Every manufacturing line and business process has variation, that’s unavoidable. The real question is whether that variation stays within limits your customer can accept. That’s exactly what process capability measures: the statistical ability of a process to produce output that consistently meets specification limits. Without this metric, you’re essentially guessing whether your process is reliable or quietly generating defects.
Process capability analysis uses indices like Cp and Cpk to put a number on how well your process performs relative to what’s required. Cp tells you about potential, what the process could do under ideal conditions. Cpk tells you the truth, what it’s actually doing right now, including any drift from center. Together, they give you a clear, data-driven picture that removes opinion from the conversation.
At Lean Six Sigma Experts, capability analysis is foundational to how we help organizations identify and reduce process variation. Whether we’re consulting on a production line, training Green Belts, or staffing a continuous improvement team, Cp and Cpk are among the first metrics we examine. This article breaks down what process capability means, how to calculate and interpret the key indices, and how to run a capability study from start to finish.
Why process capability matters
Understanding what is process capability goes beyond a technical exercise. When you measure capability, you find out whether your process is structurally able to meet customer requirements before defects ever reach the customer. Without that knowledge, you rely on inspection to catch problems after they occur, which is expensive, slow, and inherently reactive.
The gap between passing inspection and being capable
Inspection and capability are not the same thing. A process can pass final inspection on a given day and still have a Cpk below 1.0, meaning it regularly produces output outside specification limits. Inspection catches defects after they exist. Capability analysis tells you why they are being generated in the first place.
That distinction matters because fixing a root cause is far cheaper than filtering defects at the end of a production line or service cycle. Once you know your Cp and Cpk values, you know whether to adjust the process mean, reduce variation, or both, and you can prioritize that work with data instead of instinct.
A process with a Cpk of 1.33 or higher produces fewer than 64 defects per million opportunities, which is a standard benchmark many industries use to define a capable process.
The business cost of poor capability
Poor process capability drives up costs in ways that don’t always appear on a single report. Scrap, rework, warranty claims, and customer returns are direct costs you can measure. Lost contracts and damaged supplier ratings are harder to quantify but equally damaging to the bottom line.
When you run capability studies on a consistent schedule, you build documented evidence that your process is stable and repeatable, which is exactly what quality auditors and customers want to see. Organizations that treat capability as an ongoing metric rather than a one-time project consistently spend less on correction and more on improvement work that drives real growth.
The core metrics: Cp, Cpk, Pp, and Ppk
Four indices do most of the analytical work when you measure what is process capability: Cp, Cpk, Pp, and Ppk. Each one answers a different question about your process, and using them together gives you a complete picture of performance from both a short-term and long-term perspective.
Cp and Cpk: potential vs. actual performance
Cp compares your specification width to six times the within-subgroup standard deviation. It tells you how much room your process has to operate within the specification limits. A Cp above 1.33 means the process spread is comfortable, but Cp ignores centering entirely.
Cpk corrects for that blind spot. It measures the smaller distance from your process mean to either specification limit and divides by three times the within-subgroup standard deviation. When Cpk equals Cp, your process is centered. When Cpk runs lower than Cp, the mean has shifted toward one limit and you need to act.
A Cpk of 1.0 produces roughly 2,700 defects per million opportunities. Raising it to 1.33 drops that number below 64 DPMO.
Pp and Ppk: long-term process performance
Pp and Ppk use the overall standard deviation calculated from all study data rather than within-subgroup variation, which makes them reflect long-term process behavior across shifts, operators, and material lots.
Your Ppk will often run lower than your Cpk. When that gap is large, special-cause variation is entering your process over time, and the size of that gap tells you exactly how urgent the investigation needs to be.
What to confirm before you calculate Cp and Cpk
Before you run any numbers, two conditions must hold or your Cp and Cpk values will mislead you. Understanding what is process capability requires recognizing that these indices are built on specific statistical assumptions, and skipping the validation step produces results that look mathematically precise but point you in the wrong direction entirely.
Your data must be normally distributed
Cp and Cpk formulas assume a normal distribution, so if your data is heavily skewed or bimodal, the indices will understate or overstate your actual defect rate. Run a normality test such as the Anderson-Darling test before you calculate anything.
If your data fails normality, you have three practical options:
- Transform the data using a Box-Cox or Johnson transformation
- Use a non-normal capability analysis matched to the correct distribution
- Investigate whether mixed process streams are distorting the shape of your data
Your process must be in statistical control
Statistical control means only common-cause variation is present, with no unexplained spikes, shifts, or trends visible in your output. Confirm this by building a control chart and reviewing it before you run the capability study, not after.
Calculating Cpk on an unstable process produces a number that describes unpredictable behavior, not true capability, and any corrective action you base on it will fail to hold.
If your control chart shows out-of-control signals, stop and investigate those causes first. A stable process is the only valid foundation for a meaningful capability index.
How to run a process capability study step by step
Running a process capability study gives you a structured path from raw data to a clear verdict on whether your process meets requirements. The steps below apply whether you’re working through what is process capability on a production line or a transactional service process.
Step 1: Define your specification limits and collect data
Start by confirming your upper specification limit (USL) and lower specification limit (LSL) with your engineering or customer requirements documents, not from memory. Then collect a minimum of 25 subgroups with four to five measurements each, gathered during normal operating conditions across multiple shifts.
Collecting data during only your best production runs will inflate your Cp and Cpk, giving you false confidence in a process that underperforms regularly.
Step 2: Verify stability and normality, then calculate your indices
Before you calculate a single index, build a control chart to confirm your process is in statistical control. If it is, run a normality test to validate your data distribution. Once both checks pass, calculate Cp and Cpk using your within-subgroup standard deviation, then calculate Pp and Ppk using the overall standard deviation to compare short-term potential against long-term reality.
Review the gap between Cpk and Ppk. A large gap signals time-dependent variation from sources like shift changes, tool wear, or raw material differences that your team needs to identify and address.
How to interpret results and choose the right action
Once you calculate your indices, you need to translate those numbers into a specific course of action. Understanding what is process capability only pays off when the results tell you what to fix, and the Cpk value is your primary decision driver.
When Cpk is below 1.0
A Cpk below 1.0 means your process is actively producing defects beyond your specification limits right now. Your first move is to determine whether the problem is poor centering, excessive variation, or both. If your Cp is acceptable but Cpk is low, your spread is fine but your mean has drifted, and you can often recover quickly by adjusting the process target.
If both Cp and Cpk are below 1.0, you have a variation problem that requires a structured reduction effort, not just a simple adjustment.
When Cpk is between 1.0 and 1.33
A Cpk in this range means your process meets minimum capability standards but has little room for error. Any shift in the mean or increase in variation will push output outside the specification limits. Treat this zone as a warning signal that calls for monitoring and incremental improvement work rather than acceptance.
Reaching a Cpk of 1.33 or higher puts your process in a solid operating range. Shift your focus to maintaining stability and documenting the controls that keep it there long term.
Key takeaways and next step
What is process capability in practical terms? It’s a measurement system that tells you whether your process can reliably meet customer requirements, not just on a good day, but consistently. Cp and Cpk are your primary tools: Cp measures potential, Cpk measures actual performance including any shift from center. Pp and Ppk extend that view across time and reveal whether long-term variation is eating into your margins.
Before you calculate anything, confirm your process is in statistical control and your data follows a normal distribution. Those two checks protect you from making decisions on numbers that don’t reflect reality. Once your indices are valid, let them guide your action: adjust centering, reduce variation, or both, depending on what the data shows.
If you want expert help applying these methods to your operation, contact the Lean Six Sigma Experts team to discuss your specific process improvement goals.