Definition
What is a Layered Process Audit question?
An LPA question checks whether a defined process control is in place and being followed at the point where work happens. It names something the auditor can see, measure, compare, or ask to be demonstrated, and ties that observation to a stated requirement. It is not a finished-product inspection and not a general opinion about the process.
Working definition
A usable LPA question identifies a process control, gives the auditor something specific to observe, defines what it is compared against, and leads to a clear response when the observed condition does not meet the requirement.
Questions commonly come from process documentation, control-plan and PFMEA thinking, customer complaints, scrap or rework history, process changes, and previous findings. For the broader program, see the Layered Process Audit guide.
Requirements vs. examples
Before you copy any question: check your actual requirements
None of the questions on this page is presented as universally required. Customer-specific requirements can add rules that matter to question content, review, or follow-up.
General Motors example
GM's current Customer-Specific Requirements for IATF 16949 explicitly require an internal layered process audit process. The requirements include periodic review of layered-audit questions so they can be changed to focus on organizational weaknesses, and findings that cannot be corrected during the audit move to an action plan monitored to closure.
Read the current GM customer-specific requirements →AIAG CQI-8 guidance
AIAG's public CQI-8 overview describes the second-edition guideline as including expanded guidance for writing effective LPA checksheet questions and process-specific checksheets. Treat it as industry guidance, not a universal question list.
See AIAG's CQI-8 overview →Use your own source documents
If you supply an OEM or another customer with specific audit requirements, read the current requirement directly. Do not adopt a vendor blog, an example bank, or this page as a substitute.
Question anatomy
What makes a good LPA question?
Control
The process control being protected.
Torque setting
Object
What the auditor actually looks at.
Torque-driver display
Requirement
What the condition is compared against.
Current released work instruction
Method
How the auditor verifies it.
Read the display and compare
Evidence
What is useful to record.
Observed value or revision
Response
What happens on a Fail.
Finding with owner and due date
Wording rules
Specific
Name the item and the requirement.
Observable
The auditor can see, measure, compare, or ask for a defined demonstration.
Objective
A value, ID, revision, reference, or record reduces interpretation.
One control per question
Bundled questions make Pass / Fail ambiguous.
Compliant means Pass
Avoid negatively inverted wording that invites mistakes.
Short enough to repeat
A frequent audit cannot support long investigations in every question.
Worth failing
If nobody would act on a Fail, ask why the question is on the checklist.
Rewrite pattern
Weak vs. better LPA questions
| Weak question | Why it fails | Better question / action |
|---|---|---|
| Is the operator following the process? | No step is named; the answer becomes an opinion. | Does the operator perform [step name] in the sequence shown in the current standard work? |
| Are the tools OK? | "OK" is undefined and no tool is named. | Is [tool ID] the tool named on the work instruction, with its verification status current? |
| Does the operator know how to handle nonconforming material? | Knowledge cannot be verified by a vague yes/no. | Can the operator point to the designated container and identify who decides disposition? |
| Is the product free of defects? | This inspects output, not the process control. | Was the in-process check required by the control plan completed at the stated interval for this run? |
| Is the area organized? | Housekeeping is broad and not tied to a process control. | Are only the components for the current job present at the station? |
| Are the records complete? | The record and completion point are undefined. | Is the run record complete through the last operation finished? |
| Is the control plan being followed? | An entire document is bundled into one question. | Does the check at this station match the method and frequency defined for this operation? |
| Is the equipment well maintained? | No equipment, task, or standard is named. | Is the preventive-maintenance task due this period for [equipment ID] recorded as complete? |
| Is the FMEA current? | This is usually not a station-level observation. | Move this to the appropriate periodic PFMEA review or higher-level review rather than forcing it into a line-side check. |
The last example matters: some questions do not belong on a line-side LPA at all. For checklist-level design, see the Layered Process Audit checklist guide.
Question bank
59 Layered Process Audit question examples
Use this bank as a source of ideas, not a checklist to adopt whole. Replace bracketed terms and acceptance references with your own process controls. Numeric requirements should come from your approved documents, not from a generic example.
