PPAP Element 11 Initial Process Studies: What Ppk Value Passes and Why Reviewers Reject the Number Before They Read It
A supplier quality engineer opens Element 11 of a Level 3 PPAP package on a Tuesday afternoon. The part is a machined shoulder pin with one Special Characteristic, a diameter of 12.000 plus or minus 0.025. The supplier has submitted a capability summary that reads Cpk 2.14. It is a beautiful number. The reviewer does not celebrate. She looks at three things in order. There is no control chart behind the number, just a histogram. The study is labeled Cpk, not Ppk, on a brand new process that has never run production before. And the measurement system analysis in Element 9 came back at 28 percent GRR on a characteristic with a 0.050 total tolerance. She writes the kickback in ninety seconds. The Cpk of 2.14 is meaningless, because the process was never demonstrated stable, the wrong index was reported for an initial study, and the gauge eats more than a quarter of the tolerance the number claims to control.
Element 11 is the element where PPAP stops being a document exercise and becomes a statistics exercise. Most PPAP rejections that survive a first read are Element 11 rejections. A part number can be wrong on Element 1 and get corrected in a day. A capability claim that rests on an unstable process or an unqualified gauge cannot be corrected without rerunning the study, and rerunning the study means another significant production run. That is why reviewers attack Element 11 first and hardest.
This post walks Initial Process Studies from a reviewer's seat under AIAG PPAP 4th Edition. What Element 11 actually requires and on which characteristics, why an initial study reports Ppk and not Cpk, the three-tier acceptance criteria at 1.33 and 1.67 and the reaction each tier demands, the two prerequisites a reviewer verifies before trusting any index, where the characteristics under study come from, the non-normal and one-sided characteristic trap that voids an otherwise clean number, the five kickback patterns that account for most Element 11 rework, and a pre-submission QA pass a supplier quality lead can run in fifteen minutes before the package ships.
For the full submission context, PPAP Element 11 sits inside the 18 element package and every element has to survive the same reviewer. For the foundational orientation, what PPAP is and why the customer asks for it.
What Element 11 Actually Requires
AIAG PPAP 4th Edition defines Element 11, Initial Process Studies, as the demonstration that the process producing the part is capable of meeting the customer's requirements before regular production begins. The word initial is load bearing. This is not the ongoing capability the process will show after six months of production. It is the first statistical evidence, taken from a controlled production event, that the process can hold the characteristics that matter.
Element 11 does not require a study on every dimension on the drawing. It requires a study on the Special Characteristics: the characteristics the customer or the organization has designated as significant or critical because a departure affects safety, regulatory compliance, fit, function, or downstream processing. A drawing with forty dimensions might carry three Special Characteristics. Element 11 covers those three. Running capability on all forty is not thoroughness, it is noise, and a reviewer reads it as a supplier who does not know which characteristics are special.
The study has to come from a significant production run. AIAG PPAP defines this as production from the intended production process, using production tooling, production gauging, production operators, production materials, and production environment, run at the quoted rate. Parts from a prototype cell, a soft tool, or an R and D bench are not admissible. The reviewer checks the traveler and the run date to confirm the parts came from the process that will actually ship. A capability study on prototype parts is one of the most common Element 11 rejections, because it proves nothing about the process the customer is approving.
Minimum data volume for the study runs to a significant production quantity. The working standard is a minimum of 300 consecutive production parts as the significant run, from which the capability sample is drawn as at least 25 subgroups of 4 to 5 parts, or a minimum of 100 individual measurements. Fewer than 100 readings is a sample size finding on its own, before the reviewer even looks at the number the sample produced.
Ppk, Not Cpk: Why the Initial Study Uses the Performance Index
This is the distinction that decides more Element 11 kickbacks than any other, and it is the one most suppliers get backwards.
Cpk and Ppk both compare the spread of the process against the tolerance and account for how far off center the process sits. They differ in which standard deviation they use. Cpk uses the within-subgroup, short-term standard deviation, estimated from the average subgroup range divided by d2. It answers the question, how capable is this process at its best, when only the moment-to-moment variation inside a subgroup is present. Ppk uses the overall, total standard deviation calculated from every individual reading in the study. It answers the question, how did this process actually perform across the whole run, including the drift, the tool wear, and the shift-to-shift variation that lives between subgroups.
