6 Sigma Only for Giants? Deming's Thoughts?



6 Sigma Only for Giants?

About this six sigma matter, as I recall, Deming did not have a very positive attitude in his last book.Why? Does it work? Works only for giants like Motorola?


Juran is not positive about the six sigma process either.
My question is, why stop with six sigma? Why not go on with 9, 12, 15,18 . . . ?
We could go on forever, the tails of the curve will never go to "0".

Michael T

Hi Sotiris,

As I understand Dr. Deming's less than supportive comments concerning Six Sigma (this is just my interpretation/opinion), it all stems from the philosophy of conforming to specifications. Conformance to specifications does not take into consideration the loss associated with variation. While a process can be within specification, the variation around X-bar may be widely dispersed. Even with special causes of variation eliminated (the process is considered "in control"), the goal should be to reduce the common causes of variation to obtain a tight grouping around x-bar, not maintain conformance to specifications. By adding a Taguchi Loss Function, it is possible to observe that, the further a process moves from optimal, the greater the loss to the customer. That should be the goal - reducing the loss to the customer, not conformance to a specification. As Dr. Deming states, "The most important use of a loss function is to help us to change from a world of specifications (meet specifications) to continual reduction of variation about the target through improvement of processes." (The New Economics, Chapter 10).

Now, my thoughts on Six Sigma (taking Dr. Deming's System of Profound Knowledge and the philosophy of continuous improvement into consideration) are that Six Sigma focus's too much on the short term benefits derived from a Six Sigma project (and the financial gains derived from it). It lacks the long term focus necessary to fundamentally change the company culture to one of, say, Kaizen. Six Sigma seems to be looking for those big gains, overlooking the small incremental changes to a process that can reduce common causes of variation and tighten the process to get ever closer to the target value. Another problem I see with Six Sigma is that the focus is on large dollar projects. For example, at the onset of a Six Sigma program, projects will only be accepted if the expected payback exceeds a certain dollar figure, say $750,000. Eventually, those projects will be come more scarce as the number of "big ticket" projects diminishes. Then what? Lower the expected payback value? Ok... sure - drop the payback to $500,000. Eventually, those projects will also dry up. If the target payback keeps getting smaller, eventually the cost of a implementing a project will exceed it's payback. Yet, in all this hoopla about Six Sigma, where are those small changes to a process that the person on the shop floor can tell you about that can make a process a little better, get the product manufactured a little easier or a little quicker. There aren't any huge $$ gains, just the small incremental increases that lead to enhanced productivity, reduced variation and more contented line workers who feel that management is listening to them and they have control over their process.

Now - I am not saying Six Sigma is a bad thing. It is a step in the right direction. Training people in process improvement, SPC, DOE, etc., cannot be a bad thing. Like ISO or QS, it is a good start towards a process of continuous improvement, and that (continuous improvement) should be the ultimate goal of management.

So... does Six Sigma work. The short answer is, yes, Six Sigma works. Does it only work for giants? Don't know... Motorola is huge and has gained a great deal of notoriety because of Six Sigma. I don't know of any small companies that have implemented Six Sigma because you just don't hear of them.

Anyway - these are my thoughts on the subject. I'm donning my asbestos underware for the flames I may catch for this post... :D



Al Dyer


I like your style of thinking. Maybe after 6 sigma we can have 8 sigma, including all the new books, training, and associated costs.

Where will the madness end?

Top Bottom