Category AStandard work and point-of-use instructions
5 questions
Standard work and point-of-use instructions
| # | Question | What it verifies | Example evidence |
|---|---|---|---|
| 1 | Is the work instruction posted at the station the current released revision? | Posted revision against release list | Revision seen matches released version |
| 2 | Does the operator perform [named step] in the sequence shown in current standard work? | One full cycle observed | Sequence observed and compared |
| 3 | Are the visual standards at the station the current approved set? | Sample or photo ID against approved list | Reference IDs match |
| 4 | Can the operator show the reaction plan for [named failure] and identify who to contact? | Defined reaction path | Plan shown and contact identified |
| 5 | Is the operator listed as qualified for this operation on the plant's qualification record? | Qualification record | Operator appears on current record |
Category BProcess parameters
5 questions
Process parameters
| # | Question | What it verifies | Example evidence |
|---|---|---|---|
| 6 | Does the recipe or program ID loaded match the setup sheet for the part number running? | Controller vs. setup sheet | IDs match |
| 7 | Do the displayed [named parameters] fall within the approved process range? | Display vs. setup sheet or control plan | Observed values recorded |
| 8 | Is the setting protection for [named parameter] in place where the process requires it? | Lock, password, seal, or other defined control | Protection observed active |
| 9 | Is the alarm or monitoring function for [named parameter] active and not disabled? | Controller status | Status observed active |
| 10 | Is the in-process parameter record complete through the last required interval? | Record vs. defined interval | Latest entry within interval |
Category CMistake-proofing (poka-yoke)
5 questions
Mistake-proofing (poka-yoke)
| # | Question | What it verifies | Example evidence |
|---|---|---|---|
| 11 | Using the approved challenge method, does [device] reject the challenge piece? | Device response | Challenge ID and result |
| 12 | Is every device on the station's error-proofing list present and enabled? | Device list vs. station | All required devices present |
| 13 | Is the challenge-verification record complete for the current required period? | Verification record | Current entry present |
| 14 | Where a device is bypassed, is the bypass authorized under the plant's defined bypass process? | Bypass status and authorization | Authorization reference |
| 15 | Are challenge pieces identified, controlled, and stored as the plant defines? | Reference-piece control | Piece IDs and storage condition |
Category DTooling and fixtures
4 questions
Tooling and fixtures
| # | Question | What it verifies | Example evidence |
|---|---|---|---|
| 16 | Does the fixture ID installed match the fixture listed for the current part number? | Fixture vs. setup requirement | Fixture ID recorded |
| 17 | Is [tool, insert, or electrode] within the replacement limit in the tool-life record? | Counter or log vs. limit | Counter reading recorded |
| 18 | Are locating and clamping elements free of damage or wear defined as a reject condition? | Physical condition vs. defined reference | Photo or note if Fail |
| 19 | Is the tool ID in use the tool named on the work instruction? | Tool ID vs. instruction | Tool ID matches |
Category EGauges and measurement
4 questions
Gauges and measurement
| # | Question | What it verifies | Example evidence |
|---|---|---|---|
| 20 | Is the gauge in use the one named on the control plan or work instruction? | Gauge ID vs. process document | Gauge ID recorded |
| 21 | Is the gauge's calibration or verification status current? | Gauge label or system status | Due date or status recorded |
| 22 | Was the required start-of-shift gauge check completed and recorded, where applicable? | Gauge check record | Current entry present |
| 23 | Is the gauge free of visible damage defined as a reject condition? | Gauge condition | Photo or note if Fail |
Category FMaterial identification and segregation
5 questions
Material identification and segregation
| # | Question | What it verifies | Example evidence |
|---|---|---|---|
| 24 | Does the part number on each component container match the setup sheet for the current job? | Container label vs. setup sheet | Part number seen |
| 25 | Does the material lot at the point of use match the lot on the production record? | Material label vs. production record | Both lots recorded |
| 26 | Are only components for the job currently running present at the station? | Station contents vs. job | Unexpected material noted |
| 27 | Where material has a limited life, is the material in use within its defined date? | Material label date | Date recorded |
| 28 | Do containers carry the status identification the plant requires? | Released / hold / other defined status | Status tag matches |
Category GTraceability
3 questions
Traceability
| # | Question | What it verifies | Example evidence |
|---|---|---|---|
| 29 | Does the lot or date code on the most recent unit match the run record? | Unit vs. record | Both identifiers recorded |
| 30 | Are serial or lot records complete for the most recent units completed? | Traceability record | Units checked noted |
