QFINANCE Topics
• Balance Sheets and Cash Flow
• Financial Markets
• Financial Regulation and Compliance
• Funding and Investment
• Governance and Ethics
• Mergers and Acquisitions
• Operations and Performance
QFINANCE Reference
• Research Data
• Financial Reference

Home > Business Strategy Calculations > Statistical Process Control Methods

Statistical Process Control Methods

What It Measures

Statistical Process Control (SPC) is a tool for monitoring and controlling variation in processes such as manufacturing of goods, testing, or statistical results. It was created in the 1920s by Dr W. Edwards Deming, who claimed the majority of variation was due to operator over-adjustment. SPC requires an organization to:

• determine the process parameters that need to be monitored;

• create a control chart to confirm the process is under control;

• collect data to compare with the control chart to identify process variation.

Why It Is Important

SPC itself doesn’t improve your processes—it can only tell you if variation has increased beyond normal levels. However, this information enables businesses to incorporate or eliminate the changes causing an abnormal variation. This is important because it provides an understanding of business baselines, gives insights into possible process improvements, and shows the value and results of existing processes. SPC can also provide real-time analysis to establish where business processes can be improved, improving decision-making.

How It Works in Practice

The key tool in applying SPC is the control chart that illustrates what a process looks like (statistically) when it is in-control. Control charts typically measure variable data and monitor the process target (mean result) and process range. There are a number of different sorts of control chart but among the most common are the X chart (pronounced “x bar” but also known as the averages or means chart), and the R chart (also known as the range chart).

This type of control chart has three key elements: a center line, an upper control limit, and a lower control limit. The center line on an X control chart is the process mean, while the center line on an R chart is the mean range.

The upper and lower control limits (UCL and LCL) are set to represent deviation from the mean that includes 99.7% of all data points (i.e., plus and minus 3 standard deviations from the mean). Data is then collected from the process and the mean plotted on the X and R charts—which can be interpreted to determine if the process is staying in-control or is out-of-control.

In the example below, a company is manufacturing pencils that are 16mm in diameter (mean), and they want to know if the process is in-control, and the level of variation in the pencils created.

The company begins by collecting a series of sample measurements, which are placed in subgroups. Next, the mean of each subgroup is calculated by adding all the measurements together and dividing by the number of measurements in the subgroup.

Next, the company calculates the mean of all of the means—this gives an overall mean for the data. This overall mean is the centre line in the control chart—in our example this is 13.75.

Next, the company must calculate S, the standard deviation of the data points. This can be easily calculated in Excel using the formula =STDEV (data points).

Next, calculate the UCL and LCL as follows:

UCL = CL + 3 × S

LCL = CL – 3 × S

where CL is the center line or overall mean value. Draw a line on the control chart to represent the UCL and LCL.

On your X chart, the x-axis shows the subgroup means from your original calculation. The y-axis represents actual measurements from the process over time. As a rule of thumb, the process is considered out-of-control if any point falls beyond the UCL or LCL lines (i.e., more than 3 standard deviations from the mean).

Control charts may also indicate a process is out-of-control if eight consecutive points fall on the same side of the center line, or if more than two consecutive points are more than one standard deviation point from the mean.

• Using a spreadsheet application, it is possible to calculate standard deviation and mean for a set of data. In Excel this is done using the STDEV and AVERAGE functions.

• The benefits of SPC won’t be immediately realized in every organization. SPC is best applied where there are clearly defined and consistent processes, and where the organization’s leadership is willing to identify and address new problems that might be identified by SPC.

Further reading on Statistical Process Control Methods

Books:

• Oakland, John. Statistical Process Control. 6th ed. Amsterdam: Elsevier Butterworth-Heinemann, 2008.
• Stapenhurst, Tim. Mastering Statistical Process Control: A Handbook for Performance Improvement Using SPC Cases. Amsterdam: Elsevier Butterworth-Heinemann, 2005.