Your delivery times fluctuate by a day or two each week. Your machine produces parts that are never exactly identical. Further, your call center handles different volumes hour to hour. You train your team to follow the same procedure, and still the output varies slightly each time.
None of this is necessarily a problem. This is common cause variation — the natural, predictable noise that exists in every process, in every industry, without exception.
Understanding common cause variation is one of the most foundational skills in Lean Six Sigma. Not because it is complicated, but because confusing it with the other type of variation — special cause variation — leads to one of the most costly and common mistakes in process management: reacting to normal variation as if it were a crisis.
This article explains what common cause variation is, where it comes from, how to recognize it, how it differs from special cause variation, what tampering is and why it makes things worse, and how to genuinely reduce it when the process demands improvement.
Table of contents
- What is Common Cause Variation?
- How to Recognize Common Cause Variation on a Control Chart
- Common Cause vs. Special Cause Variation: The Critical Distinction
- Tampering: What Happens When You React to Common Cause Variation
- Common Cause Variation and Process Capability
- How to Reduce Common Cause Variation
- Common Cause Variation in the DMAIC Framework
What is Common Cause Variation?
Common cause variation is the inherent, random fluctuation present in any process that is operating in a stable, predictable state. It results from the combined effect of many small, everyday factors that are always present — minor differences in raw materials, slight environmental changes, small inconsistencies in equipment performance, natural variation in how people apply a process, and dozens of other routine influences that no one can fully control or eliminate.
The term was introduced by Walter A. Shewhart, a physicist and statistician at Bell Laboratories, who formalized the distinction between common cause and special cause variation in a landmark internal memo on May 16, 1924. That memo introduced the control chart as a tool to distinguish one from the other.
Shewhart’s framework was later championed and extended by W. Edwards Deming, who made it a cornerstone of his management philosophy.
Common cause variation goes by several names across the quality and statistics literature:
- Chance cause variation (Shewhart’s original terminology)
- Natural variation
- Random variation
- Noise
- Non-assignable cause variation
- Inherent process variation
They all describe the same thing: background fluctuation that is the predictable result of how the process normally operates, not a signal that something has gone wrong.
Public, Onsite, Virtual, and Online Six Sigma Certification Training!
- We are accredited by the IASSC.
- Live Public Training at 52 Sites.
- Live Virtual Training.
- Onsite Training (at your organization).
- Interactive Online (self-paced) training,
What Causes It?
No single source causes common cause variation. It is the cumulative result of many small influences acting simultaneously — the “system” in which the process operates.
Typical sources include:
Minor variation in incoming raw materials or components — not defective, just slightly inconsistent batch to batch. Small changes in ambient temperature, humidity, or other environmental conditions that affect the process in minor ways. Gradual wear in equipment that has not yet crossed any threshold requiring maintenance. Slight differences in how individual operators apply a standardized method, even when trained to the same procedure. Normal measurement system variation — the small, unavoidable imprecision in any gauge or instrument. Random fluctuations in customer demand or process input volumes.
No one of these sources dominates. No single one can be “fixed” to eliminate the variation entirely. The variation is a property of the whole system, not of any one element within it.
This is why Deming repeatedly emphasized that most variation in organizational outcomes is produced by the system itself, not by individual workers. Trying to hold an individual accountable for output that varies within normal system noise is both unfair and ineffective. Improving the system — changing the process design, materials, equipment, or environment — is what actually reduces common cause variation.
How to Recognize Common Cause Variation on a Control Chart
The primary tool for distinguishing common cause variation from special cause variation is the control chart, developed by Shewhart for exactly this purpose.
A control chart plots process data over time with three reference lines:
- The centerline (the process average)
- The upper control limit (UCL) set at +3 standard deviations from the centerline
- The lower control limit (LCL) set at −3 standard deviations from the centerline
Control limits are calculated from the process data itself — they are not specification limits set by the customer. They describe the natural range of the process’s own variation.
