Skip to content

Question list · Quality management

Layered process audit questions: 56 to choose from, and why you ask each one

Short answer

Layered process audit questions are short yes or no checks that a leader answers at the station while the process runs. Each one verifies a single control: a step in the work instruction, a quality check, an error-proofing device, a setting or the fix for a past problem.

A good layered process audit question is closed, checks one thing, and can be answered at the station in under a minute by looking or testing. Write it so that yes always means the process conforms. Trace each one to a risk: a line in the control plan or PFMEA, a customer complaint, or an error-proofing device. This page lists 56 questions in eight groups, each with the reason it earns a place on an audit. Nobody asks all 56. A team leader's audit on every shift needs about 8 to 12 and takes 5 to 10 minutes, and the higher layers add the questions that check the audit itself. AIAG, which publishes the CQI-8 Layered Process Audits Guideline, describes the questions as the heart of an LPA system and says they should target the causes of the biggest issues for the plant and its customers. Start from your last complaints and your error-proofing devices, and fill the rest of the list from the groups below.

Last reviewed:

How to use this list

This is a bank to choose from, not one audit. Layered process audit questions are closed: the auditor answers yes, no or not applicable at the station while the process is running. A short audit that is done on every shift is worth more than a long one that gets skipped, so pick the few questions that match the risks at each station and leave the rest.

The 56 questions in eight groups

  1. Questions 1 to 7

    Work instructions

    Check that the written method is at the station and the person doing the job was trained to it.

  2. Questions 8 to 15

    Quality checks

    Check that the inspections in the control plan are done on time, with the right gauge, and acted on.

  3. Questions 16 to 21

    Error proofing

    Prove that each device still catches the defect it was installed to catch.

  4. Questions 22 to 28

    Settings and tooling

    Check that the machine is running the approved setup and its tooling is inside its limits.

  5. Questions 29 to 36

    Material

    Check that the right material is in use and that bad parts cannot find their way back into the flow.

  6. Questions 37 to 42

    Past problems

    Check that the fix for the last problem is still in place and the team still knows about it.

  7. Questions 43 to 48

    Safety

    Check the few safety conditions an auditor can confirm in a minute at any station.

  8. Questions 49 to 56

    Upper layers

    Check that the layer below is auditing, that findings get closed, and that the list is still current.

Groups 1 to 7 are asked at the station. Group 8 is for the layers that audit the audit.
  • Write every question so that yes means conforming. A list where some good answers are yes and others are no produces a score nobody can read.
  • One question checks one thing. If a question contains the word and, split it.
  • Ask about what you can see or test at the station today. Whether a document was reviewed belongs in a system audit.
  • Trace each question to a risk: a control plan or PFMEA line, a customer complaint, or an error-proofing device. If you cannot say what would go wrong, drop the question.
  • Every no is fixed on the spot or gets an action with an owner and a date before the auditor leaves the station.
  • Keep error-proofing verification and any open containment on every audit.

The layered process audit checklist has 20 checks in four groups, a layer and frequency matrix, short question banks for stamping, welding, molding and assembly, and a findings table. Use it as the starting sheet and swap in questions from this page. AIAG's own articles about CQI-8 say the questions can change often and that many organizations adjust them from time to time. A person does that: change a few at a time and keep most of the list stable so the score still means something. If the method is new to you, read what a layered process audit is first.

Put your questions on the LPA checklist

A yes, no or N/A sheet for the station, a layer and frequency matrix, question banks for four processes and a findings table. Free, after one short form.

Free to use: print it, copy it and edit it for your team. Enter your name and work email once to download.

See what is on each sheet on the LPA checklist page.

Work instructions and training

Start with the method, because every other control assumes the job is done the way it was approved. Watch one full cycle before you answer any of these. Two or three belong on every audit.

  1. 1

    Is the current revision of the work instruction posted at the station?

    Why: An old revision at the station means an engineering change stopped in the office. Compare the revision on the sheet with the document list.

  2. 2

    Is the operator doing the steps in the order the instruction shows?

    Why: Watch one full cycle before you answer. A changed sequence is how a step, and the check that goes with it, gets dropped.

  3. 3

    Can the operator show you the key points for this job?

    Why: Key points are the few things that make or break the part. If the operator cannot point to them, the training covered the motions and not the reasons.

