The W. Edwards Deming Institute Blog

Statistical Techniques Allow Management to do a Better Job


In this post I discuss another wonderful paper by Dr. Deming. The W. Edwards Deming Institute makes this paper, and many more, available on our website. As you would expect from a non-profit focused on promoting the application of his ideas these articles are freely available with no barriers to downloading them. On the Teaching of Statistical Principles and Techniques to People in Industry by W. Edwards Deming, 1954:

For all four groups of people [management, statistical administration, research, front-line workers], the statistical method is more than an array of techniques. It is a mode of thought-sharpened thinking. It helps anyone in the four groups, be he a machine operator or an executive, to make better decisions, and to do his work better, than he could do otherwise.

Once again this paper shows the insight W. Edwards Deming had as he aimed to improve the functioning of the entire organization over 60 years ago. Those who continue to think he was focused on the factory floor alone have missed most of what he proposed. Improving the decision making at the executive level was always Deming’s focus. Continual improvement should be a part of everyone’s job but as executives have more authority the impact of improving their performance multiplies, or stifles, the impact of improvement anywhere else in the organization.

Management is today the most important and most neglected of the four groups, hence I will be more explicit with respect to it than the others.

with the aid of statistical techniques, much of what the subjectivity of the information that management needs will disappear, and there will emerge the possibility of the science of management. Statistical techniques do not displace management: they aid management to do a better job.

Even most of those interested in Deming’s ideas have not read this paper: I strongly recommend reading it, it is packed with valuable information and provides insight into what he was thinking in 1954. And it clearly shows the primary importance he placed on changing how executives and managers thought and acted, even all the way back in 1954. I’ll admit I get a bit tired of hearing over and over people say that Deming’s ideas only make sense on the factory floor, or only applied to manufacturing companies, and such statements. It is very hard to have such a limited view with even a cursory look at what Deming himself said and wrote.

He discusses the red bead experiment in the paper. He also touches on a topic dear to my heart, design of experiments (DoE) that Dr. Deming didn’t write about much, saying in this paper:

greatly enhance the usefulness of the experimental program… by the use of efficient designs (factorial designs, confounding, fractional representation, randomized blocks, etc.).

My father co-authored Statistics for Experimenters and worked a great deal with DoE, he also worked with Dr. Deming, including authoring the pages on the project applying Deming’s ideas in City of Madison government in Out of the Crisis (Peter Scholtes was introduction to Deming’s ideas during this project).

The paper also talks about consumer research and how that ties directly into the management of the company.

Consumer research takes the pulse of the consumer’s reactions and demands, and interprets the results for the use of management. The main use of consumer research is to feed consumer reactions back into the design of the product, so that management can anticipate rationally changing demands and requirement

Deming thoughts on consumer research are rarely discussed. I think this is because to some extent these ideas have become part of the accepted way of operating. But even in this area the full richness of Deming’s ideas are still missed today. I think the accepted ideal matches what Deming discussed fairly well today, but the practice often falls far short of the ideal. Often consumer research is more directed by marketing than by operations and the organization often is more focused on using what they learn to market than to improve products and services (though obviously it is used to improve also).

The importance of both customers and non-customers is also understood in theory but in practice often customers are surveyed but not those people that chose not to be customers. This is a huge gap in how consumer research should be done.

And most importantly gaining intimate insight into customers so that you can predict and innovate was key to Deming’s way of thinking but is often far short in businesses today. To the extent they study the potential customer base they often do not do so in a focused enough way. Of course, in some instances business do. And in fact those doing this are often given glowing press on their design efforts (see organizations like IDEO – our previous blog post: Deming’s Stage 0: By What Method?).

"Does the customer invent new product or service? The customer generates nothing. No customer asked for electric lights..."

Another topic he addresses is the idea that an organization can install statistical thinking without changing how executives and managers think and act. How often over the decades have you heard this same comment about why quality effort failed or a lean effort failed? Thinking that adopting a few tools is what is needed completely misses the point. The point is to change how management thinks and behaves and then have that way of thinking permeate the organization. That isn’t something that is bolted onto the current culture. The current culture must change.

the statistical control of quality is not installed. One sometimes hears of a company that is about to “install” the statistical control of quality, as if it were about to install a new air-conditioning system, or new linoleum, a new filing system, or even a new president. Statistical principles and techniques must be rooted and nourished with patience, support, and recognition from top management.

Related: The aim of supervision should be to help people and machines and gadgets to do a better job101 Ways to Design an Experiment, or Some Ideas About Teaching Design of ExperimentsExecutive LeadershipQuality is Made in the Board Room

Categorised as: data, Dr. Deming, understanding variation


  1. Anonymous says:

    Thanks for the great recommendation. I’ll definitely bring this up in our next strategy meeting.

  2. […] Hunter of Curious Cat suggests that Statistical Techniques Allow Management to do a Better Job: “Thinking that adopting a few tools is what is needed completely misses the point. The point […]

  3. […] discussed Deming and DoE a bit in a previous post: Statistical Techniques Allow Management to do a Better Job (about a paper Dr. Deming […]

Leave a Reply

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