Re: 6 Sigma

[Rip Stauffer <ripstaur@vabch.com>]

Joseph Kelada wrote:

"The 'sigma' scale alows each department or even person to
measure its, his or her performance 'as objectively' as possible and
compare it). The known equation is 'performance' = (number of
errors/defects or whatever, found) / (maximum number of possible errors)
multiplied by 1,000,000. You compare this with the sigma scale and see
how you rate in number of sigmas. The 6-sigma is equal to 3.4 per
million (allowing for a 1.5 sigmas shift of the mean, either way).
Whatever the level you are at gives you an indication at a later time to
measure your improvement."

And if this idea is still the foundation of Six Sigma, it is still
fundamentally flawed, still a throwback to pre-Taguchi MBO, still a
refutation of Shewhart and Demings' work, still completely ignorant of
analytic studies. These numbers only work on the mathematical plane, and
the only way to achieve them is through 100 percent inspection and a
great deal of luck (or fudging figures). 

>From Grant, Myron and others, I had thought Six Sigma had evolved to
something that was worthwhile. Maybe it has in some organizations. But
if the above paragraph is still the fundamental basis for it in others,
it is still fundamentally flawed. I attended a Motorola presentation 4
years ago which propounded the above. I thought the approach was just
"Crosby meets Stats 101" then--I see no reason to change my opinion now.

Rip Stauffer
Naval Leader Training Unit
ripstaur@vabch.com
===================================================================