MANAGING FOR QUALITY PROCESS IMPROVEMENT

1
MANAGING FOR QUALITYPROCESS IMPROVEMENT

• DR. YONATAN RESHEF
• UNIVERSITY OF ALBERTA
• SCHOOL OF BUSINESS
• EDMONTON, ALBERTA
• CANADA T6G 2R6

2
STABLE SYSTEM/PROCESS

• A PROCESS WILL BE IN STATISTICAL CONTROL WHEN,
  THROUGH THE USE OF PAST EXPERIENCE, WE CAN
  PREDICT, AT LEAST WITHIN LIMITS, HOW THE PROCESS
  WILL BEHAVE IN THE FUTURE

3
PROCESS IMPROVEMENT

• IMPROVEMENT OF A STABLE PROCESS CANNOT BE DONE BY
  TAMPERING WITH OUTPUT (E.G., MANAGING BY RESULTS)
• ACTION BASED ON RESULTS CAN ONLY BE APPROPRIATE
  IN THE PRESENCE OF SPECIAL CAUSES

4
CAUSES OF VARIATION

• SPECIAL CAUSES (SIGNAL) PROBLEMS ATTRIBUTABLE
  TO INDIVIDUALS WHO ARE OUT OF STATISTICAL CONTROL
• COMMON CAUSES (NOISE) PROBLEMS ATTRIBUTABLE TO
  THE SYSTEM (I.E., MANAGEMENT)

5
VARIATIONTWO COMMON MISTAKES

• OVER-ADJUSTMENT ASCRIBING VARIATION OR A
  MISTAKE TO A SPECIAL CAUSE WHEN IN FACT THE CAUSE
  BELONGS TO THE SYSTEM
• DOING NOTHING ASCRIBING VARIATION OR A MISTAKE
  TO THE SYSTEM WHEN IN FACT THE CAUSE IS SPECIAL

6
TAMPERING WITH A SYSTEM

• TAKING ACTION ON A STABLE PROCESS IN RESPONSE TO
  PRODUCTION OF A FAULTY ITEM OR A MISTAKE
  (OVER-ADJUSTMENT)

7
INSPECTION, OR NO INSPECTION

• IF PROCESSES ARE IN STATISTICAL CONTROL, THERE
  ARE ONLY TWO CHOICES NO INSPECTION OR 100
  INSPECTION
• IF PROCESSES ARE IN CONTROL, A SAMPLE FROM A
  BATCH CONTAINS NO NEW INFORMATION CONCERNING THE
  UNINSPECTED ITEMS IN THAT BATCH
• THE CHOICE BETWEEN THE TWO ALTERNATIVES WHETHER
  TO INSPECT OR NOT IS MADE ON THE BASIS OF
  ECONOMICS, SAFETY, ETC.

8
CHAOS

• A STATE OF CHAOS, THAT IS WHEN PROCESSES ARE
  OUT OF CONTROL, DESERVES CONSIDERATION OF 100
  INSPECTION

9
LESSONS FROM THE RED BEAD EXPERIMENT

• THE PROCESS TURNED OUT TO BE STABLE THE
  VARIATION AND OUTPUT WERE PREDICTABLE
• ALL THE VARIATION CAME ENTIRELY FROM THE PROCESS
  ITSELF. THERE WAS NO EVIDENCE THAT ANY WORKER WAS
  BETTER THAN ANOTHER

10
LESSONS

• THE WORKERS COULD DO NO BETTER. BEST PEOPLE
  DOING THEIR BEST DOES NOT ALWAYS WIN THE DAY
• UNDER SUCH CIRCUMSTANCES, RANKING IS WRONG, AS IT
  ACTUALLY MERELY RANKS THE EFFECT OF THE PROCESS
  ON PEOPLE

11
LESSONS

• PAY FOR PERFORMANCE CAN BE FUTILE. THE
  PERFORMANCE OF THE WORKERS WAS GOVERNED BY THE
  PROCESS
• DIVIDED RESPONSIBILITY THE INSPECTORS WERE
  INDEPENDENT OF EACH OTHER (A POSITIVE PRACTICE).

12
LESSONS

• KNOWLEDGE ABOUT THE PROPORTION OF RED BEADS IN
  THE INCOMING MATERIAL (20) WOULD NOT ENABLE
  ANYONE TO PREDICT THE PROPORTION OF THE RED BEADS
  IN THE OUTPUT. THE WORKLOADS WERE NOT RANDOM
  DRAWINGS. THEY WERE EXAMPLE OF MECHANICAL SAMPLING

13
SAMPLING

• EVERY BEAD MUST HAVE A CHANCE TO BE IN THE SAMPLE
• IN OTHER WORDS, RANDOM SAMPLING MUST BE
  INDEPENDENT OF ANY PHYSICAL ATTRIBUTION OF THE
  EXPERIMENT
• COLOR OF THE BEADS
• SHAPE OF THE PADDLE
• ANGLE OF THE RAISING OF THE PADDLE
• SIZE OF THE SAMPLING BOWL

14
LESSONS

• Acceptable Defects Rather than waste efforts on
  zero-defect goals, Dr. Deming stressed the
  importance of establishing a level of variation,
  or anomalies, acceptable to the recipient (or
  customer) in the next phase of a process.
  Oftentimes, some defects are quite acceptable,
  and efforts to remove all defects would be an
  excessive waste of time and money.

15
LESSONS

• THERE WAS NO BASIS FOR MANAGEMENTS SUPPOSITION
  THAT THE 1-2 BEST WORKERS OF THE PAST WOULD BE
  BEST IN THE FUTURE
• RIGID/PRECISE PROCEDURES ARE NOT SUFFICIENT TO
  PRODUCE QUALITY
• NUMERICAL GOALS CAN BE MEANINGLESS