| 31 | Is the traveler or build record completed through the last operation finished? | Traveler vs. actual process state | Last operation noted |
Category HFirst-piece and startup verification
4 questions
First-piece and startup verification
| # | Question | What it verifies | Example evidence |
|---|---|---|---|
| 32 | Was first-piece verification completed and recorded for the current run? | First-piece record | Time and lot recorded |
| 33 | Does the first-piece record show approval by a person the plant authorizes to approve? | Approver vs. authorization | Approver ID |
| 34 | After the last shutdown or restart, was startup verification completed before production resumed? | Restart time vs. verification record | Both times recorded |
| 35 | Where the plant retains a first-piece or reference sample, is it available and identified? | Reference sample control | Sample ID |
Category IChangeover and setup
4 questions
Changeover and setup
| # | Question | What it verifies | Example evidence |
|---|---|---|---|
| 36 | Is the checklist for the last changeover complete and signed where required? | Changeover record | Changeover time and record |
| 37 | Does the installed tooling match the setup sheet for the part number running? | Tooling IDs vs. setup | IDs recorded |
| 38 | Was the previous job's material and labeling cleared before the new job started? | Station clearance | Anything remaining is described |
| 39 | Is setup approval recorded before production release where the plant requires it? | Setup approval record | Approver and time |
Category JNonconforming material and rework
5 questions
Nonconforming material and rework
| # | Question | What it verifies | Example evidence |
|---|---|---|---|
| 40 | Is all nonconforming material in the designated container, tagged, and outside the production flow? | Segregation and identification | Tag reference |
| 41 | Does each nonconforming tag show the information the plant requires? | Tag content | Required fields present |
| 42 | Can the operator show where suspect parts go and who decides disposition? | Defined material-control path | Demonstration noted |
| 43 | Are reworked parts routed through the defined rework path and re-verified before returning to flow? | Rework route | Route observed |
| 44 | Is the rework instruction at the rework station the current released revision? | Revision control | Revision seen |
Category KPreventive controls and equipment condition
3 questions
Preventive controls and equipment condition
| # | Question | What it verifies | Example evidence |
|---|---|---|---|
| 45 | Is the preventive-maintenance task due this period for [equipment ID] recorded as complete? | PM record | Completion date |
| 46 | Is [critical consumable] replaced within its defined interval? | Replacement log vs. interval | Last replacement date |
| 47 | Where the process depends on it, is the environmental condition within the range the work instruction requires? | Display or environmental record | Observed reading |
Category LPackaging and labeling
5 questions
Packaging and labeling
| # | Question | What it verifies | Example evidence |
|---|---|---|---|
| 48 | Is the label revision loaded at the station the current revision for this customer and part? | Label vs. released list | Revision recorded |
| 49 | Does the label content match the customer label specification for this part and lot? | Label vs. specification | Label reference or photo |
| 50 | Does the quantity in the container match the pack instruction? | Count or approved method vs. instruction | Quantity recorded |
| 51 | Is the packaging material the type named in the pack instruction? | Packaging material vs. instruction | Material ID |
| 52 | Have obsolete labels and packaging been removed from the station? | Station condition | Anything found described |
Category MPrevious corrective actions
4 questions
Previous corrective actions
| # | Question | What it verifies | Example evidence |
|---|---|---|---|
| 53 | Is the control added under [finding or corrective-action reference] still in place? | Corrective control | Reference and observation |
| 54 | Does the control added after [complaint reference] work as intended when demonstrated? | Corrective control effectiveness | Demonstration result |
| 55 | Is the revised step added after [issue reference] being performed as written? | Changed standard work | Observed cycle |
| 56 | Where a temporary control is in use, is it within its authorized period and documented? | Temporary control authorization | End date and reference |
Category NHousekeeping only where it protects a control
3 questions
Housekeeping only where it protects a control
| # | Question | What it verifies | Example evidence |
|---|---|---|---|
| 57 | Are the sealing or locating surfaces of [fixture] free of debris as the work instruction defines? | Condition affecting the process control | Photo or note if Fail |
| 58 | Is the station free of unidentified parts that could be mixed with the current job? | Mix-prevention control | Anything found described |
| 59 | Are sensor or scanner lenses on [device] clean enough to function as the work instruction defines? | Device condition | Condition noted |
General 5S scoring usually belongs in a separate audit. Housekeeping earns a place in an LPA when it directly protects a process control—for example, debris on a locating surface or unidentified material that could be mixed into the current job.
