1. "There is no lean enterprise god but Henry Ford."
• Henry Ford's industries (his automotive plants plus the industries that grew to support them) were directly responsible for making the United States the wealthiest and most powerful nation on earth.
• "What has Six Sigma done for Motorola lately?" (This is not sour grapes, I have an ASQ Six Sigma Black Belt certification.) In contrast, Ford's system did not let him down during the post-World War I depression. It even held up pretty well during the Great Depression.
• Taiichi Ohno, the developer of the Toyota production system, says openly that his ideas came from Henry Ford (with American supermarkets, in which products are replaced as they are purchased, as the inspiration for kanban). Waste (muda) reduction was a centerpiece of Ford's methods.
• Ford's business system included every element of what we now know as lean enterprise, except for the statistical quality control methods that support it. (Their introduction began during the 1920s, with Walter Shewhart as the pioneer of control charting.)
• Continuous improvement (kaizen), in combination with standardization (making the new "one best way" the standard for the job) and best practice deployment (applying the improvement to every similar operation in the business) was a paramount aspect of everything that Ford did.
• The ability to identify waste (muda) that most people would overlook was among Ford's most important success secrets.
• Ford's processes included error-proofing (poka-yoke), and self-check systems such as those later described by Shigeo Shingo.
• Ford described design for manufacture (DFM) explicitly.
• Ford used the elements of 5S-CANDO, especially continuous cleaning and preventive maintenance.
• Supply chain management, including freight managment systems (FMSs) and supplier development, played major roles in Ford's operations. Just-in-time delivery was a specific goal because Ford realized that even materials in transit constituted inventory. The FMS could track a railroad car to within one hour, even without the aid of modern information systems.
• Ford also included the "soft sciences" that are a prerequisite for the successful implementation of lean manufacturing. Ford described the key elements of Dr. Stephen Covey's excellent Principle-Centered Leadership very explicitly.
• The entire workforce was apparently trained and empowered in lean manufacturing, as shown by numerous worker initiatives to suppress waste. The attitude at the River Rouge plant regarding waste was, "It worried the men." This little phrase shows that lean thinking was for everyone, not only technical support personnel.
2. "I am His Prophet."
•  Levinson, 2002, Henry Ford's Lean Vision: Enduring Principles from the First Ford Motor Plant (Productivity Press), is apparently the first comprehensive reconstruction of the Ford lean enterprise system. Searches in Amazon.com and Abebooks.com (The Advanced Book Exchange, a network of used book dealers) do not unearth any similar references. Both sources were used to purchase the references that went into Henry Ford's Lean Vision!
• This book is actually a comprehensive survey of every available book about the Ford production system. Most of them are, incidentally, out of print.
• Arnold and Faurote's Ford Methods and the Ford Shops (1915)
• The notorious Harry Bennett's Ford: We Never Called Him Henry (1951) shows how Ford recognized the waste in a rust-flecked pile of slag near the Detroit, Toledo, & Ironton Railroad.
• Henry Ford's own My Life and Work (1922), Today and Tomorrow (1926), and Moving Forward (1930)
• Benjamin Franklin, whose writings influenced Ford, described some elemental principles of lean manufacturing: (1) lost time is lost forever (a cardinal principle of Goldratt's theory of constraints) and (2) buying inventory simply to get a good price is a bad idea.
• Edwin Norwood's Ford: Men and Methods (1931), the story of the River Rouge plant. It underscored the fact that waste reduction was part of the company culture, and it also showed that Ford was using lockout-tagout for safety during maintenance.
• Charles Sorensen's My Forty Years with Ford (1956). Sorensen, the designer of the Willow Run bomber factory, was Ford's production chief.
• Frederick Winslow Taylor's The Principles of Scientific Management (1911) and his less-well-known Shop Management. (1911). A table of metalworking activities in Shop Management clearly separates non-value-adding setup from value-adding machining-- and Shigeo Shingo was influenced by Taylor's writings.
• When compared with modern sources with which I am also familiar, the Ford system came together as a complete picture. It is this complete picture that I can offer to clients, in the form of a system of synergistic and mutually supporting elements.

See my page on this subject for more information. Few if any SPC textbooks show what to do when the data don't follow a normal distribution. Furthermore, the central limit theorem (which says that the averages of large samples follow a normal distribution no matter what the underlying distribution looks like) works for SPC only if large samples are indeed available. Furthermore, it doesn't help at all for process capability, which measures the process' ability to meet specifications. Specialties include:

1. SPC and process capability indices for nonnormal distributions, such as the gamma and Weibull distributions
• A one-sided specification is evidence that your process may be nonnormal.
• Impurity levels often follow the gamma distribution.
• I've seen plenty of statistical quality reports in which nonnormal data was plugged into conventional SPC software, which assumes a normal distribution. Many points are outside the control limit, simply because the control limit is wrong. On the other hand, the process capability (Cp, Cpk) estimates from such data are almost always overoptimistic. The distribution's long tail usually extends toward the specification limit, so the estimate of the nonconforming fraction (e.g. ~1 ppb for six standard deviations between the nominal and the spec limit) can be off by several orders of magnitude!
2. SPC and process capability indices for processes with nested levels of variation
• This is typical of batch processes, because there is within-batch and between-batch variation. Another strong argument for single-unit processing (where feasible)!
3. Confidence limits for capability indices
• Process capability indices that are based on small samples have very wide confidence limits. The confidence interval is fairly wide even when 30 to 100 measurements are available.
• If your customers are asking for monthly Cpk reports whose basis is a few measurements, I can educate them.

Some related publications

Levinson, W. A. "Statistical Process Control in Microelectronics Manufacturing," Semiconductor International, November 1994.

Levinson, W.A. "Exact Confidence Limits for Process Capabilities" Quality Engineering, 9(3), 521-528, 1997 (March)

Levinson, W. "Watch Out for Nonnormal Distributions of Impurities," Chemical Engineering Progress, May 1997, pp. 70-76.

Levinson, W. "Approximate Confidence Limits for Cpk and Confidence Limits for Non-Normal Process Capabilities," in Quality Engineering, 9(4), 635-640 (1997)

Levinson, William A. "Using SPC in Batch Processes." Quality Digest, March 1998, 45-48

Levinson, William and Polny, Angela. "SPC for Tool Particle Counts," Semiconductor International, June 1999.

Levinson, "SPC for Real-World Processes: What to do when the Normality Assumption Doesn't Work." Presented at the ASQ's Annual Quality Conference (2000)

Levinson, W.A., Stensney, Frank, Webb, Raymond, and Glahn, Ronald. 2001. "SPC for Particle Counts," Semiconductor International, 10/01