Last quarter, our quality team rejected three Cpk reports from different stamping suppliers—all showed impressive numbers, but none included control charts 1, raw data, or proof of process stability.
To request valid Cpk and Ppk process capability reports from stamping parts suppliers, you must define a complete capability study protocol in your RFQ—specifying Critical to Quality features, sample sizes, measurement methods, required output formats, and minimum index targets like Ppk ≥ 1.33—rather than simply asking for a number.
A single capability index on a PDF tells you almost nothing. What matters is the evidence behind that number. This article walks you through exactly what to include in your request, how to evaluate your supplier’s understanding, what values to require, and how to verify the data is real.
What Should I Include in a Cpk and Ppk Request to Avoid Confusion with My Supplier?
One lesson we learned early in our supply chain management 2 work was this: vague requests produce vague reports. When we simply wrote "provide Cpk data" on a purchase order, we received everything from a single number scribbled on a sticky note to a 40-page report with no context. The confusion was entirely our fault.
Your Cpk and Ppk request must include the drawing revision, Critical to Quality dimensions with specification limits, required measurement equipment, minimum sample size of 125 parts, production stage, target index values, required output formats like control charts and histograms, and a reaction plan for out-of-control conditions.
/12-05-2025-Pictures%20-%20Name/Hydraulic%20press%201.jpg)
Build a Capability Study Protocol, Not Just a Request
The root problem is that most buyers treat capability reports like a checkbox. They ask for "Cpk" and expect the supplier to know what they mean. But stamping suppliers vary widely in their statistical process control 3 maturity. Some have full SPC software. Others measure five parts with calipers and calculate an average.
You need a capability study protocol embedded in your RFQ. This protocol removes ambiguity. It tells the supplier exactly what to measure, how to measure it, how many parts to measure, and what format to deliver.
The 9-Point RFQ Checklist
Here is the checklist we now include in every stamping RFQ that requires process capability data:
| # | Item | What to Specify | Why It Matters |
|---|---|---|---|
| 1 | Drawing & revision | Part number, revision level, material grade, thickness | Ensures data matches current design |
| 2 | CTQ features | Tolerance limits (USL/LSL), datum references | Focuses study on critical dimensions only |
| 3 | Measurement method | Equipment type, fixture, operator procedure, units | Prevents gage-induced variation |
| 4 | Sample size | Minimum 125 consecutive parts | Provides statistical validity |
| 5 | Subgroup size & frequency | e.g., 5 parts every 30 minutes | Captures time-based variation |
| 6 | Production stage | Trial run, pilot, or stable production | Prevents trial-run data substitution |
| 7 | Required output | Cp, Cpk, Ppk, X-bar/R chart, histogram, raw data | Enables independent verification |
| 8 | Capability target | e.g., Ppk ≥ 1.33, Cpk ≥ 1.67 for safety features | Sets clear pass/fail criteria |
| 9 | Reaction plan & retention | What to do if out-of-control; how long to keep records | Ensures ongoing accountability |
Why Each Item Prevents Misunderstanding
Item 6 is particularly important for stamping parts. Stamping dies experience thermal expansion 4 and progressive wear. A trial run of 50 parts on a cold press will produce tighter data than a 10,000-piece production run. If you do not specify "stable production," the supplier may submit trial-run Cpk values that look excellent but do not reflect real manufacturing conditions.
Item 3 matters because Measurement System Analysis 5 must come before capability analysis. If the gage cannot resolve to at least one-tenth of the tolerance, the measurement error will inflate the standard deviation and corrupt the capability index. We always require a Gage R&R report alongside the capability data.
Item 7 is where most requests fail. A Cpk number without a control chart is meaningless. The control chart proves process stability. The histogram shows distribution shape. The raw data lets you run your own calculations. Without these, you are trusting a number you cannot verify.
When we prepare a Production Part Approval Process 6 submission for our U.S. clients, we include all nine items as mandatory deliverables. This eliminates back-and-forth and sets the expectation from day one.
How Do I Know If My Stamping Parts Supplier Actually Understands Process Capability Requirements?