Process-specific wording
How question wording changes by manufacturing process
The practical LPA examples page shows full audit scenarios. This table focuses only on wording.
| Process | Example question | Why the wording works |
|---|---|---|
| Final assembly | Does the fastening tool ID at the station match the tool listed on the current work instruction? | Names the object and reference. |
| Final assembly | Does the sub-assembly part number match the part number on the job traveler? | Compares two identifiers. |
| Final assembly | Using the approved missing-component challenge, does the presence check produce the expected stop or reject? | Builds the test method into the question. |
| CNC machining | Does the tool-life counter for [tool position] show the tool within the limit on the setup sheet? | Points to a counter and a limit. |
| CNC machining | Does the first-piece record show the critical characteristics checked with the specified gauge? | Links the record to the measurement control. |
| CNC machining | Is the part seated on the locating features defined in the setup sheet? | Anchors 'correctly seated' to a reference. |
| Injection molding | Does the resin lot on the hopper label match the lot on the production record? | Two records, one comparison. |
| Injection molding | Does the mold ID on the tool match the scheduled job? | Single-ID check. |
| Injection molding | Where drying is required, does the dryer show the approved setting and is the required drying record complete? | Conditional wording avoids false failures. |
| Welding / fabrication | Is the filler or consumable identification the one named in the work instruction? | Names the consumable and reference. |
| Welding / fabrication | Where the plant or applicable process requires a welding procedure, is the referenced procedure the one in use? | Conditional; does not invent a requirement. |
| Welding / fabrication | Does the first-article record show the required inspection completed before production continued? | Checks sequence using a record. |
| Paint / coating | Does the oven setpoint fall within the approved range for the selected part and recipe? | The range lives in plant documents. |
| Paint / coating | Was coating-thickness verification done at the defined interval and recorded for the current run? | Checks timing and record, not a made-up number. |
| Paint / coating | Is the mix batch identified with the time or expiry the process requires, and is it still within that limit? | Identification plus limit. |
| Packaging | Does the label revision in the print queue match the released label list for this customer part? | Revision against release list. |
| Packaging | Does the customer pack requirement at the station show the current revision? | Document control at point of use. |
| Packaging | Does the quantity in a sampled container match the pack instruction? | Defines sample and reference. |
Layer perspective
How questions can change by audit layer
Titles and layer counts vary by plant. Using mistake-proofing as one thread, the same control can be examined at different levels without merely repeating the same signature exercise.
| Layer | Perspective | Example question | What it adds |
|---|---|---|---|
| Layer 1 | Station-level direct check | Using the approved method, does [device] produce the expected reject right now? | Confirms the control works now. |
| Layer 2 | Area-level follow-up | Did the control added after the last [device] finding hold, and does it work now? | Re-verifies a previous fix. |
| Layer 3 | Program-level pattern | Which mistake-proofing checks failed more than once this period, and are any related findings overdue? | Looks for recurrence and stalled action. |
Higher-layer prompts may be answered partly from records and reports. That can be useful when the layer adds a genuinely different perspective. For cadence and rotation, use the LPA schedule guide.
Shared core vs. additions
Should every layer use the same questions?
