The broad use of the words “Six Sigma” today means many things and it has evolved from its beginnings into a Quality Management System. “Six Sigma” Systems have also been combined with “Lean” Systems, based on the Toyota Production System, as well as the “Theory of Constraints”. Each has their thinking approach, followers, advantages, tools and techniques.
The thoughts and ideas that led to Six Sigma and Design for Six Sigma (DFSS) were actually to address issues of product reliability, not quality. The Originator and Father of “Six Sigma”, Bill Smith (1929 – 1993) first shared notes discussing the reliability of products in two internal to Motorola papers in 1984. The first, from mid 1984, really notes copied and distributed, was called “Early Life Failures in Electronic Equipment”. The second was “Reliability Predictions and Customer Perceptions” dated September 4, 1984. Bill stated, “We are trying to improve the overall Reliability. Units would go through environmental testing in repeated loops of 5 to induce failures. Many failures matched what was going on in the field. Most were Early Life Failures due to latent defects.” The Latent Defects/Unit was proportional to the defects created in components and in product manufacturing.
Within the next year, Bill had convinced Bob Galvin, then the CEO of Motorola, that we had to improve our products and needed new ideas, and he had some. Also in 1985, Bill Smith coins the term “Six Sigma” to describe an expected level of design margin and product quality. It was known by Bill and others that the rolled yields on products, or the number of products going through the entire manufacturing process defect free, were lower (less products defect free) than similar complexity Japanese products observed in factories near Chicago. This was a competitive threat that was clearly recognized. Bill’s concepts were incorporated into existing Motorola new product testing programs and were detailed in a paper by Ken Enger on February 27, 1986 called “Defects per Unit Relationship to Product Performance Reports and Warranty”.
In 1986, Bill, Ken, and Dr. Mikel Harry helped create a course and book called “Design for Manufacturability” intended for Product Designers, Process Designers, Quality Control, Marketing, Manufacturing, Sourcing, and generally all the support people who participated in New Product launches. This course was widely participated in and was rolled out to all business units within Motorola. The concepts of developing more robust products that could function with “wider tolerances” and in parallel, developing processes that had narrower “Variation” and the resulting lower defects per unit counts, led to the first widespread use of the terms “Six Sigma” and “Six Sigma” capability.
Finally, Mikel Harry clearly defined “Six Sigma” in a chapter of his book, “Achieving Quality Excellence: Strategy, Tactics, and Tools” (1986) and the internal Motorola booklet titled “The Nature of Six Sigma Quality” (1986). Mikel explained what Six Sigma Quality is, how to relate capability to sigma level and process yield, and laid the foundation for the strategies and tactics on achieving “Six Sigma”. From these original documents of Bill, Ken, and Mikel, a clear picture is obtained of how “Six Sigma” was invented, defined, and rolled out to a larger organization.
The concepts from these Original sources have been used to document, for historical purposes, the origins and initial concepts of Six Sigma. In some situations, the original text and examples have been simplified, paraphrased, and grammar corrected. It is also to share with people the fact that the initial first thoughts that led to the creation of “Six Sigma” were a result of Bill Smith’s focus was on Early Life Product Reliability. This focus then led to an effort to reduce defects per unit which led to the use of defects per million opportunities to compare products of various complexities.
Efforts from this point on were to develop and communicate a six sigma framework to help utilize common measures and to reduce defects per unit by process and by product systematically with the goal of 10 times improvement over a period of time. Many quality improvement tools and methodologies already existed, but the framework of Six Sigma provided a system to quantify and compare internal and competitive products.
Examples of the initial applications of these initial Six Sigma methods are also included in section 2. To determine defect rates on critical characteristics where specification limits were well beyond a value of +/- 3 sigma, additional Z tables to determine areas under the normal curve had to be developed out to +/- 6 sigma using spreadsheet software. Almost all popular books on Statistical Process Control at the time typically had Z values only out to 3 or 3.5 sigma.
This blog entry was prepared by John Forsberg. His book is published at: