Dorian Shainin - Red X and Pareto

Red X and Pareto

Shainin's development of the “Red X” concept originated from his association with Joseph Juran. In the 1940s Juran coined and popularized the notion of “the vital few and trivial many,” also known as “The Pareto Principle,” recognizing the uneven impact of problems on business performance to be the same phenomenon that Vilfredo Pareto had observed in respect to the distribution of wealth. As suggested by Juran, “I observed (as had many others before me) that quality defects are unequal in frequency, i.e., when a long list of defects was arranged in the order of frequency, a relative few of the defects accounted for the bulk of the defectiveness.”

In the 1950s Shainin recognized that the Pareto principle could be applied effectively to the solving of variation problems. Shainin concluded that, amongst the thousands of variables that could cause a change in the value of an output, one cause-effect relationship had to be stronger than the others. Shainin called this primary cause the “Big Red X” and demonstrated that the cause can exist as an interaction among independent variables. The effect of the Red X is then magnified by the square-root-of-the-sum-of-the-squares law, thereby isolating the root cause.

Shainin asserted that his application of statistical methods was more cost-effective and simpler than Taguchi methods. In order to determine the "Red X," Shainin would swap pairs of parts between functional and faulty equipment until the one part responsible for the failure is discovered. Shainin would claim that he could often find the primary defective part within a dozen paired swaps.

Shainin's policy of "talking to the parts" was the primary distinguishing factor that set his methods apart from Taguchi's. In classical or Taguchi DOE (Design of Experiments), engineers would brainstorm to form hypotheses regarding possible causes of a problem. Shainin's methods postpone this theoretical step, requiring first the diagnosis of causes via one or more of four clue generation techniques designed to determine, through the empirical testing of the actual parts in question, the root cause, or "Red X".

In the 1940s Leonard Seder, an MIT classmate and friend, developed the Multi-vari chart, a graphical method for analysis of variance. Shainin was an early adopter of this method, discovering that with Multi-vari charts, he could quickly converge on the root cause of a problem. Multi-vari charts also played an influential role in Shainin’s development of the Red X concept.