Statistical Process Control

Statistical Process Control, or SPCClosed Statistical Process Control, or SPC, is an industry-standard methodology for measuring and controlling quality during the manufacturing process., is a statistical method used to improve and control processes over time. It involves running a statistical analysis on a process to control the generated output. With the process analyzed, irregular processes can be brought in control, or made stable. A stable process can then be improved upon to generate less errors over time and/or an output with a smaller degree of error.

The processes that can be improved with SPC include manufacturing processes and outputs, financial and accounting processes, shipping and inventory management processes.

Statistical Analysis

Statistical analysis follows a simple procedure:

  1. Run a process.

  2. Measure the output.

  3. Graph the results.

This creates SPC Charts that can tell you at a glance if a process is stable.

A complete statistical analysis also requires collecting and understanding the following information:

  • Variation

  • Standard Deviation

  • and Process Capability Index


Variation is what causes the output to vary and can be considered as one of two groups: 

  • Common causes - Random variation in the process output that is expected. This is often from a non-assignable source.

    For example: tool wear, part size, die rolls, time to get to work

  • Special causes - Intermittent and unstable variation in the process output that is unexpected. This is often from an assignable source.

    For example: tool malfunction, incorrect measurements, loaded dice, random acts of nature (like a hurricane or an earthquake)

By tracking the measured output on a chart, you can identify variation present due to special causes and then work to eliminate it. Removing special causes from the process means the process is Stable, or In Control.

Common Causes Special Causes

Example deviation graph for a stable process:

Example deviation graph for an unstable process:

Note: Stable, or In Control processes will display a Standard Deviation graph with a bell-shaped curve.

Standard Deviation

Standard deviation, measured in sigma (σ), is used to measure the output variation spread for a process. The current industry standard for standard deviation of a process is 3σ, which is a stable process that has a ~99.8% of occurring.

Your deviation value determines your upper control limit and lower control limit:

  • UCL - Upper Control Limit: how high the measured output value can go before the process is unstable
  • LCL - Lower Control Limit: how low the measured output value can go before the process is unstable

Process Capability Index

Process Capability Index, or CpkClosed The Interchange Control Header, or ISA, is the first segment in an X12 EDI data transaction. Just like the addresses located on a piece of mail, the ISA segment contains sender and receiver information., is how well a process produces an output that is within the specification limits. Cpk can only be used on stable processes.

Cpk measures how much "natural variation" a process experiences relative to its specification limits and allows different processes to be compared with respect to how well an organization controls them.

Control Charts

By logging your output measurements, you can create a control chart, which allows you to analyze the process output and see what the deviation is from the intended value. Control charts are a useful tool for avoiding defects and preventing unnecessary process adjustments while providing easy to use and understand diagnostic and capability information:

MR Control Chart Example

Standard Deviation Graph Example

Control Charts help you monitors an important characteristic of a process over time. This examines the process to consistently and predictably make the process achieve higher quality, lower unit cost, and a higher effective capacity. The chart does this by answering the following:

  • Is the process predictable and stable over time, i.e. in control?

  • If the process is in control, can it meet the requirements?

  • Were there changes to the process made by your team that had an impact on the process?

The three elements of a control chart are:

  • A time series graph - added to display the measurement change over time.

  • A central line (X) - added as visual reference for detecting shifts or trends. This is also called the process location.

  • Upper and Lower Control Limits - computed from available data to easily view when a process is unstable.

Control Chart Prerequisites

Before building a control chart, you should:

  • Identify and define the process

  • Determine characteristics to be charted, considering:

    • The customer’s needs

    • Current and potential problem areas

  • Define/ establish the measurement system

  • Minimize unnecessary variations, i.e. Special Causes

Implementing Control Charts

Once you have your prerequisites mapped, follow the below process to implement the control chart:

  1. As the day progresses, in real time, gather data and plot it on a chart.

  2. Calculate the limits based on the process data.

  3. Identify any special causes and take corrective action.

  4. Recalculate limits.

  5. Once the process is stable, calculate the capability.

  6. Make process adjustments that lead to improved quality, cost, efficiency, etc.

There are multiple types of SPC charts. OnRamp focuses on the Individual Moving Range or X-MR chart.

Once you see an unstable process, investigate as to what occurred to have the output out of the limits so that you can know what changes were made that altered the process and you can make the required changes to make the process stable again and improve the process.

Recalculating Control Limits

When the following occurs, OnRamp recommends recalculating your upper and lower control limits:

  • There is a known cause to the process that has caused a deviation to the previously desired output, such as a new hire, or a machine recalibration.

  • There has been a statistically significant change to the mean or standard deviation.

  • There is an ongoing change to the process performance that is expected to continue.

Note: Automated control limit recalibration features may hide, or not take into account, important process changes that would require manually reviewing the limits.

SPC Reporting

In OnRamp, the Quality > Statistical Process Control folder contains all the SPC related items, including Review screens and SPC Reports.

You can also use SPC to track and plot preventative maintenance tasks. For more information, see