A buyer we work with in the Midwest once shared a frustrating story: he asked three stamping suppliers for Ppk reports, and two of them confused Ppk with Cpk. The third sent a dimensional inspection report with averages and ranges—no capability index at all.
You can assess a supplier's understanding by asking them to explain the difference between Cpk and Ppk, requesting a sample capability report from a previous project, verifying they use proper SPC software, and checking whether their quality management system includes documented procedures for capability studies.

Test Their Knowledge Before Placing the Order
The cheapest quality investment you can make is a 15-minute call with the supplier's quality engineer before you award the contract. Ask direct questions. Do not accept marketing answers. Here are the questions we use during supplier audits in Vietnam and across Asia:
| Question | What a Good Answer Sounds Like | Red Flag |
|---|---|---|
| What is the difference between Cpk and Ppk? | "Cpk uses within-subgroup variation; Ppk uses total variation including between-subgroup shifts" | "They are the same" or no answer |
| What sample size do you use for capability studies? | "Minimum 100–125 parts from a production run" | "We measure 5–10 parts" |
| Do you run control charts before calculating capability? | "Yes, we verify process stability first with X-bar/R or individual-moving range charts" | "We just calculate the number" |
| What software do you use for SPC? | Names a specific tool (Minitab, QI Macros, InfinityQS, 1Factory) | "We use Excel" (not necessarily a red flag, but needs follow-up) |
| How do you handle non-normal data? | "We check normality and apply Box-Cox or Johnson transformation if needed" | "What do you mean by non-normal?" |
Look at Their Quality Management System
Suppliers certified to IATF 16949 7 are more likely to have embedded SPC and capability study procedures. But certification alone does not guarantee competence. During our supplier audits, we review their quality manual and look for documented procedures covering:
- How they identify Critical to Quality characteristics on a drawing
- Their standard operating procedure for conducting a capability study
- How they define subgroup size and sampling frequency
- Their criteria for rejecting their own process when capability falls below target
If these procedures do not exist, the supplier is likely generating capability data ad hoc. That means the data quality depends on whoever happens to run the study that day.
Request a Sample Report
Before placing the order, ask the supplier to share a redacted capability report from a previous stamping project. Look for these elements: a header showing part number, revision, and date; identification of the measured characteristic with specification limits; a histogram overlaid with spec lines; an X-bar/R or I-MR control chart; the calculated Cp, Cpk, and Ppk values; sample size; and the measurement instrument used.
If the sample report is just a table of Cpk numbers with no supporting charts, the supplier does not understand what a valid capability study looks like. That is your signal to either invest time training them or find a different source.
We have found that spending one hour on a capability study kickoff call with the supplier's quality team saves weeks of rework and re-measurement later. It is far easier to align expectations before production than to argue about data after shipment.
What Cpk and Ppk Values Should I Require for Critical Stamping Dimensions?
When our team reviews a new stamping project for a U.S. client, one of the first decisions is which dimensions need formal capability targets—and how high those targets should be. Setting the bar too low invites defects. Setting it too high for every feature wastes time and money.
For critical stamping dimensions, require a minimum Ppk of 1.33 for standard characteristics and Ppk of 1.67 for safety-critical or fit-critical features. For initial Production Part Approval Process submissions, a Cpk of 1.67 is common to demonstrate short-term process potential before long-term monitoring begins.

Understand What the Numbers Mean in Practice
A Cpk or Ppk of 1.00 means the process just barely fits within the specification limits, with roughly 2,700 parts per million 8 falling outside tolerance. A value of 1.33 tightens that to about 63 ppm. A value of 1.67 reduces it to less than 1 ppm. For stamping parts—where a single out-of-tolerance bracket can cause assembly failure—1.33 is the practical minimum.