The rule for Element 11 follows from what an initial study can honestly claim. To report Cpk validly, you have to have demonstrated that the process is in a state of statistical control, because the within-subgroup sigma is only a fair estimate of process spread when the between-subgroup variation is negligible. A brand new process on its first significant run has not earned that claim. It has one run of data. It cannot yet show that tomorrow's run will look like today's. So AIAG PPAP directs the initial study to report Ppk, the performance index built on total variation, because Ppk makes no assumption about long-term stability. It reports what happened, all of it.
Once the process has accumulated enough history to be shown stable and in control, ongoing capability migrates to Cpk. Element 11 is the initial study. It reports Ppk. A supplier who submits Cpk on a first-article significant run is either claiming a stability they have not demonstrated or does not understand the difference. Either reading costs credibility.
The deeper mechanics of when each index applies, how the two sigmas are calculated, and what a 1.33 Cpk and a 1.33 Ppk actually mean in defect terms are covered in full in Cpk vs Ppk: which process capability index to use for PPAP. Element 11 is the specific place that distinction gets graded.
The Three-Tier Acceptance Criteria
AIAG PPAP sets the acceptance criteria for initial process studies in three tiers, and each tier carries a required reaction, not just a pass or fail label.
Index greater than 1.67. The process meets requirements. Submit the study with the package, note the result on the Control Plan, and move the characteristic to ongoing monitoring at the Control Plan frequency. For critical characteristics, 1.67 is the working floor, not a stretch goal, because a 1.67 index still corresponds to a nonzero long-term defect rate once the process is allowed to shift the customary 1.5 sigma.
Index between 1.33 and 1.67. The process may be acceptable, but not on the supplier's say-so. AIAG PPAP requires the supplier to contact the customer and agree on the disposition before submission. That disposition usually means enhanced controls: tightened Control Plan reaction rules, increased sampling frequency, or additional containment until the index improves. Submitting a 1.45 index with no evidence of a customer conversation and no enhanced controls in the Control Plan is a kickback. The number is in the conditional band, and the package treats it as if it passed.
Index below 1.33. The process does not meet requirements. The supplier cannot self-approve out of this band. AIAG PPAP requires containment, which in practice means 100 percent inspection of the characteristic until capability is restored, a documented corrective action plan with a timeline, and customer notification. The part can still get PPAP approval, but as an interim approval, not a full approval, and the interim carries the containment and the corrective action commitment as conditions. A sub-1.33 index submitted as if it were a full approval, with no containment and no corrective action, is the single most consequential Element 11 rejection, because it means the supplier intends to ship an incapable process to production without a safety net.
The thing to internalize is that Element 11 is not a gate that reports a binary. It is a gate that assigns a reaction. The reviewer is checking that the reaction in the package matches the tier the number landed in.
The Two Prerequisites a Reviewer Verifies Before Trusting the Number
A capability index is a ratio built on two measured quantities: the spread of the process and the location of its center. Both are measured through a gauge, and both assume the process holding still enough to be measured. If either assumption fails, the index is a number with no meaning. Reviewers know this, so they verify two prerequisites before they read the index at all.
The Measurement System Has to Be Acceptable First
Every reading in the capability study passed through a gauge. If the gauge itself contributes a large share of the observed variation, then some of the spread the study attributes to the process is actually measurement error, and the capability number is corrupted. This is why Element 9, Measurement System Analysis, is a prerequisite for Element 11, not a parallel deliverable.
The working rule from AIAG MSA is that a Gauge R and R under 10 percent is acceptable, 10 to 30 percent is conditional depending on the application and the cost of the characteristic, and over 30 percent is unacceptable. A capability study run through a 28 percent GRR gauge on a tight tolerance is reporting a process spread that is partly the gauge. The reviewer cross-checks the gauge ID on the Element 11 study against the accepted gauge in Element 9. If the MSA is not acceptable, or if the gauge used for the study is not the gauge that was qualified, the capability number is void regardless of how good it looks.