  4. 4

    Is the operator signed off for this job on the training matrix?

    Why: Cover for absence puts people on jobs they were never trained for. This catches it on the day it happens, not after the defect.

  5. 5

    Is the work instruction at the station for this part number, not a similar one?

    Why: Stations that run several variants collect instructions. The wrong one within reach is a mix-up waiting for a busy shift.

  6. 6

    Are the visual aids and limit samples present, legible and in date?

    Why: A faded photo or a limit sample with no date gets ignored. Operators then judge borderline parts from memory.

  7. 7

    After the last changeover, was the first-piece check done and recorded?

    Why: The first part after a changeover carries the highest risk of a wrong setup. The record shows that someone looked before the run started.

Quality checks at the station

These questions test whether the checks in the control plan are happening at the station today. Bring the control plan line with you, and read the check sheet against the clock, not against yesterday's page.

  1. 8

    Are the checks in the control plan being done at the stated frequency?

    Why: Compare the check sheet with the clock. A check due every hour with the last entry three hours ago has stopped.

  2. 9

    Are today's results written down as they happen?

    Why: A sheet filled in ahead of time or at the end of the shift is a record of intention. Look for entries made at the times the checks were due.

  3. 10

    Is the gauge named in the control plan the one in use?

    Why: A substitute gauge may read differently or not check the feature at all. Match the gauge number to the plan.

  4. 11

    Is the gauge within its calibration date?

    Why: An out-of-date sticker puts every reading since the due date in doubt. That can turn one finding into a sort.

  5. 12

    Does the operator know the limits for the check?

    Why: Ask for the number. An operator who records values without knowing the limits cannot react to a bad one.

  6. 13

    Does the operator know what to do when a check fails?

    Why: Listen for stop, segregate and call. If the reaction plan lives in a binder, the next out-of-limit part will keep moving.

  7. 14

    Was the last out-of-limit result followed by the reaction plan?

    Why: Find the last bad reading on the sheet and ask what happened next. This checks the reaction plan in use, not in theory.

  8. 15

    Is the master or reference part available and identified?

    Why: Without the master, nobody can prove the gauge or the vision system still tells good from bad.

Error proofing and detection devices

A device that has stopped working looks exactly like one that works, which is why these questions are tested and not just looked at. Keep error-proofing verification on every audit, whatever else changes on the list.

  1. 16

    Was each error-proofing device verified at the start of the shift?

    Why: A sensor that failed overnight looks the same as one that works. Only a check with a known bad part tells them apart.

  2. 17

    Does the device reject the challenge part when you test it now?

    Why: Ask the operator to run the challenge part while you watch. A record says it was checked. The test shows it works.

  3. 18

    Are the challenge parts marked and kept in their own place?

    Why: An unmarked defective part kept near the line is a defect waiting to ship. A marked one in a fixed place is a tool.

  4. 19

    Is every device switched on, with no bypass in use?

    Why: Devices get bypassed to keep the line running during a fault and stay that way. Look at the key switch and the screen, not the logbook.

  5. 20

    If a device was bypassed, is the approved backup check in place?

    Why: A bypass can be legitimate for a short time. Without the extra inspection that goes with it, the line runs unprotected.

  6. 21

    Are rejected parts going into a locked or marked reject container?

    Why: An open reject chute lets a part be picked out and put back in the flow. The container is the last part of the device.

Process settings, tooling and maintenance

Parts made on the wrong setting can pass every check at the station and fail later. Read the settings from the machine yourself, and take the setup sheet with you so you compare against the approved values and not against memory.

  1. 22

    Do the machine settings match the setup sheet?

    Why: Read three settings from the screen and compare them. Drift from the approved window usually starts as a small adjustment to get through a shift.

  2. 23

    Is the approved program or recipe loaded for this part number?

    Why: The wrong recipe can make parts that look right and fail later. Check the program name against the setup sheet.

  3. 24

    Were today's operator maintenance checks done?

    Why: Daily lubrication and cleaning checks are the first to be skipped when the line is behind.

  4. 25

    Is the tool or die within its life count?

    Why: A tool run past its count makes a defect that grows slowly. The counter and the limit should both be visible at the machine.

  5. 26

    Are torque tools set to the specified value and verified today?

    Why: A torque tool can drift with no sign at the joint. The daily verification is the only evidence the fastener is right.

