Why results differ.
WHAT VARIATION IS
The natural spread in any repeated process.
WHY IT MATTERS
Reacting to normal variation as though it were a problem makes things worse.
WHAT THE TWO TYPES ARE
Variation inherent in the process Variation from a specific cause
WHY THE DISTINCTION IS THE KEY IDEA
They require opposite responses.
WHAT TO DO ABOUT INHERENT VARIATION
Change the process, if the spread is unacceptable.
WHAT TO DO ABOUT A SPECIFIC CAUSE
Find and address that cause.
WHAT HAPPENS WHEN YOU CONFUSE THEM
Adjusting a stable process in response to normal fluctuation increases variation.
WHAT THAT LOOKS LIKE IN PRACTICE
A manager reacting to every weekly figure.
WHAT IT PRODUCES
Constant change, worse results and demoralised staff.
HOW TO TELL THE DIFFERENCE
Plot the measure over time, and look at the pattern rather than individual points.
WHAT TO LOOK FOR
Points far outside the usual range Runs consistently above or below the average Trends
WHAT A STABLE PROCESS LOOKS LIKE
Variation within a consistent range, with no pattern.
WHAT TO DO WITH A STABLE BUT INADEQUATE PROCESS
Redesign it.
WHY NOT EXHORT PEOPLE
The result is the process behaving normally.
WHAT TO DO WITH AN UNSTABLE PROCESS
Find the specific causes first.
WHAT TO AVOID
Explaining every fluctuation.
WHY
Most have no specific cause.