Shared core
Repeating a small set of critical controls across layers creates independent verification of the same operational risks.
Core + layer-specific additions
Higher layers can add recent-finding, recurrence, or follow-through questions that match their broader span of control.
Both models can be reasonable. The useful test is whether each layer adds real verification or insight rather than reproducing paperwork.
Question count
How many questions should an LPA contain?
There is no standard number. Published recommendations vary, and they are not universal requirements. Start with the time a competent auditor can actually spend at the process, rank candidate controls by consequence of failure, and include only what can be observed properly in that time.
Coverage vs. repeatability
More questions cover more possible risk but slow the audit and invite rushed answers. Fewer questions run faster but can omit a critical control. Pilot the checklist, then adjust it from actual use rather than a generic target number.
Answer design
Writing Pass / Fail / N/A questions
Pass
Observed condition meets the stated requirement.
Fail
Observed condition does not meet the requirement.
N/A
The question genuinely does not apply to this audit.
- Give the requirement. For numeric checks, recording the observed value can be more useful than a bare Pass.
- Define N/A in advance. “I could not find the answer” is not the same as not applicable.
- Capture an N/A reason where useful. Repeated N/A can reveal poor scope.
- Do not let N/A hide a control. If the control should have been verified, fix the question or the audit process.
After a failed question
What happens when an LPA question fails?
The LPA finding supports follow-up; it does not replace the plant's nonconformance or containment procedures. The original audit answer stays as submitted, while the finding carries the changing status. The main LPA guide covers the broader lifecycle.
Learn from recurrence
Questions after a recurring failure
Repeat failures are one of the strongest reasons to review an LPA question set. Suppose a customer complaint is traced to a wrong label revision at packaging. The plant might add: “Is the label revision loaded at the station the current revision for this customer and part?”
| Trigger | Control behind it | Question action |
|---|---|---|
| Wrong label complaint | Label revision control | Add or strengthen a revision-at-station check. |
| Repeat failure | Control not holding | Rewrite for clarity or add higher-layer re-verification. |
| Process change | New or changed step | Add a question for the new control where recurring verification adds value. |
Revision control
Change questions without rewriting history
Suppose Revision 2 had no changeover-verification question. After a production issue, Revision 3 adds one. Audits completed under Revision 2 should remain tied to Revision 2; future audits use Revision 3 after its effective date.
REVISION 2
Historical audits stay here
REVISION 3
New question applies going forward
The LPA template guide shows a practical way to separate checklist definitions, audit records, and findings.
Failure modes
Common LPA question mistakes
| Mistake | What it looks like | Fix |
|---|---|---|
| Vague wording | "Is the process followed?" | Name the step and reference. |
| Yes/no with no reference | "Is the setting correct?" | State what it is compared against; record the value where useful. |
| Housekeeping overload | Half the form is cleanliness. | Keep housekeeping only where it protects a process control. |
| Copied from another process | Torque question appears on a coating line. | Rebuild from the actual process controls. |
| Cannot be observed | "Are operators trained?" | Check an appropriate record or demonstration instead. |
| Every audit always passes | Years of clean results. | Review whether the question or the observation quality is weak. |
| Too many questions | Audits run long and get rushed. | Cut to controls that matter. |
| No response to a Fail | Failures sit on the form. | Define owner, due date, and verification flow. |
| Stale after process change | Instruction references are obsolete. | Tie review to process changes. |
| Paperwork only | Every check is record-based. | Include direct observation where the control is physical. |
Audit quality
How to avoid audit theater
An LPA loses value when people complete forms without verifying anything. The warning signs are practical: unrealistically fast audits, long stretches of perfect results, auditors who cannot describe what they saw, higher-layer work that keeps slipping, and failed checks that never create follow-up.
Record observed values where useful
A number can preserve more evidence than a bare Pass.
Use real demonstrations
Where safe and defined, verify the control rather than only reading a record.
Capture context on every Fail
The owner should not have to reconstruct what the auditor saw.
Calibrate auditors together
Two people should interpret the same question the same way.
Review questions that never fail
They may protect a stable critical control—or they may be too vague to detect anything.