  6. 27

    Are fixtures and nests clean and undamaged?

    Why: Chips or weld spatter in a nest shift the part by a fraction. That shows up later as a dimensional problem with no clear source.

  7. 28

    Is the last setup change recorded with who made it and why?

    Why: An unrecorded adjustment cannot be linked to the defect it caused or cured.

Material, labelling and nonconforming parts

Wrong-part and mixed-part complaints rarely come from the process itself. They come from what sits around it: containers, labels, hold areas and the scrap bin. These questions take a minute each and are answered with your eyes.

  1. 29

    Is the material at the station the right part number for this job?

    Why: Read the label and compare it with the build sheet. Look-alike components are an easy way to build the wrong part.

  2. 30

    Is every container at the station labelled?

    Why: A container with no label holds parts of unknown status. Sooner or later someone decides they are good.

  3. 31

    Is the lot or batch being recorded where traceability requires it?

    Why: If a supplier lot turns out bad, this record is what limits the sort to one day of production.

  4. 32

    Is material being used first in, first out?

    Why: Old stock pushed to the back ages or misses an engineering change. Check the dates on the front and back containers.

  5. 33

    Are nonconforming parts tagged and in the marked hold area?

    Why: A bad part on the corner of the bench has one tired moment between it and the customer.

  6. 34

    Is the scrap bin free of parts that look good?

    Why: Good parts in the red bin mean the operator is unsure of the standard, or the bin is being used as storage.

  7. 35

    Is rework being done only to an approved instruction?

    Why: Informal rework has no check behind it. If rework is happening at the station, ask to see the instruction.

  8. 36

    Do finished containers hold the right quantity with the right label?

    Why: Wrong count and wrong label are complaints that have nothing to do with the part, and they are the easiest to catch here.

Past complaints and corrective actions

This is the group that changes most often. Each time a complaint or an internal escape is closed, write one question that checks the fix at the station, and keep it on the audit until the fix has clearly held. A person decides when it comes off.

  1. 37

    Is the quality alert for the latest customer complaint posted here?

    Why: The people making the part should know what the customer found. If the alert is not at the station, the complaint was closed on paper.

  2. 38

    Can the operator describe the defect in that alert?

    Why: Posting is not the same as knowing. Ask what the defect looks like and where on the part to look.

  3. 39

    Is the corrective action from the last complaint still in place?

    Why: Corrective actions fade when attention moves on. Go to the fixture, the check or the instruction that was changed and see it working.

  4. 40

    Is the containment check still running if its exit criteria are not met?

    Why: Extra inspections get dropped quietly after a few clean days. Check the agreed end date or quantity against the record.

  5. 41

    Is the top defect for this station shown with its current trend?

    Why: A team that can see its biggest defect is working on it. An empty or month-old chart means nobody is looking.

  6. 42

    Was the open finding from the last audit closed by its due date?

    Why: An audit whose findings stay open teaches the area that findings do not matter.

Safety and workplace

A layered process audit is not a safety inspection, but a leader who walks past a missing guard to check a gauge sends the wrong message. Two or three of these on each audit is enough.

  1. 43

    Is the operator wearing the protective equipment the job requires?

    Why: The quickest check on the list, and the one people notice a leader skipping.

  2. 44

    Are guards, light curtains and interlocks in place and working?

    Why: A guard taken off to clear a jam and not put back is an injury on hold.

  3. 45

    Is the emergency stop reachable and unobstructed?

    Why: Pallets and carts drift in front of emergency stops. Reach for it from the operator's position.

  4. 46

    Are aisles, exits and electrical panels clear?

    Why: Blocked access is easy to see and fix on the spot. Recording it shows whether the same spot fills up every week.

  5. 47

    Is the lockout procedure for this machine posted and readable?

    Why: People clearing a jam need the steps in front of them, not in an office.

  6. 48

    Is the station free of leaks, trip hazards and loose items?

    Why: Oil on the floor is a slip risk and an early sign of a failing seal.

Questions for layers 2 and above

A supervisor or manager repeats a few station questions, then audits the audit. These questions are what make the method layered: each level checks that the level below it is looking, recording honestly and closing what it finds.

  1. 49

    Were all of this week's layer 1 audits done on schedule?

    Why: The first job of a higher layer is to check that the layer below is auditing. A missed audit is a finding in its own right.