Here is how different capability values translate to real-world performance:
| Capability Index Value | Approximate Defect Rate (ppm) | Sigma Level | Typical Application |
|---|---|---|---|
| 0.67 | 45,500 | 2σ | Unacceptable for most stamping |
| 1.00 | 2,700 | 3σ | Marginal; requires 100% inspection |
| 1.33 | 63 | 4σ | Standard for non-safety features |
| 1.50 | 6.8 | 4.5σ | Preferred for fit-critical features |
| 1.67 | 0.57 | 5σ | Required for safety-critical features |
| 2.00 | 0.002 | 6σ | World-class; rarely required |
How to Decide Which Dimensions Need Higher Targets
Not every dimension on a stamping part needs a capability study. Focus on characteristics that affect fit, function, or safety. On a typical mounting bracket like the ones we source—with bent flanges, punched mounting holes, and slotted features—the Critical to Quality dimensions usually include:
- Hole position relative to datum
- Hole diameter (especially if press-fit)
- Bend angle and bend radius
- Overall length or width that interfaces with mating parts
- Flatness of mounting surfaces
Non-critical features like edge trim dimensions or cosmetic radii typically do not need formal Cpk/Ppk tracking. Requesting capability on every dimension increases cost and lead time without improving quality where it matters.
Cpk for Trial vs. Ppk for Production
One important distinction: during initial Production Part Approval Process submissions, you may accept Cpk data from a controlled trial run. This Cpk reflects short-term process potential under ideal conditions. But once you move into volume production, switch to requiring Ppk. Ppk captures all variation—tool wear, material lot changes from different coils, operator shifts, and thermal steady-state differences in the stamping press.
For stamping, Ppk is the more honest index. A die that stamps beautifully for the first 500 parts may drift as it wears. Material from a new heat lot may have slightly different mechanical properties that affect springback. Ppk catches these shifts. Cpk does not.
We recommend specifying both in your quality agreement: Cpk ≥ 1.67 at PPAP, and Ppk ≥ 1.33 for ongoing production monitoring. This gives you a tight gate at approval and a realistic standard for sustained performance.
How Can I Verify That the Cpk and Ppk Data My Supplier Provides Is Accurate and Not Fabricated?
Early in our work managing supply chains across Asia, we encountered a situation that made us permanently skeptical of unsupported capability numbers. A stamping supplier submitted a Ppk of 1.85 for a critical hole position on a mounting bracket. The number was suspiciously high. When we asked for raw data, the supplier hesitated. When we finally received it, the 125 measurements had almost zero variation—statistically impossible for a stamping process running across multiple coils.
To verify Cpk and Ppk accuracy, require raw measurement data in production sequence, demand control charts and histograms alongside the indices, confirm an acceptable Measurement System Analysis was completed before the study, check the data for statistical normality, and look for suspicious patterns like unnaturally low variation or rounded numbers.
/ICHI-%20Photo/150-ton%20Stamping%20Press%20Machine%201.jpg)
The Three-Check Validation Framework
Before you accept any capability report, run three foundational checks. If any one fails, reject the report and request a re-study.
Check 1: Measurement System Analysis (MSA)
The measurement system must be capable before the process data means anything. Request the Gage R&R study. Total gage R&R should be below 10% of the tolerance for the measured characteristic. The gage resolution must be at least one-tenth of the tolerance range. If the supplier used a caliper with 0.01 mm resolution to measure a tolerance of ±0.05 mm (total range 0.10 mm), the resolution is exactly 1/10—barely acceptable. A micrometer or CMM would be better.
Check 2: Process Stability
Look at the control charts. Before calculating Cpk or Ppk, the process must be in statistical control. That means no points beyond the control limits 9, no runs of 7+ consecutive points above or below the center line, and no obvious trends. If the control chart shows instability, the capability index is mathematically invalid—even if the number looks good.
Check 3: Distribution Shape
Look at the histogram. The data should approximate a normal distribution 10. If it is skewed, bimodal, or truncated (which can happen when parts are sorted or 100% inspected), the standard Cpk/Ppk formula does not apply. The supplier should use a Box-Cox or Johnson transformation and report the adjusted capability, or use a nonparametric method.
