Quality By the Numbers
---“In God we trust, all others bring data.”
I’ll admit it, I’m a numbers person. Numbers are our friend.
They help us make better decisions. Here is a set of 7 numbers that have helped me in my 47 year career.
1.5 for a Cpk minimum
In six sigma a Cpk of 1.5 equals 3.4 ppm. This can be concerned zero defectis. If you are under 1.5 do the following sequence:
1. Remove assignable (special cause) variation
2. Get the process on target (not costly)
3. Reduce sigma (more costly)
4. Question the specification (if possible).
3 minimum for cross-functional decision making
Supplement: 3 for cross-generational brainstorming
Most organizations use a concept of a Material Review Board (MRB). A typical membership is Production, Engineering, and Quality. The purpose is to come up with an agreed upon decision from 3 different functions. Each has a different perspective. Makes use of a “checks and balances” system.
Most organizations have 3 different generations working at the same time. Tap into that diversity. Realize your generation is outnumbered by the other two.
10% of your time on improvement
Many of us spend our entire day on control or “fire fighting”. We should be spending close to an hour a day on improvement or “fire prevention”. The fire department does both those functions.
20% are the vital few
Compliment: 80% are the trivial (useful) many
Any list can be prioritized by those having the most impact on the results. For instance, on any given process 20% of the varibles give you 80% of the overall process variation. This provides where you should put your effort. It is called “the biggest bang for your buck”.
Trying to work all the trivial many can spread yourself to thin and is not as cost effective.
6-12 for ideal team size
Less than 6 could limited variety of ideas. More than 12 could cause not everyone contributing. This applies to project teams as well as training sessions. I heard of one person having 12 apostles -- and that was already one too many (grin)!
21 days to develop a habit
Corollary: 21 days to break a habit
Joe Juran told me I need to develop the “habit of improvement”. Research shows it takes 21 days to develop a habit. It is about reprogramming brain cells. This would be the same time to break a bad habit. Now you know why New Year’s resolutions very seldom work!
30 data points
From any population, all you need is 30 data points. For those readers who know about test of hypothesis, the t-distribution becomes very close to the z-distribution at n=30. The key is the 30 data points have to be completely random and represent an entire homogenous population.
If a large population is from a continuous process, it is better to use rational subgroups. A control chart would be the preferred tool of choice. How many subgroups do you need to calculate control limits? You guessed it – 30.
I will leave you with one final tip. Enhance any number by using the $ sign.
This is especially helpful when talking with upper management.
Dr. Juran called it the language of upper management.
Good luck!
Bio:
Ron Sedlock has close to 50 years of quality experience. He began his career studying under the personal tutorage of Dr. W. Edwards Deming and Dr. Joseph M. Juran.
He has been a member of ASQ or ASQC since 1976 and has held most ASQ Certifications.
He is a military veteran serving with the 1st Air Cavalry in Vietnam.
He can be reached at rsedlock@msn.com