Control Charts: How to Know Whether Your Process Is in Control
A control chart (also known as a Shewhart chart or statistical process control chart) is a graphical tool for monitoring a process over time and distinguishing between two types of variation: common cause variation (the natural, expected variation inherent in any process, which does not require intervention) and special cause variation (variation that signals a genuine change in the process, which requires investigation and, if negative, correction).
The distinction matters because the appropriate response to common cause variation is different from the appropriate response to special cause variation. Reacting to common cause variation as if it were special cause -- 'tampering' with a process that is in control -- increases variation rather than reducing it. Control charts give you an objective basis for knowing which type of variation you are looking at.
The Anatomy of a Control Chart
Centre line (CL): the mean of the process data over the baseline period.
Upper control limit (UCL): the mean plus three standard deviations. Calculated from the process data -- not set externally.
Lower control limit (LCL): the mean minus three standard deviations.
Data points: individual measurements plotted over time, connected by a line.
Control limits are calculated from the data, not from specification limits or customer requirements. A process can be in statistical control but still not meet specifications.
Signals That a Process Is Out of Control
A process is considered out of control -- a special cause is present -- when any of the following patterns appear:
A single point falls outside the upper or lower control limits.
Two out of three consecutive points fall in the outer one-third of the chart (between two and three standard deviations from the mean).
Four out of five consecutive points fall on the same side of the centre line beyond one standard deviation.
Eight or more consecutive points fall on the same side of the centre line (a run).
A trend of six or more consecutive points steadily increasing or decreasing.
How to Use a Control Chart in Practice
Choose the right chart type. Different control chart types are appropriate for different data types. An Individuals and Moving Range (I-MR) chart is appropriate for individual measurements taken over time (e.g., cycle time for each transaction). An X-bar and R chart is appropriate for measurements taken in subgroups. A p chart is appropriate for proportion defective data.
Collect baseline data. A control chart requires at least 20-25 data points to calculate stable control limits. Collect baseline data before attempting to monitor the process.
Investigate signals -- do not tamper. When a signal appears, investigate its cause. If the cause is identified and is genuinely special (a machine malfunction, a staff change, a policy change), address it. If no cause is found, treat the point as common cause variation. Do not adjust the process in response to individual data points that fall within the control limits.
XNM applies Lean Six Sigma tools including statistical process control to quality improvement in public-sector and capital-project environments. Reach out to XNM's strategic advisory team to discuss process control and quality improvement for your organisation.