Skip to content
CorrectiveActionReport.comSupplier CAR Response Desk — home

CAR & SCAR Response

Answering a Customer Corrective Action Request: The Supplier's Guide

A request arrives, the clock starts, containment goes out, and the 8D is built, signed off and returned. This guide walks the whole job from the supplier's side of the desk, across every customer you answer.

Rovaryn Digital · · 17 min read

When a supplier corrective action request lands on your desk

A customer's corrective action request rarely arrives at a convenient moment. Take a tier-1 customer that rejects a lot of formed brackets for burrs on a flange edge. The request form carries a part number, a defect photo, a short description and a reference number you have not seen before. Somewhere in that customer's supplier manual is a window for your first response, and it is already running.

The people and records you need are seldom in one place. The engineer who knows that customer may be off site, and the shop floor may not have heard about the rejection yet. A different customer's request from last month may mention a similar burr, but that file sits in another inbox, in another format, under another name.

Most of the work in a supplier corrective action request is not the analysis. It is reading the right rules, starting the right clocks, finding the affected material and keeping every request's history in one place you control. This guide walks that job from start to finish, from the supplier's side of the desk: what to read, what to log, how to contain, how to build the 8D and how to close it with evidence.

What a customer CAR, SCAR or 8D request asks of you

Customers use several names for the same basic act. A corrective action request (CAR), a supplier corrective action request (SCAR) and an 8D request all tell you that a customer has found a nonconformance attributed to your product or process and expects a documented response. If you want the plain definitions, the guide to what a corrective action request is covers them, and the SCAR versus CAR comparison explains how the labels differ between customers.

What matters in daily work is that the request asks you to do four things:

  1. Contain the problem, so suspect product does not keep reaching the customer.
  2. Explain the problem, with a root cause that covers both how it happened and why it escaped detection.
  3. Correct it, with permanent actions you can show are in place.
  4. Prove it, with evidence that the fix worked and a record the customer can accept.

The request is issued by the customer and runs inside the customer's process. Some customers send it by email with a form. Others require you to answer inside a portal they operate, such as a supply-chain platform's 8D-based problem-solving module. That is simply a channel the customer requires, and nothing about your own record changes that.

Who sits where

Three roles show up on nearly every request. The customer's supplier quality engineer (SQE) or buyer owns the request on their side. Your quality engineer or Supplier Quality Engineer usually owns the response on yours. The Quality Manager approves what goes out and answers for the pattern across customers. In a smaller shop, the Operations Manager often holds all three jobs for customer complaints.

Tier matters too. A tier-2 supplier may be answering a tier-1 customer that is itself passing down an OEM's complaint. In that case the tier-1's window can be shorter than the OEM's, because the tier-1 needs time to respond upward. Check what your customer expects, not what their customer expects.

Read your customer's rules before the clock gets ahead of you

Customers usually keep their response windows, fee terms, required formats and escalation steps in a supplier manual, in customer-specific requirements (CSRs), or in both. These documents are revised over time. Windows and fee terms differ from one customer to the next, and they can change between revisions of the same manual. Nobody can give you a safe general number for containment, root cause or closure. The only reliable number is the one in the current revision of your customer's document.

For IATF 16949 suppliers, the guide to IATF 16949 customer-specific requirements explains how CSRs fit alongside the standard. For timing in particular, the article on 8D response time and customer requirements shows why one customer's window cannot be assumed for another.

A checklist for reading the response section of a supplier manual

Run this checklist once per customer, then again whenever the manual is revised. Keep the answers in one place, not in a PDF you have to reopen each time.

  1. Revision and date. Which revision are you reading, and is it the one your customer currently expects you to follow?
  2. Trigger. What event starts the clock: a rejection, a notification level, a portal posting, an email?
  3. Containment window. How long do you have for an initial containment response? Is it hours, days or tied to the severity of the issue? Is it counted in calendar or working days?
  4. Plan window. How long do you have for root cause and a corrective action plan?
  5. Closure window. When must the final response, implementation and verification be complete?
  6. Format. Does the customer require its own form, a portal entry, a specific 8D layout or specific field order?
  7. Required content. Must you show an occurrence and an escape root cause? Must you attach evidence such as inspection records or photos?
  8. Fees. Does the manual list administrative, late-response or debit memo terms? Under what conditions, and can they repeat?
  9. Escalation. What happens after a missed window: a scorecard hit, a status change, controlled shipping, a new-business hold?
  10. Contacts. Who is the SQE or buyer, and who is the backup?
  11. Approval. Does the customer need to accept the response before the request is closed, and how is that acceptance communicated?

