Control Charts and Capability Analysis

What you don’t know can confuse you

May 28, 2019

Current business conversation often focuses on data and big data. Data are the raw information from which statistics are created and provide an interpretation and summary of data. Statistics make it possible to analyze real-world business problems and measure key performance indicators that enable us to set quantifiable goals. Control charts and capability analysis are key tools in these endeavors.

Control charts

Developed in the 1920s by Walter A. Shewhart, control charts are used to monitor industrial or business processes over time. Control charts are invaluable for determining if a process is in a state of control. But what does that mean?

“We’re really just talking about a process or system that’s stable and predictable,” explains Matt Savage vice president of product development for PQ Systems. “If it’s in a state of statistical control, it just means that the processes are expected in the future to do what [they have] been doing in the past. Control charts help us understand, statistically, whether a process is in fact stable or whether you’ve got some anomalies, or special causes, or out-of-control points. A control chart is just a simple run chart if you will, with a little bit of math or statistics that help give you clues when the process may have changed, or if in fact it did change, it’s going to give you some clues about that.”

The chart below is an actual chart of a restaurant’s weekly sales data:

The upper graph (X chart) displays data points (weekly sales averages) during a 25-week period. The calculated average is then used to calculate the upper and lower control limits. The lower graph displays the moving range (mR chart) with its average and upper control limit. There is no lower control limit as the value of the difference between consecutive observations is recorded as an absolute value (positive number).

“There are different ways you can use these data, but basically if you have a point above the upper limit or below the lower limit you would consider [it] out of control or worth investigating,” says Savage. “Most companies institute rules like: If a datum is seven consecutive points above or below the mean, that indicates that that process is statistically different, and therefore it’s worth your time to investigate and see what exactly happened.”

By plotting data and looking at where they fall, you’re able to get a feel for how your process is doing. If the data remain within the assigned limits for a period, you’re going to expect that pattern to continue. You can say with some degree of certainty this is what you expect to have happen going forward.

Capability analysis

Capability analysis is a set of calculations used to assess whether a system is able to meet a set of specifications or requirements.

“Capability analysis is nothing more than a tool to evaluate how well your process is doing rather than how stable it is,” says Savage. “You can think of it like your kids’ report card. They get letter grades A, B, C, etc. In the same way, capability analysis tells us how the processes are doing, if [a] process [is] capable of producing a part within spec, for instance.”

In process improvement efforts, the process capability index or process capability ratio is a statistical measure of process capability. In other words, the ability of a process to produce output within specification limits. Two common analysis indexes are centered capability ratio, Cpk, and centered performance ratio, Ppk.

The concept of process capability only holds meaning for processes that are in a state of statistical control.

“When you talk about capability analysis, one of the prerequisites is that your process is already stable,” explains Savage. “So, get your process stable and predictable first, and then assess whether it’s doing what it should be doing. If your control charts show your process is still all over the place, I would focus effort and energy more on getting things under control before analyzing its capability to deliver good results.”

Check out the full interview with Matt Savage on the
May 1, 2019, Quality Digest Live podcast.