The full acceptance framework, including why percent GRR alone is not enough and where the number of distinct categories matters, is in Gauge R and R acceptance criteria: percent GRR, NDC, and what AIAG MSA requires. Element 11 is where an unqualified gauge quietly poisons a capability claim.
The Process Has to Be in Statistical Control
The second prerequisite is stability. A capability index computed on a process that is drifting, cycling, or reacting to special causes is describing a distribution that does not exist as a stable thing. This is why AIAG PPAP requires control chart evidence with the study, not just a histogram and a computed index.
The reviewer looks for an X-bar and R chart, or an individuals and moving range chart for low-volume characteristics, with the study data plotted in production sequence. The chart has to show statistical control: no points beyond the control limits, no runs of seven or more on one side of the centerline, no trends, no obvious cycling. If the chart shows an out-of-control condition, the supplier has to have investigated it, identified the special cause, and either removed the affected data with documented rationale or rerun the study. A histogram alone is a red flag on Element 11, because a histogram hides the time order, and time order is exactly where stability lives or dies. A capability index without a control chart behind it is an index the reviewer cannot trust, because they cannot see whether the underlying process was even stable enough to characterize.
Where the Characteristics Under Study Come From
Element 11 does not invent its own scope. The characteristics it studies are the Special Characteristics, and those are designated upstream, in a chain the reviewer can trace.
Special Characteristics originate in the design and process risk analysis. A characteristic gets flagged as significant or critical in the Design FMEA and the Process FMEA because its failure mode carries a high severity, often coupled with a real occurrence risk. That designation flows into the Control Plan, which lists each Special Characteristic with its specification, its measurement method, its sample size and frequency, and its reaction plan. Element 11 then runs capability on exactly those Control Plan Special Characteristics, using the Control Plan sampling as the basis for the study.
This traceability is something a reviewer checks. The Special Characteristics studied in Element 11 have to match the Special Characteristics on the Control Plan, which have to match the significant and critical items carried out of the PFMEA. A characteristic that is special on the Control Plan but missing from Element 11 is an under-scoping finding. A characteristic studied in Element 11 that is not special anywhere upstream is either over-scoping or a sign the designation flow is broken. When the chain is clean, Element 11 reads as the statistical proof of the risk decisions the PFMEA and Control Plan already made.
For how special and critical designations get established and propagated in the process risk analysis, see the practical guide to Process FMEA. For how the Control Plan carries those designations into the sampling that Element 11 draws from, see what goes in a manufacturing Control Plan and how to build one.
The Non-Normal and One-Sided Characteristic Trap
The standard Ppk formula assumes the data follows a normal distribution. That assumption holds for a lot of dimensional characteristics: a machined diameter, a length, a bore. It fails, quietly and often, on geometric characteristics, and when it fails, the computed index is wrong even though every step of the arithmetic was correct.
The most common offenders are the one-sided geometric characteristics. True position, flatness, runout, perpendicularity, and concentricity are bounded at zero. You cannot have negative true position. Their distributions are not normal, they are folded or half-normal, piled up against the zero boundary with a long tail toward the tolerance. Feeding that data into a normal Ppk calculation produces an index that can look either falsely good or falsely bad, because the formula is fitting a symmetric bell to a distribution that is nothing like a bell.
The correct handling is one of two paths. Either use a non-normal capability method that fits the actual distribution, such as a Weibull-based or a percentile-based approach, or, for a one-sided characteristic where only the upper bound matters, report the upper capability alone rather than forcing a two-sided index onto a one-sided requirement. Either way, the study has to include a non-normal data assessment, which AIAG PPAP calls out explicitly. The reviewer looks for evidence that normality was checked, not assumed. A true position characteristic with a two-sided Ppk of 1.9 and no normality assessment is a characteristic where the reviewer already knows the 1.9 is fiction.
This is the trap because the number looks clean. Nothing on the summary sheet signals a problem. The failure is invisible unless the reviewer knows to ask whether the characteristic is one-sided and whether normality was tested. Experienced reviewers ask it every time.
Five Kickback Patterns on Element 11
Across a stack of Element 11 rejections, the same five patterns account for most of the rework.
One, wrong index for an initial study. Cpk reported on a first significant run where stability was never demonstrated. The fix is to report Ppk on the total variation, or to first establish and document stability if a Cpk claim is genuinely intended.
