EP241 – Cp and Cpk: The Small Difference That Makes a Big Difference
EP241 – Cp and Cpk: The Small Difference That Makes a Big Difference
The truth is, many people treat Cp and Cpk as if they’re the same, but they’re not. Mixing them up can cost you money, customers, and credibility. Simply put, Cp measures variation, while Cpk shows how well the process is centered and performing.
#AdvancedQualityPrograms #JuanNavarro #ProcessCapability #CPK
A few years ago, a production manager invited me to review his performance indicators. He was proud, and he had good reason to be. His team had worked hard. He showed me the screen and said, “Look, Juan. Cp is above 1.33 on all three lines. This is under control.”
I looked at him with respect, because the numbers were not lying. The Cp was solid. But I asked him one question: “And what about Cpk?”
He said, “We do not check it as often. Cp always seemed good enough.”
Three weeks later, that same process delivered a batch that was out of specification. The customer rejected it. When we sat down to review what had happened, we both learned the same lesson: the Cp was excellent, but the process had been slowly moving toward the upper limit for months, millimeter by millimeter, and no one noticed because no one was watching the Cpk.
That day I understood something that changed the way I think: trusting only Cp is like looking at the speedometer without looking where the steering wheel is pointing. You may be going at the perfect speed and still drive off the road.
That experience is why I want to explain two ideas that are often taught poorly in Six Sigma courses: Cp and Cpk. I am not going to start with formulas. You already have manuals for that. Instead, I want to give you a simple picture you can remember the next time you look at a capability report.
The truth is that many people use Cp and Cpk as if they were the same thing. They are not. Confusing them can cost you money, customers, and credibility. The research is clear: Cp measures variation; Cpk measures how well the process is centered and performing (Dong et al., 2021; Springer, 2023).
I also say this from experience. When I started in quality, I confused them too. It took a couple of customer rejections and several long nights for me to truly understand the difference. So, I will explain it the way I wish someone had explained it to me.
Let’s use a garage as an example.
Imagine the garage door is the customer’s specification limit: the range of results the customer will accept. The door is 2.5 meters wide, and your car is two meters wide. That leaves twenty-five centimeters of space on each side. At first glance, everything looks safe.
This is what Cp tells you. It answers one question: does the car fit in the garage? In process terms, Cp checks whether the total process variation is smaller than the tolerance. If the car is narrower than the garage door, Cp says there is enough room.
That matches the formal definition: Cp evaluates variation compared with tolerance (Dong et al., 2021). It tells you whether the process has enough potential space to operate within the limits.
But there is one important thing Cp does not tell you: where the car is positioned inside the garage.
You might drive in close to the left wall one day and close to the right wall the next. The car is still the same width, so Cp does not change. But if you keep drifting toward one side, sooner or later you hit the wall.
That is where Cpk comes in. Cpk asks a different question: is the process centered, or is it drifting toward one of the limits?
The research defines it this way: Cpk measures how far the process average is from the closest specification limit (Springer, 2023). In simple words, it tells you how close you are to the nearest wall.
If the process is perfectly centered, Cpk = Cp. If the process is off center, even a little, Cpk becomes lower than Cp. That is why a good Cp can still hide a problem: the process may have enough room, but it may be using that room badly.
Now let’s move from the garage to a real manufacturing example. Imagine a shaft that should measure 10.00 mm, with a tolerance of ±0.10 mm. That means every shaft must be between 9.90 mm and 10.10 mm.
Your process has low variation, so the measurements are very consistent. Cp is high, let’s say 1.67. On paper, that looks excellent.
But the process average is 10.08 mm. The parts are still inside the specification, but they are very close to the upper limit. That is why Cpk drops to 0.8.
This is exactly what the studies warn about: a high Cp can hide an off-center process, which is why Cpk is essential (Dong et al., 2021).
That is the trap. The process may be capable in theory, but it is not safe in practice because it is no longer centered. Cp shows the potential. Cpk shows the current reality.
Here is the golden rule:
- High Cp and low Cpk means your process has potential, but it is off center. Adjust the average.
- If Cp and Cpk are the same, your process is centered. Good. But do not get comfortable. Check it regularly.
- Cpk never, ever goes above Cp. If you see that in a report, someone made a mistake with the data. Cp is the ceiling. This is documented in the literature: Cpk ≤ Cp always (Springer, 2023).
Deming said you cannot inspect quality into a product; you have to build it into the process. Cp and Cpk are not just numbers to fill out a report. They are warning signals. But a warning signal only helps if you understand what it is telling you.
So here is the simple summary:
- Cp measures potential. Would the process fit inside the tolerance if it were centered?
- Cpk measures reality. Is the process inside the limits and centered right now?
- If you only look at Cp, you are only seeing half the picture.
- If Cpk is low and Cp is high, do not change the whole process right away. First, center it. Then optimize it.
Now bring this back to your own plant, team, or reports. Are you looking at Cp, Cpk, or both? And how many decisions are being made with only half the information?
Pause the episode and think about the last quality problem you had. Was it truly a capability problem, or was the process simply drifting away from the center? If you do not know, you already have your first task for tomorrow.
Thank you for all the ratings on my books: The Quality Mindset, Quality Principles, and Quality Life Projects. If this episode helped you understand something that had been confusing you for a while, please like it, subscribe, and share it with a colleague who only looks at Cp.
That is all for now. Stay excellent, keep improving, and do not be the manager who discovers too late that the process had been drifting for months and no one noticed.
REFERENCES
- Dong, P., Wang, Y‑B., Peng, D‑Z., Wang, J‑J., Cheng, Y‑T., Deng, X‑Y. & Zheng, B. (2021). Utility of process capability indices in assessment of quality control processes at a clinical laboratory chain. Journal of Clinical Laboratory Analysis, 35(8), e23878. https://doi.org/10.1002/jcla.23878
- Springer (2023). Method for determining process capability indices for non‑normal data. Applied Sciences / Manufacturing Quality Series.
- Springer (2023). Beyond Regular SPC: Bridging the Cpk Capability Index for Advanced Manufacturing