6 Sigma and Cpk are both used to evaluate process performance, variation, and the risk of producing nonconforming parts. Although they are closely related, they do not describe exactly the same thing. Cpk is primarily a short-term process capability index, while 6 Sigma is both a statistical performance target and a broader process-improvement methodology.
What Is 6 Sigma?
Statistical Meaning
In statistics, sigma (σ) represents the standard deviation and is used to describe the amount of variation in a data set. A smaller standard deviation indicates that the process output is more tightly concentrated around its mean, while a larger standard deviation indicates greater variation.
A 6 Sigma process is one in which the process output is located at least six standard deviations from the nearest specification limit, represented by the upper specification limit (USL) or lower specification limit (LSL). When the process mean is centered between the specification limits and does not shift, the theoretical defect rate at the 6 Sigma level is approximately 0.002 ppm, or two defects per billion opportunities.
Management Philosophy and Improvement Method
6 Sigma is not limited to a numerical index. It is also a process-improvement philosophy and management methodology promoted by companies such as Motorola and General Electric. Its objective is to reduce process variation, control critical sources of defects, and improve consistency through structured problem-solving methods.
One commonly used framework is DMAIC:
- Define: Define the problem, customer requirements, process scope, and improvement objectives.
- Measure: Collect reliable process data and establish the current performance baseline.
- Analyze: Identify the causes of variation and determine which factors contribute to defects.
- Improve: Modify the process to reduce variation and improve its capability.
- Control: Maintain the improved condition through monitoring and process controls.
Under the conventional long-term 6 Sigma interpretation, the target defect rate is 3.4 ppm. This value considers the commonly used assumption that the process mean may drift by approximately 1.5σ over the long term. Therefore, the short-term centered statistical model and the long-term operational target should not be confused.
What Is Cpk?
Cpk, or the process capability index, measures whether a process can meet its specification limits while also considering whether the process mean is offset from the specification center.
Unlike an index that only evaluates overall process spread, Cpk examines the distance from the process mean to both specification limits and uses the smaller distance. In general form:
Cpk = min[(USL - μ) / 3σ, (μ - LSL) / 3σ]
Here, μ is the process mean, σ is the process standard deviation, USL is the upper specification limit, and LSL is the lower specification limit. Because Cpk uses the nearer specification limit, it decreases when the process mean shifts toward either limit, even if the overall variation remains unchanged.
Typical Cpk Interpretation
- Cpk ≥ 1.33: The process capability is generally considered adequate for the stated requirement.
- Cpk = 1.00: The process is marginal and may require improvement.
- Cpk < 1.00: The process capability is insufficient, with a significant risk of producing nonconforming output.
These thresholds are commonly used engineering guidelines, but their application still depends on the product requirement, measurement system, sampling method, and process-control conditions.
Cpk is normally used to describe short-term process capability under the assumption that the process is stable and that the measured data approximately follow a normal distribution. If the process is unstable, or if the data are strongly non-normal, a conventional Cpk value may not represent actual long-term performance accurately. In such cases, Ppk or an appropriate nonparametric method may be more suitable.
Relationship Between 6 Sigma and Cpk
6 Sigma and Cpk are different views of the same basic normal-distribution model. Both relate process variation to the distance between the process mean and the specification limits. The direct connection is the Sigma level, often expressed as a Z value.
Short-Term Relationship Without Process Shift
When the process mean is centered between the specification limits, the distance from the mean to the nearest specification limit is the same on both sides. Under this condition, the short-term Sigma level is related to Cpk by:
Zshort-term = 3Cpk
Accordingly, a process with Cpk = 1.00 has a short-term capability of approximately 3σ, while a process with Cpk = 1.33 corresponds to approximately 4σ of short-term capability. A Cpk of 2.00 corresponds to a short-term distance of approximately 6σ, provided that the process is stable, centered, and evaluated using valid data.
However, this relationship does not automatically mean that Cpk = 2.00 proves long-term 6 Sigma performance. The calculation describes the capability represented by the data set and the assumed short-term variation. It does not, by itself, prove that the process will remain centered and stable during extended production.
Long-Term Drift and the 1.5σ Assumption
Real processes may experience changes in equipment condition, material properties, operating temperature, setup, tooling, measurement conditions, or other process inputs. These changes can cause the process mean to shift over time. The conventional 6 Sigma methodology accounts for this practical issue by using an assumed long-term mean shift of approximately 1.5σ.
Under this convention, a short-term 6σ capability corresponds to a long-term effective distance of approximately 4.5σ after the assumed shift. The associated long-term defect target is approximately 3.4 ppm. This explains why the theoretical centered value of about 0.002 ppm and the commonly cited 3.4 ppm target are both associated with 6 Sigma, but refer to different assumptions.
Comparing the Two Measures
| Measure | Primary Focus | Typical Assumption |
|---|---|---|
| Cpk | Short-term process capability, including process centering | Process is stable and data are approximately normal |
| 6 Sigma level | Long-term defect-rate target and process-improvement objective | Long-term drift is considered through the conventional 1.5σ shift |
Cpk is therefore a snapshot of current short-term capability. It focuses on whether the present process output is likely to remain within the specification limits. The 6 Sigma level is a broader performance objective that includes a long-term view of defects, process variation, and potential mean drift.
Key Points to Consider
- Cpk assumes that the process is stable and that the data distribution is suitable for the calculation. If these conditions are not satisfied, the result may be misleading.
- 6 Sigma is generally treated as a long-term performance objective that includes the possibility of process mean drift, whereas Cpk is commonly reported as a short-term capability index.
- Cpk ≥ 2.00 should not be mechanically interpreted as proof of long-term 6 Sigma performance. Long-term stability, centering, sampling conditions, and the validity of the statistical model must also be confirmed.
- 6 Sigma is more than a statistical index. It also includes a structured improvement methodology, project management, process-control practices, and an organizational approach to reducing defects.
- A high capability index obtained from an unstable process may not translate into equivalent production performance. Stability and trend analysis should be considered before drawing conclusions from Cpk.
In practical process evaluation, Cpk can be used to assess the current relationship between process variation and specification limits. The 6 Sigma framework can then be used to reduce the underlying sources of variation and control the process over the long term. Used together, the two concepts provide both a measurement of present capability and a structured approach for achieving more consistent performance.
