The W. Edwards Deming Institute Blog

Specification-based Management is Not Sufficient


Guest post by Bill Bellows (originally written as a comment about the question of how Dr. Deming viewed six sigma).

While Dr. Deming was known for his appreciation of continuous improvement, he was also know for encouraging us to think of variation and improvement from a systems perspective. As such, I do not believe he would be a Black Belt.

Note that Dr. Deming ended his last book, The New Economics, with this quotation (first stated by Don Wheeler);

Conformance to requirements, Zero Defects, and Six Sigma [Quality], and all other specification-based nostrums, all miss the point.”

As to what point, I suggest his focus was the degree to which a specification-based focus, which is integral to Six Sigma Quality, Zero Defect quality, and even most lean efforts, which each carry on the world-wide 200+ year-old tradition of “interchangeable parts.” That is, the belief that all parts that meet requirements are good and equally good. Such models ignore variation within meeting requirements and implement variation reduction efforts to achieve zero defects as the end goal. Yet, are all newly degreed medical doctors the same?

As pointed out by John Betti, a Ford VP of Operations in the early 1980s, this assumption (all parts which meet requirements are good and equally good) has not been “rigorously and not so rigorously challenged.” Betti’s realization of the limits of “specification-based” thinking is based upon Ford’s 1983 discovery of the dramatic difference in warranty claims between an automatic transmission designed by Ford and built by both Ford and Mazda.

Ford’s design was produced using a classic “conformance to requirements” strategy, very much the world standard. As documented by Ford (and Henry Neave, in his book, The Deming Dimension), Mazda’s design followed the thinking of Genichi Taguchi’s Quality Loss Function, wherein they focussed their attention on the “gap” between the bore (hole) diameter and the outer diameter of the valve, realizing that variation in the ideal gap between them was essential to maintain. In so doing, Mazda managed variation as a system, to improve the function of the transmission. Ford managed variation of the parts, under the model of “conformance to requirements.” According to Neave’s book, Ford collected over 1,000,000 data points on 10 transmissions from Ford and Mazda, all chosen randomly, to realize that Mazda’s “mind the gap” focus was remarkably different (and better for customers) than Ford’s “mind the part” focus.

The last chapter of The New Economics explains Dr. Taguchi’s Loss Function concept and how this strategy challenges “conformance to requirements.” Dr. Deming promoted the practice of managing variation as a system, which does not “miss the point” of “conformance to requirements” he was addressing.

This video from Ford includes many details of this account. Henery Neave’s book adds a great deal more.

If anyone is interested, I have written extensively on the difference between “conformance to requirements” and “minding the gaps“; see a collection of articles available online.

Related: Lean Management Journal Articles by Bill BellowsUsing Deming’s Management Methods to Enhance the Application of Taguchi’s Ideas

Categorised as: continual improvement, systems thinking, understanding variation


  1. Tony says:

    Excellent. Great video. Pity all the Six Sigma mob don’t watch it.

  2. […] if the amateur orchestra does not make a mistake. Deming’s point was that even if the producer meets specifications it doesn’t guarantee a quality experience for the customer. The professional orchestra and the […]

Leave a Reply

Your email address will not be published. Required fields are marked *