- slide 1 of 2
Data Analysis Tool
The Control chart is popularly in use as a data analysis tool in Six Sigma DMAIC projects. The most basic type of control chart, the individuals chart, plots data over time. The chart includes statistical limits that represent common cause variation, which is variation in process performance due to random factors inherent in the process itself. Any points lying outside these control lines are due to special cause and signal a need for investigation.
Project leaders also use control charts to look for other types of special cause, such as trends over time and process shifts. If data shows no signs of special cause, the process is said to be stable or "in control". Note, however, that it is possible to have a process that is in control yet does not mean customer or business specifications.
Use control charts to uncover special causes in your data that provide information about opportunities for process improvement. Also use them to understand the amount of variation inherent in your process, which if too large may indicate the need for process improvement as well.
For additional details on the creation and use of control charts, read our Introduction to Control Charts. To learn more about the different control chart versions, check out this article on Types of Control Charts.
- slide 2 of 2
Control Chart Example
The image shown here is a control chart that you can download from this link. It charts data that might be collected over an approximate two-year period representing monthly web page views. Someone collecting such data might be interested in whether page views are increasing, and whether any specific months had higher or lower traffic than is typical.
The green line is the average of the data. Some people prefer to show the median, whereas others prefer the mean. The red lines are the statistical control limits, which represent the amount of variation you could expect based solely on common cause variation, which is random variation in the process. That is, even if nothing changes in terms of how the website is run or how traffic is generated, monthly hits could fall anywhere between the lower and upper control limits.
If a project manager determines that the amount of common cause variation is unacceptable, the next step would be to determine what process features lead to this variation and implement improvements to tighten up the process.
The blue line represents the lower specification limit or minimum target performance. Remember that spec limits indicate the customer requirements, whereas control limits indicate actual process performance. In this case, the site owner has determined that he or she wants to achieve at least 1,800 page views a month. (Note that this chart does not show an upper specification limit, but a site owner could conceivably set an upper limit based on the bandwidth the site server can handle.)
According to the control chart, although the process is in control it is not meeting the site owner's requirements. So a project manager would need to figure out ways of increasing the monthly page view volume in order to meet the business requirement.
This chart was created using Minitab software, and the red data points indicate evidence of special cause. The red point near the top of the chart is labeled "1" meaning it indicates the type of special cause that Minitab tracks as its first rule. This first rule checks for outliers, which are data points outside the specification limits. In this month, the number of page views is more than we would expect from random variation alone, so they site owner and project manager would want to investigate to find out what led to this higher level of traffic. Ideally the information gleaned from this investigation would allow them to increase traffic overall.
The two red points on the left side of the graph signal evidence of special cause matching Minitab's third rule. This rule tests for at least six data points in a row either increasing or decreasing. Minitab has found that seven points in a row progressively increased, indicating a trend in the data. A prudent Six Sigma practitioner knows that the obvious next step is to determine the cause for this trend and how to leverage it.
Your data might show evidence of other types of special cause, and software programs typically let you select which types of special cause to test for.