Two, no control chart. A histogram and a computed index with no evidence the process was in statistical control. The fix is the X-bar and R or individuals and moving range chart plotted in production sequence, with any out-of-control condition investigated.
Three, unqualified measurement system. The gauge used for the study is not the gauge accepted in Element 9, or the MSA came back above the acceptable GRR threshold for the tolerance. The fix is to qualify the measurement system first and rerun the study on the qualified gauge.
Four, conditional or failing index submitted as a pass. An index between 1.33 and 1.67 with no customer conversation and no enhanced controls, or an index below 1.33 with no containment and no corrective action plan. The fix is to match the reaction in the package to the tier the number landed in.
Five, non-normal data treated as normal. A one-sided geometric characteristic run through a two-sided normal Ppk with no normality assessment. The fix is a non-normal method or a one-sided index, with the distribution check documented.
Notice that four of the five are not about the process being bad. They are about the study being unable to support the claim it makes. A capable process with a broken study fails Element 11 exactly as hard as an incapable process, because the reviewer cannot approve a number they cannot trust.
A Pre-Submission QA Pass for Element 11
A supplier quality lead can run this pass in about fifteen minutes before the package ships, and it catches the five patterns above before a customer reviewer does.
- Confirm the scope. The characteristics in Element 11 match the Special Characteristics on the Control Plan, which match the significant and critical items on the PFMEA. Nothing extra, nothing missing.
- Confirm the source. The study parts came from a significant production run on production tooling, gauging, operators, and materials, not from prototype or soft-tool parts. Check the traveler and the run date.
- Confirm the sample size. At least 100 individual readings, or at least 25 subgroups of 4 to 5. Fewer is a finding on its own.
- Confirm the index type. Ppk on an initial study, not Cpk, unless stability has been demonstrated and documented and a Cpk claim is deliberate.
- Confirm the control chart. An X-bar and R or individuals and moving range chart is present, plotted in sequence, and shows statistical control. Any out-of-control point was investigated.
- Confirm the measurement system. The gauge on the study matches the accepted gauge in Element 9, and the MSA passed for the tolerance in question.
- Confirm the distribution. Any one-sided geometric characteristic has a normality assessment and a correct capability method, not a forced two-sided normal index.
- Confirm the reaction matches the tier. Above 1.67, clean submit. Between 1.33 and 1.67, customer agreement and enhanced controls in the package. Below 1.33, containment, corrective action plan, and interim approval, not a silent pass.
If all eight hold, Element 11 will survive the first read. If any one fails, fix it before the package ships, because every one of these is cheaper to fix on your desk than in a customer rejection that stalls the launch.
How QualityEngineer.ai Handles Initial Process Studies
The reason Element 11 fails so often is that the study is assembled by hand from three disconnected sources: capability math in one spreadsheet, the MSA in another, the Control Plan sampling in a third, and the reviewer is the first person to check whether they agree. QualityEngineer.ai closes those gaps by keeping the pieces connected.
The Analyze module runs the capability math directly, computing Ppk and Cpk with the correct sigma for each, generating the X-bar and R and individuals and moving range charts from the study data, and flagging non-normal distributions before they get forced into a normal index. The Control Plan built in the document cascade carries the Special Characteristic designations forward from the PFMEA, so the Element 11 scope matches the Control Plan by construction rather than by manual reconciliation. And when the study is assembled into the submission, the Package module runs gap analysis across the 18 elements, so an Element 11 index that sits in the conditional band without the matching enhanced controls, or an index resting on an unqualified gauge in Element 9, gets caught before the package leaves the building instead of after the customer opens it.
Element 11 is the element where a PPAP package proves it can actually make the part. It is worth getting right, because a rejection here does not cost a day. It costs another significant production run, and the launch waits on both.
For where Element 11 fits in the full submission, see the PPAP 18 elements checklist. For the index distinction it rests on, see Cpk vs Ppk. For the measurement system it depends on, see Gauge R and R acceptance criteria. For aerospace suppliers deciding whether a characteristic study belongs in a PPAP or a First Article, see FAI vs PPAP.