  2. 50

    Do the layer 1 answers match what you see at the station?

    Why: Repeat two questions from the last layer 1 audit. Straight yes answers next to an obvious problem mean the sheet is being ticked.

  3. 51

    Does every no from the past week have an action with an owner and a date?

    Why: Finding without fixing is how a layered audit dies. Pick three findings and trace them.

  4. 52

    Are overdue findings being escalated?

    Why: An overdue action that nobody raised has been accepted.

  5. 53

    Is this station free of findings that keep coming back?

    Why: A repeat means the last fix treated the symptom. It should go to problem solving, not be corrected a fourth time.

  6. 54

    Do the questions cover the latest complaints and process changes?

    Why: A checklist written two years ago audits two-year-old risks. Compare the list with the last three quality alerts.

  7. 55

    Are the audit results posted where the team can see them?

    Why: People who never see the results assume nothing is done with them.

  8. 56

    Did a leader from each layer audit this area in the period?

    Why: The method depends on every layer of management going to the process. If the plant manager's audits are the ones slipping, operators notice.

Which questions each layer asks

Every layer audits the same stations, but not with the same list. The layers, frequencies and times below are the ones on the checklist template. They are common choices, not a rule, and a customer requirement overrides them.

Which question groups each layer takes

  • Layer 1: team leader

    How often and how long
    Every shift, 5 to 10 minutes
    Questions to take
    The core station questions: work instructions, quality checks, error proofing and any open containment.

  • Layer 2: production supervisor

    How often and how long
    Weekly, 10 to 15 minutes
    Questions to take
    A shorter set of station questions, two of them repeated from the last layer 1 audit, plus whether layer 1 audits were done and every no has an action.

  • Layer 3: area manager

    How often and how long
    Monthly, 15 to 20 minutes
    Questions to take
    A few station questions, then overdue findings, repeat findings and whether the list covers the latest complaints.

  • Layer 4: plant manager

    How often and how long
    Monthly, 15 to 20 minutes
    Questions to take
    A few station questions on the highest risks, then whether every layer audited this area and results are posted.
The higher the layer, the fewer station questions it takes and the more it checks the audit itself.

Weak audit questions, and how to rewrite them

A weak question lets the auditor say yes without looking at anything. Each one below is rewritten into a question from the bank that has to be checked at the station.

Weak audit questions and better ones

  • Weak question

    Is quality OK at this station?

    What goes wrong: Nobody can answer no. There is nothing to look at.

    Rewrite

    Are the checks in the control plan being done at the stated frequency?

  • Weak question

    Is the gauge OK?

    What goes wrong: A glance at the gauge says yes.

    Rewrite

    Is the gauge within its calibration date?

  • Weak question

    Are the settings correct and the checks recorded?

    What goes wrong: Two checks in one line. A yes for one hides a no for the other.

    Rewrite

    Do the machine settings match the setup sheet? Then, as its own line: are today's results written down as they happen?

  • Weak question

    Is error proofing working?

    What goes wrong: Answered by looking at the device, which looks fine either way.

    Rewrite

    Does the device reject the challenge part when you test it now?

  • Weak question

    Are there any nonconforming parts lying around?

    What goes wrong: Yes is the bad answer, so the score reads backwards.

    Rewrite

    Are nonconforming parts tagged and in the marked hold area?

  • Weak question

    Was the PFMEA reviewed?

    What goes wrong: A document question that belongs in a system audit, not at the station.

    Rewrite

    Is the corrective action from the last complaint still in place?

A good audit question can be answered yes or no by looking, in under a minute.

Worked example: choosing 12 questions for a weld cell

Illustrative example: the plant, the cell and the numbers are made up to show how risks turn into questions. A quality engineer builds the layer 1 audit for a projection weld cell that welds nuts to brackets. She lists the six things most likely to go wrong, notes where each one comes from, and takes one question from the bank for each.

From risk to question for a weld cell

Example numbers

  • Missing weld nut

    Where it comes from
    A customer complaint three months ago
    Question on the audit
    Does the device reject the challenge part when you test it now? The device is the nut presence sensor.

  • Weak weld

    Where it comes from
    Control plan: destructive test every two hours
    Question on the audit
    Are the checks in the control plan being done at the stated frequency?

