Two types of variation can be distinguished using control charts: the variations inherent in a process, such as product design or choice of materials, and special causes of variation, i.e., variations from unexpected sources, such as human error or a power failure. When the process is in control, it is solely governed by common causes and is indicated through data points lying within control limits. In this case, the only way to improve a process is to make a fundamental change, for example, upgrade equipment, change procedures, or provide more training. However, the presence of a special cause would result in data points falling outside the control limit or show trend. Special causes of variation must be removed before the control chart can be used to monitor the performance of a process to make it stable and predictable.