Understanding Variation
Understanding variation is fundamental to effective process management. Based on the pioneering work of Dr. Walter Shewhart and refined by Dr. Donald J. Wheeler, we can learn to distinguish between different types of variation and make better decisions about when to act and when to leave a process alone.
Why Every Process Has Variation
Variation is not a defect or failure—it's a fundamental characteristic of all real-world processes. No matter how carefully controlled, every process will produce slightly different results each time it runs. This is because:
- No two inputs are ever identical: Materials, information, and conditions vary slightly each time, even when they appear the same
- Human involvement introduces variability: People naturally vary in their actions, decisions, and interpretations, even when following the same procedures
- Environmental factors constantly change: Temperature, humidity, time of day, and countless other factors affect process outcomes
- Measurement itself has variation: Even our tools for measuring outcomes have inherent imprecision
- Complex interactions create unpredictability: Multiple factors interact in ways that produce natural fluctuations in results
The goal is not to eliminate variation—that's impossible. Instead, the goal is to understand it, distinguish between its types, and reduce it where appropriate while avoiding tampering with stable processes.
Two Types of Variation
Dr. Wheeler's work emphasizes that while all processes contain variation, not all variation is the same. There are two fundamental types that require different responses:
Common Cause Variation
This is the natural, inherent variation that exists in every process. It's predictable, stable, and forms the "voice of the process." Common cause variation is always present and represents the routine, everyday fluctuations in your system.
- Predictable within statistical limits
- Consistent over time
- Requires system-level changes to reduce
- Cannot be eliminated by adjusting individual instances
Special Cause Variation
This is variation that comes from outside the normal process. It's unpredictable, unstable, and represents exceptional circumstances or events that are not part of the routine operation.
- Unpredictable and sporadic
- Caused by specific, identifiable events
- Can and should be investigated and addressed
- Signals that something unusual has occurred
The Danger of Tampering
One of Dr. Wheeler's most important insights is the concept of "tampering" - making adjustments to a process in response to common cause variation as if it were special cause variation. This actually increases variation and makes the process less predictable.
When you react to every fluctuation in your process metrics, you're likely treating common cause variation as if it were special cause. This leads to:
- Increased overall variation
- Reduced process stability
- Wasted effort on non-issues
- Decreased team morale and trust in the process
Using Control Charts
Control charts, as developed by Walter Shewhart and refined by Dr. Wheeler, provide a statistical method to distinguish between common and special cause variation. By plotting your process data over time with calculated control limits, you can:
- Identify when special causes are present
- Avoid tampering with stable processes
- Make data-driven decisions about when to investigate
- Track the impact of process improvements
Practical Application
In workflow management, understanding variation helps you:
- Distinguish between normal fluctuations in cycle time and genuine problems
- Avoid overreacting to single data points
- Focus improvement efforts on systemic issues rather than random events
- Build more predictable and stable delivery processes
For a deeper understanding of these concepts, we highly recommend Dr. Wheeler's books, particularly "Understanding Variation: The Key to Managing Chaos" and his numerous papers on statistical process control.