Process Capability - Part 1: Difference between Process Capability (Cp, Cpk) and Process Performance (Pp, Ppk)

One of the most common questions that comes up during my discussion with manufacturing and quality managers has to do with the difference between Process Capability Analysis (Cp, Cpk) and Process Performance Analysis (Pp, Ppk), and how these techniques relate to Statistical Process Control.

While there is a lot written about the difference in calculations between these indices, very little has been written about the difference in application i.e. where each of these techniques is used, and what manufacturing problems they help us solve. Therefore, over the next few blog posts, I will explain how to make the best use of each of these techniques in a manufacturing environment.

(Note that these blogs are written from the manufacturing practitioner’s perspective, and I may refer you to other more detailed explanations for some of the underlying statistics.)


Difference between Cp and Pp

[Potential] Process Capability Analysis (Cp)

A process capability study uses data from a sample to PREDICT the ability of a manufacturing process to produce parts conforming to specifications.

This prediction enables us to “qualify” a new manufacturing process as being fit for use in production. Therefore, Process Capability Analysis is best used during New Product Introduction (NPI), or at the start of each day or each shift of production as a method of checking whether a process is in control and capable of producing parts that meet specifications.

Note that when we talk about process capability, we tend to omit an important word: “potential”. The index Cp provides a measure of potential process capability i.e. how well a process can perform if there is no change in the underlying process conditions.

However, we all know that in the real world, underlying process conditions are constantly changing – tools wear, operators change, machines heat up, maintenance is skipped, raw materials vary from lot-to-lot - and therefore, we need a mechanism to monitor, control, and correct our performance over time. That’s where Statistical Process Control and specifically Control Charts come in.

Statistical Process Control and Control Charts

SPC is a tool for monitoring and controlling a production process to prevent the production of non-conforming product. The key term here is “control”. Unlike Process Capability and Process Performance Analyses where we measure and analyze without changing settings until after the study is complete, SPC requires that we measure, analyze, and act on the analysis real time. SPC helps us detect instability in the process due to things like tool wear, a drop in clamp pressure etc, and therefore requires that we make the necessary adjustments to ensure that the process is returned to stability and continues to produce parts within specifications.

[Actual] Process Performance Analysis (Pp)

A process performance study is used to EVALUATE a manufacturing process and answers the question: “how did the process actually perform over a period of time?” In daily use we tend to omit an important word, "actual", from the name of this analysis.

The missing word “Actual” from Process Performance and the missing word "Potential" from Process Capability result in significant confusion among practitioners.

The Process Performance Analysis is best used when you have inspection data, and you want to use the inspection data to understand how your manufacturing process actually performed. This is a historical analysis rather than a predictive analysis, but can still be used to drive process improvements.

A Critical Difference

Note that one of the most important distinctions and a commonly made mistake occurs during data collection. Process Capability Analysis requires that data be captured in the production sequence i.e. in the order the parts were produced.

This is because Process Capability Analysis is predicated on the process being stable over time and we must be able to measure and compare process variation over time in order to determine stability. On the other hand, Process Performance Analysis does not require data in production sequence as it uses the total standard deviation and it does not matter what order the parts were produced in (more on this in a subsequent post).

Cp and Cpk require that the manufacturing process be stable, while Pp and Ppk do not require process stability. So Cp and Cpk are effectively the best the process can perform when it is stable, and Pp and Ppk is the actual performance against that theoretical best.