Red Flags in Supplier-Submitted Data
Here are specific patterns that suggest the data may be fabricated or non-representative:
| Red Flag | What It Suggests | What to Do |
|---|---|---|
| All measurements cluster at the nominal with near-zero spread | Data may be fabricated or from a very small sample | Request raw data in production sequence; verify sample size |
| Measurements are all rounded to the same decimal place | Gage resolution may be too coarse, or data is manually entered/edited | Ask for the measurement equipment model and resolution |
| No control chart or histogram provided | Supplier may not have run a proper study | Reject the report; require full output |
| Cpk and Ppk are identical or nearly identical | Suggests no between-subgroup variation was detected; data may be from a single short run | Require data from multiple shifts, coils, or setups |
| Sample size is below 100 | Statistical confidence is too low | Require minimum 125 parts |
| Data is not in production order | Cannot verify stability or detect trends | Require time-stamped or sequence-numbered measurements |
Demand Contextual Evidence
Beyond the three checks, require the supplier to submit contextual information that traces the data back to real production conditions. This includes:
- The production date and shift
- The press and die identification
- The material lot or coil number
- The operator name or ID
- The measurement instrument serial number and last calibration date
This traceability to the 5Ms+1E (Man, Machine, Method, Material, Measurement, Environment) makes fabrication much harder. A supplier who can provide all of this is demonstrating a mature quality management system. A supplier who cannot is telling you something important.
Move Toward Real-Time Access
One trend we are seeing among forward-thinking buyers is a shift from static PDF reports to real-time SPC access. Some buyers now request API or cloud access to the supplier's SPC software so they can monitor control charts and capability indices in real time. This eliminates the lag between production and reporting, and it makes fabrication nearly impossible.
We are not there yet with all our suppliers, but it is the direction the industry is moving. For now, the three-check framework plus raw data and contextual evidence is the most reliable way to verify accuracy.
For high-volume stamping, we also recommend requesting stroke-interval capability snapshots. These are capability studies taken at defined intervals during a long production run—say every 10,000 strokes. They show how dimensional drift correlates with die wear and help predict when maintenance is needed. This kind of tolerance analysis goes beyond a single capability number and gives you a dynamic picture of process health.
Conclusion
Requesting valid Cpk and Ppk reports from stamping suppliers requires a defined protocol—not just a line item. Specify your CTQ features, sample requirements, output formats, and validation checks upfront. The effort you invest in the request directly determines the quality of what you receive.
Frequently Asked Questions
What Should I Include in a Cpk and Ppk Request to Avoid Confusion with My Supplier?
One lesson we learned early in our supply chain management 2 work was this: vague requests produce vague reports. When we simply wrote "provide Cpk data" on a purchase order, we received everything from a single number scribbled on a sticky note to a 40-page report with no context. The confusion was entirely our fault.
How Do I Know If My Stamping Parts Supplier Actually Understands Process Capability Requirements?
A buyer we work with in the Midwest once shared a frustrating story: he asked three stamping suppliers for Ppk reports, and two of them confused Ppk with Cpk. The third sent a dimensional inspection report with averages and ranges—no capability index at all.
What Cpk and Ppk Values Should I Require for Critical Stamping Dimensions?
When our team reviews a new stamping project for a U.S. client, one of the first decisions is which dimensions need formal capability targets—and how high those targets should be. Setting the bar too low invites defects. Setting it too high for every feature wastes time and money.
Footnotes
1. Authoritative guide on control charts from the NIST Engineering Statistics Handbook. ↩︎
2. Comprehensive overview of supply chain management principles and practices. ↩︎
3. Authoritative NIST handbook introduction to statistical process control. ↩︎
4. Scientific explanation of thermal expansion in materials and engineering. ↩︎
5. Comprehensive Wikipedia overview of measurement system analysis. ↩︎
6. Detailed explanation of the Production Part Approval Process (PPAP) requirements. ↩︎
7. Detailed Wikipedia entry for the IATF 16949 automotive standard. ↩︎
8. Scientific definition of parts-per notation from Wikipedia. ↩︎
9. Technical definition of control limits in statistical process control. ↩︎
10. NIST Engineering Statistics Handbook guide to the normal distribution. ↩︎
