The most common sources of process variation fall into five well-established categories: people, methods, materials, equipment, and environment. These are the root causes that quality professionals investigate first when a process produces inconsistent results. W. Edwards Deming estimated that roughly 85% of process problems trace back to the system itself, not to individual workers, meaning variation is usually built into how work is designed, supplied, and maintained.
Two Types of Variation
Not all variation is the same. Quality improvement distinguishes between common-cause variation and special-cause variation, and knowing the difference determines how you respond.
Common-cause variation is the random, predictable fluctuation present in any stable process. It comes from the ordinary interaction of all the factors involved: slight differences in materials, small swings in temperature, normal human inconsistency. You can reduce it, but you can’t eliminate it entirely without redesigning the process.
Special-cause variation is an unpredictable deviation from something that isn’t a normal part of the process. A machine breaking down, a contaminated batch of raw material, or a new untrained employee working without supervision would all qualify. These events stand out from the background noise and call for immediate investigation rather than system-level redesign.
People
Human beings are one of the most frequently cited sources of process variation. Differences in training, experience, fatigue, attention, and motivation all introduce inconsistency. Two operators following the same procedure will produce slightly different results, and a single operator will perform differently at the start of a shift compared to the end.
Error-provoking conditions in the workplace amplify this natural variability. Time pressure, understaffing, and inexperience don’t just make mistakes more likely; they make the range of outcomes wider and less predictable. Addressing human variation typically means improving training, simplifying tasks, and designing systems that make the correct action the easiest one, rather than simply blaming workers for errors.
Methods and Procedures
How work is performed, and how consistently those steps are followed, is a major driver of variation. When procedures are vague, overly complicated, or left unwritten, each person interprets the process differently. The EPA notes that a significant part of the variability between laboratories analyzing the same samples comes from slight differences in how analytical methods are actually performed, even when everyone cites the same reference procedure. Small adjustments allowed by a general method can meaningfully change final results.
Standard operating procedures exist specifically to minimize this kind of drift. Well-written procedures should be detailed enough that someone with limited experience can reproduce the process unsupervised. But in practice, field conditions vary, and workers often make judgment calls that introduce inconsistency. The gap between the documented method and what actually happens on the floor is one of the most common and most overlooked sources of variation.
Materials and Supplies
Raw materials are rarely perfectly uniform. Batch-to-batch differences in physical and chemical properties can ripple through a process and change the final output. Research on pharmaceutical manufacturing found that variation in the processability of 131 full-scale batches of a single ingredient was driven by differences in the combined effects of particle size and packing behavior. Individually, each property appeared to fall within acceptable limits, but their interaction caused meaningful variation in how the material performed during production.
Supplier changes compound this problem. Different production sites, synthesis routes, and crystallization methods produce materials with different physical characteristics even when they meet the same chemical specifications. Small levels of impurities can alter surface properties in ways that standard chemical purity tests won’t catch. This is why incoming material inspection matters: chemical identity alone doesn’t guarantee consistent processing behavior.
Equipment
Machines degrade over time, and that degradation introduces variation. Cutting tools experience gradual wear during use, which progressively changes the characteristics of the finished product and can eventually lead to catastrophic failure. Calibration drift is equally insidious. A measuring instrument that was accurate six months ago may have shifted enough to produce systematically biased readings without anyone noticing.
Measurement systems themselves are a distinct source of variation worth understanding. When you measure something, the result reflects not just the actual part but also variation from the instrument, the person taking the measurement, and the measurement process itself. Formal measurement studies break this down into repeatability (does the same person get the same reading twice?) and reproducibility (do different people get the same reading?). If your measurement system is noisy, you may be reacting to phantom variation that doesn’t actually exist in the product.
Environment
Temperature, humidity, vibration, lighting, and noise all affect process outcomes. Research on environmental conditions found that temperature and noise interact in ways that influence both perception and cognitive performance. At a practical level, a process that runs well in a climate-controlled facility may behave differently in a shop floor exposed to seasonal temperature swings or in a room with significant vibration from nearby equipment.
Environmental variation is particularly tricky because it often changes slowly. A gradual seasonal temperature shift won’t trigger an obvious alarm, but it can systematically bias a process over weeks or months. Monitoring ambient conditions alongside process outputs helps separate environmental drift from other causes.
How Variation Is Detected
Control charts are the primary tool for distinguishing common-cause variation from special-cause variation. A control chart plots process measurements over time against statistically calculated upper and lower control limits. Points falling within those limits represent normal variation. Points falling outside them signal something unusual has entered the process.
Beyond individual outliers, several patterns within a control chart indicate special causes. Eight or more consecutive points on one side of the center line suggest the process has shifted. Six or more points trending consistently upward or downward indicate a systematic change. Two out of three consecutive points near the outer edge of the control limits warrant closer scrutiny. These rules give you a structured way to separate signal from noise rather than chasing every random fluctuation.
Run charts serve a similar purpose in early-stage improvement projects but are less sensitive to subtle changes. For ongoing monitoring where you need to catch small shifts before they become big problems, control charts are the stronger tool.
Why System Thinking Matters
The instinct when something goes wrong is to look for a person to hold responsible. Deming’s work pushed back hard against this. His 85/15 rule states that 85% of problems can only be fixed by changing the system, which is under management control. Only 15% are within the direct control of individual workers. Improving systems, not replacing people, holds the greatest potential for reducing variation and eliminating mistakes.
This reframing changes how you approach every source of variation. If an operator makes frequent errors, the productive question isn’t “why is this person careless?” but rather “what about the procedure, the tools, the training, or the work environment is making errors likely?” The answer almost always involves one or more of the categories above, acting together in ways that make consistent performance difficult even for skilled, motivated people.