Review higher-layer completion
A layered program is not layered if the upper layers quietly disappear.
Pre-publish test
A question-review checklist
Control
You can name the process control it protects.
Object
The auditor knows exactly what to look at.
Requirement
A document, value, ID, sample, or state defines the expected condition.
Method
The auditor knows how to check it.
Observable
It can be answered at the process in the time available.
Single control
It does not bundle several independent checks.
Polarity
The compliant condition maps cleanly to Pass.
Response
The plant has decided what happens on a Fail.
Owner
Someone is responsible for maintaining the question.
Simple editorial rule
If a proposed question fails several of these tests, rewrite it or remove it before the checklist is published.
Managing the question set
Excel vs. LPA software for managing questions
Excel can be entirely adequate for a small program: one site, a few processes, a few auditors, and a checklist that changes infrequently. The strain usually comes from coordination rather than the question text itself.
When coordination grows
When does LPAWorks help?
LPAWorks is focused LPA software for small manufacturers moving from Excel, paper, email, or shared folders. It does not write your questions or decide which controls matter. The wording work on this page still applies.
Focused LPA workflow
Keep the questions process-owned while software handles the coordination around them
LPAWorks supports checklist creation and supported CSV/XLSX import, controlled published revisions, recurring schedules, fixed or round-robin assignment, Pass / Fail / N/A execution, notes and private evidence, findings with owners and due dates, resolution, verification, reopen, reminders, overdue visibility, audit and finding history, dashboards, reports, and PDF/CSV exports.
Historical audits remain tied to the checklist revision used at the time. LPAWorks is not a full QMS and does not replace ERP, MES, CAPA, document-control, or LMS systems.
Current access
30-day free trial
No card is required for the 30-day trial. After the trial, the workspace is read-only until you subscribe.
Import your checklistFrequently asked questions
Layered Process Audit questions FAQ
What are good Layered Process Audit questions?
Good questions are specific, observable checks of a meaningful process control, tied to a reference the auditor can compare against. They are answerable in the time available and lead somewhere when the result is Fail.
What questions should be in an LPA?
Questions should protect controls where a lapse matters. Typical sources include process documentation, PFMEA and control-plan thinking, complaints, scrap or rework history, process changes, and previous findings. Which questions belong depends on the actual process.
How many questions should an LPA have?
There is no standard number. Use only the questions an auditor can genuinely observe within the audit's time budget, then adjust after real use.
Should all audit layers use the same checklist?
Not necessarily. A shared core of critical controls can provide repeated verification, while higher layers may add questions about recent findings, recurrence, or program execution.
Can LPA questions be yes/no?
A binary outcome can work when the criterion is objective. For numeric checks, recording the observed value as well as Pass or Fail often preserves useful evidence.
What makes an LPA question objective?
It compares an observable condition with a stated reference such as a value, revision, ID, sample, or record, reducing interpretation between auditors.
Should housekeeping questions be included?
Only where housekeeping directly protects a process control, such as debris on a locating surface or unidentified material that could be mixed into the current job.
What happens when an LPA question fails?
Record the observation, follow the plant's reaction or containment process where needed, create tracked follow-up with an owner and due date, then resolve and verify it. The original audit answer remains historical.
Can I use the same LPA questions for every process?
Usually not. A common format can help, but the actual questions should reflect each process's controls and risks.
How often should LPA questions be reviewed?
There is no universal interval. Review them after process changes, recurring failures, complaints, corrective actions, and when questions stop producing useful information. Customer-specific requirements may add their own expectations.
Should an old audit change after the checklist is revised?
No. Historical audits should remain tied to the checklist revision used at the time.
Can LPA questions be managed in Excel?
Yes. Excel can work well for a small program. Coordination becomes harder as checklist revisions, auditors, schedules, findings, evidence, and reporting multiply.
The takeaway
A good LPA question earns its place.
It names a real control, gives the auditor something objective to compare against, fits inside the time available, and leads somewhere when the answer is Fail. Use the 59 examples here as patterns, then rebuild them around your own process.