A process showing only common cause variation displays all of the following:
- All data points fall within the upper and lower control limits
- No long runs of consecutive points on one side of the centerline (typically 8 or more in a row signals a non-random pattern)
- No consistent trends, cycles, or unusual patterns
- Points scattered randomly above and below the centerline
When these conditions are met, the process is said to be in statistical control, or stable. This does not mean the process is producing good output. It means the process is predictable. Its future behavior can be estimated from its historical data. A stable process producing poor output is a capability problem — not a stability problem — and requires a different response.
Common Cause vs. Special Cause Variation: The Critical Distinction
This distinction is the most important concept in statistical process control, and getting it wrong in either direction carries a real cost.
Common cause variation is expected, always present, and random within a stable range. It is produced by the system. Reducing it requires changing the system.
Special cause variation is unexpected, intermittent, and non-random. It is produced by a specific, identifiable factor that is not part of the normal process — a machine breakdown, a bad batch of material, an operator error, an equipment miscalibration, an unusual demand spike. It shows up as a point outside the control limits, a run of points trending in one direction, a suspicious cluster, or any other pattern that is statistically unlikely to occur by chance.
The practical implication is straightforward:
When a process shows special cause variation, find the cause, remove it, and prevent recurrence. The investigation should be immediate — the cause is usually temporary and identifiable.
When a process shows only common cause variation, do not react to individual data points. The variation is normal. Reacting to it as if something has gone wrong — adjusting a machine setting, retraining an individual, escalating a brief dip in performance — is called tampering, and it makes variation worse.
Also Read: Lean Six Sigma in Remote Work: Eliminating Waste and Reducing Variation
Tampering: What Happens When You React to Common Cause Variation
Tampering is the term Deming used for adjusting a stable process in response to individual data points that are simply normal variation. It is one of the most widespread and damaging mistakes in process management.
The mechanism is counterintuitive but consistent: when you adjust a process that is already in statistical control, you introduce a new source of variation on top of the existing common cause variation. The adjustments themselves become a source of additional noise. The process becomes less stable, not more.
Deming demonstrated this vividly through his Funnel Experiment — a physical demonstration using a funnel and a marble. When an operator adjusts the funnel’s position in response to each marble’s landing point (treating each deviation as a signal), the marbles scatter further and further from the target over time. When the operator leaves the funnel in a fixed position and ignores individual landing points (treating the variation as common cause noise), the marbles cluster tightly around the target.
The lesson is direct: trying to correct normal variation causes more variation.
In practice, tampering looks like:
A supervisor adjusting machine settings every time output drifts slightly from nominal — even though the drift is within control limits and within specification. A manager asking an employee to explain why their productivity was lower yesterday when yesterday’s performance was within the normal range of day-to-day fluctuation. A production team recalibrating a tool after every piece, when the tool’s variation is normal and expected.
All of these interventions feel productive. They are not. They add variation without addressing any real cause, because there is no assignable cause to address. The variation is a property of the system, and the system has not changed.
Common Cause Variation and Process Capability
A process can be in statistical control — showing only common cause variation — and still fail to meet customer requirements. These are two separate questions.
Stability asks: is the process predictable? Is its behavior consistent over time?
Capability asks: does the process produce output within the customer’s specification limits?
A process must be stable before capability can be meaningfully assessed. If a process is erratic, capability indices (Cp, Cpk) calculated from its data are misleading — the indices change from sample to sample because the process itself is changing. Stability is the prerequisite for capability analysis.
Once a process is confirmed stable, the capability analysis compares the natural spread of the common cause variation to the width of the specification range. If the common cause variation fits comfortably within the specification limits, the process is capable. If the natural variation is too wide for the spec range — even though the process is stable and in control — the process is not capable.
This is the situation that requires genuine process improvement: a stable but incapable process. Reducing common cause variation enough to fit within specifications requires changing the system — the materials, equipment, methods, or environment that produce the variation.
How to Reduce Common Cause Variation
Because common cause variation is a property of the system, reducing it requires changing the system. No amount of retraining, exhortation, or individual consequence will reduce it. The following approaches are most effective.
Reduce input variation. Work with suppliers to tighten incoming material specifications, or qualify materials more consistently before they enter the process. If raw material variation is a major driver of output variation, improving incoming quality is direct and often high-impact.
