@example.com .