Control charts are used for monitoring the outputs of a particular process, making them important for process improvement and system optimization. Although these statistical tools have widespread applications in service and manufacturing environments, they do come with some disadvantages.
Control charts are designed to measure variation in processes, including common cause variation and special cause variation. Common cause variation is considered normal, random variation within a process, while special cause variation is due to broken machinery or some other process defect. A control chart sometimes may indicate that a process is out of control and that there is special cause variation where none exists. These false alarms can cause unnecessary downtime and delays, which can cost a business money.
There are two main assumptions underlying control charts, which dictate the accuracy of the information provided to users. The first is that the measurement function monitoring a process parameter has a normal distribution. In reality, though, this may not be the case, meaning a control chart will fail to produce meaningful data. The second assumption is that measurements are independent of each other, which also may not be true. If both assumptions are in some way flawed, then control charts will fail to be useful.
Although control charts are not difficult to understand mathematically, they do require special training to create and use. Control charts use basic statistics, such as mean and standard deviations. Small organizations with limited training resources and limited experience with quality-assurance techniques will likely have difficulty implementing and using control charts. Businesses have to decide whether or not they can train their employees on lean and Six Sigma tools before using these quality tools to help improve their processes.
Misplaced Control Limits
Upper and lower control limits are added to control charts to help determine when a process is out of control. Control limits may be set too close or too far away from the process mean, distorting the information produced by control charts. If control limits are set too far away, then operators may be unaware that special cause variation is affecting the quality of process outputs. Similarly, limits that are too close to the mean may set off false alarms when a process is still in control.