Improve equipment consistency. Preventive maintenance programs, equipment upgrades, and precision tooling all reduce the process variation that comes from equipment inconsistency. Measurement system analysis (MSA/Gage R&R) is also essential — before assuming the process is varying, confirm that the measurement system is not itself contributing significantly to the observed variation.
Standardize the work method. When operator-to-operator variation is a significant source of common cause noise, standardizing how the work is done — clear documented procedures, consistent training, physical fixtures and guides that reduce the opportunity for method variation — reduces this source systematically.
Redesign the process. Sometimes the only way to meaningfully reduce common cause variation is to fundamentally change how the process works. This may mean changing the technology, the process sequence, the materials used, or the environmental conditions under which the process runs. This is the territory of a full DMAIC improvement project.
Use designed experiments (DOE). When it is unclear which sources of variation are dominant, designed experiments allow teams to systematically test the effects of multiple factors and their interactions. This identifies which system changes will produce the greatest reduction in variation — and avoids wasting effort on changes that have little effect.
Common Cause Variation in the DMAIC Framework
Common cause variation appears throughout all five DMAIC phases, but it is most central in the Measure and Analyze phases.
Define: When defining the problem, teams distinguish between processes with chronic common cause issues (consistently incapable processes) versus those with intermittent special cause problems. The type of variation determines the type of project.
Measure: Establishing the baseline requires a stable process. Teams use control charts to confirm the process is in statistical control before calculating capability indices. If special causes are present, they must be identified and eliminated before a meaningful baseline can be set.
Analyze: When a process is stable but incapable, root cause analysis focuses on identifying the dominant sources of common cause variation — the system factors whose variation produces the most output spread. Multi-vari analysis, graphical tools, and hypothesis testing help isolate these sources.
Improve: Solutions target the root causes of common cause variation identified in the Analyze phase. Changes to materials, equipment, methods, or environment are tested, validated, and implemented.
Control: After reducing common cause variation, new control charts are established to monitor the improved process. The control plan documents the new process settings and the response rules for both common cause patterns (no reaction required) and special cause signals (investigate immediately).
Also Read: Six Sigma in SaaS: Reducing Defects, Churn, and Process Variation
The 85/15 Rule: Where Responsibility for Variation Lies
Deming observed across decades of consulting work that the large majority of quality problems are caused by system factors — the processes, materials, equipment, and conditions that management controls — rather than by individual worker behavior.
This observation is sometimes summarized as the 85/15 rule: approximately 85% of process problems are caused by the system (common cause variation, the responsibility of management to address), and only about 15% are caused by special causes that workers can influence directly.
The exact percentages vary by situation, and Deming himself cited different figures at different times. The underlying principle is consistent and well-supported: in most organizations, most variation is common cause variation embedded in the system. Improvement comes from changing the system, not from pressuring the people operating within it.
This framing has direct implications for how organizations manage performance. Holding individuals accountable for output variation that falls within normal system noise is not just ineffective — it damages trust, creates fear, and reduces the candid reporting of problems that is essential for genuine improvement.
Learn Common Cause Variation in Our Lean Six Sigma Training
Understanding the difference between common and special cause variation, and knowing how to respond correctly to each, is one of the most important practical skills in process improvement. Getting it right avoids wasted effort on tampering. Getting it wrong costs real money in unnecessary adjustments, misdirected investigations, and process changes that add variation instead of reducing it.
At Six Sigma Development Solutions Inc., common cause variation, control charts, and the full statistical process control toolkit are taught in all of our Lean Six Sigma certification programs:
- Onsite training at your facility, with examples drawn directly from your own processes and data
- Live virtual classroom with a live instructor, real-time Q&A, and structured project work
- Online self-paced certification you can complete on your own schedule
Our Yellow Belt and Green Belt programs both cover common and special cause variation, control chart interpretation, and the correct management response to each type. Our Black Belt program adds advanced variation analysis, measurement system analysis, and designed experiments for reducing common cause variation at its source.