  • Left and right brackets mixed

    Where it comes from
    PFMEA: the two brackets look alike
    Question on the audit
    Is the material at the station the right part number for this job?

  • Worn electrode tips

    Where it comes from
    Maintenance history
    Question on the audit
    Is the tool or die within its life count? Here that is the tip dress counter.

  • Untrained cover on nights

    Where it comes from
    Training matrix: two new operators
    Question on the audit
    Is the operator signed off for this job on the training matrix?

  • Suspect parts back in the flow

    Where it comes from
    A finding in the last internal audit
    Question on the audit
    Are nonconforming parts tagged and in the marked hold area?
Every question on the sheet traces back to a line on the risk list.

The other six come from the fixed part of the list: work instruction revision, first-piece check after changeover, protective equipment, guards, the containment check from the complaint, and whether the last finding was closed. That makes 12 questions and about ten minutes. The first 14 audits give 168 answers. Nine are no, so the score is 159 ÷ 168, or 95 percent. Five of the nine are the tip dress counter. A finding that repeats five times in 14 audits is not corrected a sixth time. It goes to problem solving, starting with the 5W1H questions.

What to do with the answers

The audit is finished when every no has somewhere to go. Fix what can be fixed at the station and record that it was fixed. Everything else gets an owner and a date before the auditor walks away.

  • Score the audit as yes answers divided by yes plus no. Leave out not applicable. Post the score and the open findings where the team meets.
  • A no on an error-proofing device stops the audit. Contain the parts made since the last good check, then log it on the error-proofing verification log. The poka-yoke guide covers how devices are verified.
  • A finding that repeats goes to CAPA, not back onto next week's list unchanged.
  • When a complaint is closed, add one question that checks its fix. When a question has been yes for months and the risk has gone, a person swaps it for a newer one.

Closed questions verify a standard. They do not tell you why the standard is hard to follow. For that, use the open gemba walk questions on a separate visit. For workplace organization, the 5S audit questions are scored 1 to 5 and are better kept as their own audit.

Layered Process Audit Software

Run layered process audits in LeanSuite

It schedules each audit to every layer of leadership, puts the questions on a phone at the station, scores the answers and turns every No into an action with an owner.

See Layered Process Audit SoftwareBest layered process audit software

Go deeper

This page is the short version. These guides cover each part in full.

All our articles on this subject are in the quality management topic.

Free templates and tools

Key terms

Lean glossary

Doing this in LeanSuite

See it on your own lines

Bring the problem you are working on to a 45-minute demo. Or look around with sample data first, or pilot LeanSuite on your own lines for 90 days, with the setup done by our team.

Book a 45-minute demoClick through the interactive demoStart a 90-day pilot

FAQ

Questions about layered process audit questions

  1. Short yes or no questions that verify one control each: whether the current work instruction is at the station and followed, whether quality checks are done on time with the right gauge, whether error-proofing devices reject a challenge part, whether settings match the setup sheet, and whether the fix for the last complaint is still in place. Higher layers also ask whether the audits below them were done and the findings closed.

  2. Few enough that the audit is done on every shift. A layer 1 audit on the checklist template takes 5 to 10 minutes. With questions that take under a minute each, that is about 8 to 12. The checklist template has 20 checks in four groups to choose from. If audits are being skipped, cut the list before you cut the frequency.

  3. Yes. Each question is answered yes, no or not applicable, and it is written so that yes always means the process conforms. That makes the audit quick, lets different auditors give the same answer, and gives a score: yes answers divided by yes plus no.

  4. Whenever the risks change. AIAG's articles on its CQI-8 guideline say the questions can change often and that many organizations adjust them from time to time. Add a question when a complaint is closed or a process changes, and retire one that has been yes for months. Change a few at a time, and keep error proofing and open containment on every audit.

  5. Layered process audit questions are closed and verify that a standard is being followed, so the answer is yes or no. Gemba walk questions are open and help a leader understand how the work is done and what gets in the way. Use the audit to verify and the walk to learn.

Related guides and tools

More method guides

Or browse the topics, the full library and the lean glossary.

Rather see it on your own problem?

Bring the problem you are stuck on, and we will show you on a call how LeanSuite handles it.

Schedule Demo

Pass it on

We want LeanSuite to be the best place on the internet for lean help. If this was useful, send it to someone on your team or in your network who needs it.