What is Process Capability?
A quick introduction to understanding Process Capability

Whenever we try to understand any process, we must consider two different “voices”

The Voice of the Process (VOP) This represents how much the process varies and whether this variation is within the expected limits or outside of them.

The Voice of the Customer (VOC) This represents how much the customer wants the process to vary and whether this variation is within the expected limits or outside of them.
Statistical Process Control (SPC) focuses on the first voice, while Process Capability focuses on the second.
So, how does Process Capability work? In simple terms, Process Capability uses statistics to assess how reliable a process is. The higher the values we obtain for capability metrics, the more reliable the process becomes.

Many companies today still operate with an “old school manufacturing” mentality. This means people check whatever is being produced to see if it falls within specifications. If it does, they might say, “Don't touch it; it ain't broke!” While this approach may work in the short term, statistics have shown that it is not sustainable in the long run.
Famous statistician Dr. Genichi Taguchi introduced a concept to explain this effect: the Taguchi Loss Function. According to Dr. Taguchi, the cost to a company is directly related to how close the process outcomes are to the specification limits. The more “centered” a process is, the closer the cost tends toward zero. However, as the process drifts closer to the upper or lower specification limits, the cost increases significantly.

A Practical Example

Imagine you work in a restaurant and promise to serve the best burger in under 10 minutes. If it takes 10 minutes or more, the customer gets it for free. Your last four preparation times are: 9:57, 9:59, 9:58, 9:58

Even though our process is centered around 10, there is a high probability over time we will fall outside, because we are close to an upper specification limit.
Now, consider the following times instead: 6:58, 6:55, 6:57, 6:59

If our process is centered and far from specs, the probabilities of it being out become extremely low. These times are much more consistent and significantly below the customer's expectation. This reliability provides a buffer, reducing the risk of missing the target.
In Process Capability, we aim to understand how well we meet customer expectations. It's not enough to stay within specifications; we must also ensure the process is reliable and consistent.
How Do We Measure Process Capability?
There are four key metrics to consider: Cp, Cpk, Pp, and Ppk.
Cp and Cpk typically indicate short-term capability.
Cp measures the potential of a process to meet specification limits, assuming the process is centered.
Cp = (USL - LSL) / 6σ
Where USL is the upper specification limit, LSL is the lower specification limit, and σ is the standard deviation. It does not take into account whether the process is centered or not, just how wide the process variability is in relation to the specification limits.
Cpk evaluates process capability while accounting for the means position relative to the limits.
Cpk = min((USL - μ) / 3σ, (μ - LSL) / 3σ)
Where μ is the mean of the process.
Pp and Ppk, often referred to as process performance metrics, typically indicate long-term capability. The formulas are similar to Cp and Cpk, with the key difference that they are using a different data for calculating the standard deviation. See the table below for more details.
Aspect | Process Capability (Cp, Cpk) | Process Performance (Pp, Ppk) |
---|---|---|
Calculation of σ | σ estimated from subgroup data | σ calculated from whole population data |
Purpose | For short-term capability. Often used when a process is stable and in control | For long-term capability. Often used for processes too new to determine if they are stable |
What is the best value for each? The answer is: It depends! These metrics are often negotiated with customers and vary depending on the industry and supplier-customer agreements. Some agreements might focus solely on Cpk, but for all metrics, the general rule is: the higher, the better. A common benchmark is for Cp, Cpk, Pp, and Ppk to be greater than 1.33. In industries with critical requirements, this threshold may be raised to values greater than 1.67.
I'm David, Lean Six Sigma Master Black Belt
With over 15 years of experience in various manufacturing industries, I've gained valuable knowledge and insights. I'm happy to help with any questions or comments you might have about this post or statistics in general—feel free to reach out!

I'm David, Lean Six Sigma Master Black Belt

With over 15 years of experience in various manufacturing industries, I've gained valuable knowledge and insights. I'm happy to help with any questions or comments you might have about this post or statistics in general—feel free to reach out!
Dominion
©2025 Dominion SPC, Inc.