The One-Page Memo That Changed Everything

The only words on a one-page memo were :

The attached form of report is designed to indicate whether or not the observed variations in the percent of defective apparatus of a given type are significant; that is, to indicate whether or not the product is satisfactory (1).

About a third of that memo was devoted to a simple hand-drawn diagram we now recognize as a control chart. That one-page memo altered the course of industry on May 16, 1924. Shewhart's superiors at Bell Laboratories, namely George D. Edwards, received this memo. In this memorandum, Shewhart introduces the first control chart to track process performance over time and identify parameter changes that may indicate quality deterioration.

Through carefully designed experiments, Dr. Shewhart developed the control chart and the concept of statistical control. Although Shewhart drew on pure mathematical statistical theories, he understood that physical processes typically produce "normal distribution curves" (also called "bell curves"). Manufacturing data did not always behave like nature's data (Brownian motion of particles). Dr. Shewhart concluded that while every process exhibits variation, some processes exhibit controlled variation that is inherent to the process. In contrast, others display uncontrolled variation not always present in the process causal system. Today, this is called Statistical Process Control (SPC.), a tool used for anything from toasters to nuclear power plants, the least of which is automobiles.

Bell Labs engineers realized the importance of reducing variation in manufacturing processes by 1920. Furthermore, they realized that continual process adjustment in response to non-conformance increased variation and degraded quality. Dr. Deming would later describe this as tampering. Shewhart framed the problem in terms of chance and assignable cause variation. Assignable cause variation is an uncontrolled variation that is not always present, while chance variation is considered a controlled variation inherent to the process. Deming would later call these common and special cause variations. As Shewhart emphasized, it is imperative to bring a production process into a state of statistical control, where there is only common-cause variation, and maintain it there to predict future output and manage a process economically.

The example diagram in the memo monitored a production process over some time using the percentage of defective products as an index. In the diagram, Shewhart introduced two tolerances above and below the arithmetic mean and suggested that values exceeding these limits were problematic. Shewhart's diagram indicates trouble against a point above the upper limit. It was not until 1924 that Shewhart published his first paper in the prestigious Bell System Technical Journal. The paper comprised 44 pages of closely argued statistical theory with few references to Western Electric products.

To this day, Shewhart's memo is remembered as a life-changing moment in his life and the entire industry.

1) Cultures of Control, Miriam R. Levin, Routledge, 2005 - Social Science — p156-174\