Quality Remix: Six Problems I have with Six Sigma, Part II
I visited Border’s bookstore to buy a book about Six Sigma. There were several to choose from, but as near as I could tell, none defined Sigma correctly. Some of the ways they defined Six Sigma included, “meeting the requirements of the marketplace,” “statistical measurement of process performance,” “goal for near perfection,” “system of management,” and “how much variability is within a group of items.” The last one is close, but no cigar. I didn’t buy a book. On yeah, one book described Six Sigma as TQM on steroids. That’s cute; it may be the best I saw.
I see Six Sigma as an out-of-the-box Juran-type quality control program with a problem solving technique and a superb marketing program.
Another thing that bugs me is the use of Six Sigma to talk about attributes. Sigma is based on variable data: measurement like inches, millimeters, ounces and kilograms. Those measurements produce variable data, so you can calculate Sigma. But Six Sigma practitioners apply the same Six Sigma performance standard to attributes.
What is the standard deviation of a dent or of a bunch of dents? Attributes are characteristics that are either there or not there. You can count them, you can chart them, and you can even calculate control limits for them--but there is no variable data, and, thus, no standard deviation.
To read Part I of Six Problems I have with Six Sigma, Click Here.