Process Variation

Check these links out:

http://www.isixsigma.com/library/content/c030120a.asp

-and-

http://www.isixsigma.com/library/content/c000806a.asp

-and-

http://www.isixsigma.com/offsite.asp?A=Fr&Url=http://www.qualitydigest.com/aug98/html/spctool.html

Do you have a target? The larger the variation the farther you will be, on average, from that target.

Examples:

I want this part to be 10mm long. If the manufacturing process deliver parts between 9 and 11mm, I am farther from the target (on average) than if the process delivers parts within 9.9 and 10.1mm.

I want to deliver pizza ASAP. If there "slowest" deliveries take 20 minutes longer than the "fastest" ones, even without telling me the "average delivery time" I can guess that it can not be better than 10 minutes. If the fastest ones take 2 minutes longer than the slowest ones, then the average delivery time can be pretty shorter. It is (hard but) possible that the average is still 20 minutes, but it is also possible that the average is 2 minutes. In the first case, an average 20 minutes is posible but 2 minutes is not. Note that even if the average was 20 minutes in both cases, the slowest deliveries would take 30 minutes in the first case and 22 minutes in the second case.

Think of errors per order, bugs per program, waiting time in a call center, stock in a wearhouse, dimensions in manufactured parts, delays in flights, etc. Wouldn't you want to reduce variation?

After his presentations, Deming used to say something like "If I had to say the whole thing in 2 words I would say: Reduce variation".

Just trying out..tell him what chaos his customer creates by introducing variation in his production plans...(if at all..)

You're applying for a new job.

Employees in that job position at Company A have an average salary of $38,500.

Employees in that job position at Company B have an average salary of $47,000

  If averages by themselves were sufficient for summarizing data then Company B would be first choice for a place of employment.  The salary structure (in thousands of dollars) for that position is as follows:

Company A:

25, 30,30,40,40,40,40,45,45,50   - standard deviation = 7.8

Company B:

5,10,10,20,20,20,30,30,100,225 - standard deviation = 68.1

The average gives you an estimate of the "typical" salary.  Knowledge of the variation, as measured by the standard deviation, gives you a sense of how far from "typical" your starting salary might be.  In short, by understanding the variation you are able to make intelligent decisions concerning the numeric differences between averages.  In the above example, an understanding of the variation in salaries strongly suggests that,as a new hire, you will probably make more by signing with the company whose average salary is less.