Resources What's New Support Download Contact Sales: Arthur-Industrial@outlook.com
Home › Resources › A short history of SPC

A short history of SPC

From Shewhart's 1924 memo at Bell Labs to Deming, the Western Electric rules, Six Sigma and today.

Shewhart and the first control chart

On May 16, 1924, Walter A. Shewhart of Western Electric (later Bell Telephone Laboratories) wrote a short memo containing the first control chart. His insight was that variation has two sources: chance causes built into the system, and assignable causes that can be found and removed. His 1931 book, Economic Control of Quality of Manufactured Product, laid out the method.

Deming and Japan

W. Edwards Deming worked with Shewhart and spread his ideas. After World War II, Deming lectured to Japanese engineers and managers, starting in 1950 at the invitation of the Union of Japanese Scientists and Engineers (JUSE). Japanese industry adopted statistical methods widely; the Deming Prize was created in 1951. Deming's later work stressed management's responsibility for the system and warned against tampering with stable processes.

The Western Electric handbook

Western Electric's Statistical Quality Control Handbook (1956) published the zone rules for reading control charts that are still taught today. In 1984 Lloyd S. Nelson published a set of eight tests in the Journal of Quality Technology, now known as the Nelson rules.

Other milestones

  • 1954: E. S. Page introduced the CUSUM chart.
  • 1959: S. W. Roberts introduced the EWMA chart.
  • 1950s and 1960s: Kaoru Ishikawa popularized the cause and effect (fishbone) diagram and quality circles.
  • 1980s: Genichi Taguchi's ideas on robust design and the loss function reached the West.
  • 1986: Motorola started Six Sigma; General Electric adopted it on a large scale in 1995.
  • 1990: The Machine That Changed the World brought the term "lean production" to a wide audience.
  • Today: automotive (AIAG), aerospace (AS9100) and medical device (ISO 13485) quality systems expect statistical methods for key characteristics.

Questions and answers

Is SPC outdated now that we have sensors and machine learning?

No. SPC answers a basic question that more data does not remove: has this process changed, or is this just normal variation? Automated data collection makes SPC easier, not less relevant.

← All articles