Many teams use run charts and control charts interchangeably. They are not the same tool—and using the wrong one to analyze your process can lead to completely flawed operational decisions.
Using a run chart when you actually need statistical control limits will leave you chasing shadows, while throwing a control chart at a brand-new, unstable process will generate mathematical noise. In this guide, you will learn the exact differences between the two and see interactive visual examples of how they behave.
What Is a Run Chart?
A run chart is a simple line graph that plots data over time, with a center line representing the Median (the middle value of the dataset). It does not require complex statistics.
We use run charts to quickly detect shifts, trends, or non-random patterns in a process before we have enough data to calculate standard deviation. In my experience, a run chart is the perfect “first step” tool when you are trying to baseline a new production line or a freshly implemented Lean improvement.
Interactive Run Chart
Toggle the median line and shift detection. A “shift” occurs when 6 or more consecutive points fall on the same side of the median.
What Is a Control Chart?
A control chart takes the basic concept of a run chart and upgrades it with formal statistics. Instead of a median, it uses the Mean (average). More importantly, it introduces statistically calculated Upper and Lower Control Limits (UCL / LCL), usually set at 3 standard deviations (Sigma) from the mean.
Crucial Distinction: Control limits are not specification limits. Specification limits are what your customer wants. Control limits are what your process is actually capable of producing. A control chart tells you if the variation you are seeing is normal (Common Cause) or if something is fundamentally broken (Special Cause).
Interactive Control Chart
Toggle the statistically generated Control Limits. Any point falling outside these red lines is a “Special Cause” variation requiring immediate root cause analysis.
Side-by-Side Comparison
| Feature | Run Chart | Control Chart |
|---|---|---|
| Center Line | Median (Middle value) | Mean (Average value) |
| Control Limits | No | Yes (UCL & LCL) |
| Detects Trends | Yes | Yes |
| Detects Special Cause | Limited (Relies on run rules) | Strong (Statistical bounds) |
| Best Used For | Early stage, small datasets | Stable process monitoring |
When to Use a Run Chart
Use a run chart when you are at the very beginning of a Lean or Six Sigma project, or when you don’t have enough data points to calculate a reliable standard deviation (generally fewer than 20 points).
- To get quick visual insight on the shop floor.
- Communicating basic trends to operators.
- Before calculating formal process capability.
Real Plant Scenario
Tracking daily scrap rates on a newly installed assembly cell for the first 3 weeks. Because the process is brand new and highly variable, calculating formal control limits would result in massive, useless boundaries. A run chart easily shows if the scrap trend is improving over time.
When to Use a Control Chart
Use a control chart for ongoing process monitoring of an established, mature process. You need at least 20-30 data points for the math to be statistically valid.
- Validating that a process is stable over time.
- Holding the gains after implementing an improvement.
- Triggering a Root Cause Analysis when a point breaks the UCL/LCL.
Real Plant Scenario
Monitoring the extruded diameter of a rubber hose after a formal calibration adjustment. The process is stable, and we need a mathematical trigger (an outlier crossing the control limits) to tell the operator exactly when to halt the line and adjust the extruder.
Common Mistakes
Limits on Unstable Data
Applying statistically calculated Control Limits to a process that is wildly unstable or actively changing. Your limits will be so wide that they become meaningless.
Confusing Spec Limits with Control Limits
Drawing customer Specification Limits on a Control Chart. Control charts measure the “Voice of the Process”, not the “Voice of the Customer.”
Ignoring the Rules
Failing to recalculate Control Limits after a major, permanent process improvement (like buying a more precise machine) has been successfully implemented.
Over-complicating Early Data
Demanding a full control chart when an operator just needs a simple run chart to track their shift-by-shift hourly output.
Frequently Asked Questions
What is the main difference between a run chart and a control chart?
The primary difference is statistical rigor. A run chart simply plots data around a median to identify basic trends. A control chart plots data around a mean and includes mathematically calculated upper and lower control limits to detect special cause variation.
Can you add control limits to a run chart?
No. If you add statistically calculated limits to a chart, it mathematically becomes a control chart. Run charts specifically avoid control limits because they are intended for use when data is limited or non-normally distributed.
How many data points are needed for a control chart?
As a general rule of thumb in Six Sigma, you need a minimum of 20 to 25 data points to calculate a reliable mean and standard deviation for your control limits. Any fewer, and your limits will be inaccurate.
Is a control chart always better than a run chart?
Not always. A control chart is superior for monitoring a stable process, but a run chart is actually better for early-stage problem solving, tracking rapid improvements, or when you have very small sample sizes.
What is the difference between control limits and specification limits?
Control limits are calculated based on how your process actually performs (historical standard deviation). Specification limits are determined by what the customer wants (engineering tolerances). A process can be perfectly in control (within control limits) but still producing scrap (outside specification limits).