Some customers answer several of these in a single page. Others scatter them across a manual, an appendix and a CSR addendum. If an answer is not in the documents, ask the SQE and write down what they say, with the date.

Calendar days, working days and the clock you actually have

A clock that says "five days" is only useful once you know which kind of day it counts. The same start time can produce different due dates.

Question Why it changes the due date
Calendar or working days? Weekends and customer holidays may or may not count.
Whose holidays? A customer's plant holiday may differ from yours.
Does the clock start on issue or on your receipt? An email sent Thursday evening may not reach the right person until Friday.
Is the window "or sooner" depending on severity? A high-severity issue can shorten the window the manual lists for routine ones.
Is a portal timestamp the official start? The portal may stamp a different time than the email.

Write the interpretation next to the due date, with who confirmed it. Interpretations that live in one engineer's head are the first thing lost when that engineer leaves.

Where the standards sit behind the request

A customer request does not come from nowhere. For automotive suppliers, IATF 16949:2016 clause 10.2.3 requires a documented problem-solving process. In paraphrase, that process covers containment and interim actions, root cause analysis with its methodology and results, and systemic corrective actions. The same clause also says that where a customer has prescribed processes, tools or systems for problem solving, the organization uses them unless the customer approves otherwise. That is the standard-level reason you answer in each customer's format. The text is in the IATF 16949 Sanctioned Interpretations (Nov 2025), and you should read the clause itself instead of relying on a summary.

The standard also has separate clauses on nonconformity and corrective action (10.2.1) and on customer complaints and field failure test analysis. Check the current edition and your certification body's guidance for the numbers and wording.

Auditors do look at this area. According to a simpleQuE analysis of IATF nonconformance data, clause 10.2.3 (problem solving) was the most frequent major nonconformance in IATF 16949 audits, at 10.06% of major nonconformances as of November 25, 2024. Clause 10.2.1 was second, at 5.65%. These figures describe what auditors found across the industry. They say nothing about any individual supplier's audit.

For aerospace, SAE AS13000A, "Problem Solving Requirements for Suppliers", defines the process a supplier uses to respond to a customer's request for corrective and preventive action. It was issued by the G-22 Aerospace Engine Supplier Quality (AESQ) Committee, and its SAE listing is the place to start. If you hold AS9100 certification, read the current edition's corrective action clauses and your prime's supplier requirements together. The RCCA approach for aerospace is covered in the aerospace root cause and corrective action article.

A request is easier to answer when you treat the customer's manual as the source of the rules and the standard as the reason those rules exist.

Log the request the hour it arrives

Before anyone starts investigating, create one record for the request. A single record per request, kept by you, protects you from the most common loss in this work: the details sitting in an email thread that only one person can find.

Fields to capture when a request arrives

  • Customer name and the SQE or buyer contact
  • Customer's reference number for the request
  • Your internal reference number
  • Part number, revision and description
  • Quantity rejected and quantity in the customer's inventory, if stated
  • Defect description and photos, as the customer gave them
  • Date and time received, and how it arrived (email, portal, phone)
  • Severity or notification level, if the customer assigns one
  • Clock start, with the rule you applied from the manual
  • Containment due date and plan due date
  • Final response due date
  • Internal owner and backup owner
  • Fee terms from the manual, if any
  • Link or attachment of the manual revision you relied on

This list is deliberately plain. It works in a spreadsheet and works just as well on paper in a binder. What matters is that every request gets the same fields, filled the same way, so you can compare them later.

The first hour also needs three actions that are not about paperwork. Tell the shop floor, so the second-shift lead knows before the next lot ships. Acknowledge receipt to the customer using whatever channel they require. Assign the owner, so no one assumes someone else has it.

Contain first, and write down what you did

Containment protects the customer from receiving more suspect product while you work out the cause. It is not the fix. The distinction matters, because a response that describes only containment tells the customer you have stopped the bleeding but not yet found the wound.

Four terms get confused, so it helps to keep them apart:

Term What it does Example
Containment Stops suspect product from reaching the customer 100% sort of finished goods, quarantine of in-process lots
Correction Fixes the nonconforming parts already made Deburring and re-inspecting a sorted lot
Corrective action Removes the cause so it does not recur Tooling change, process parameter control, revised inspection method
Prevention Extends the fix to similar parts, lines or customers Applying the same change to other dies running similar parts

The detailed method for the D3 step is in the guide to 8D D3 containment action. Whatever customer you are answering, the containment record answers the same questions.

A containment log layout

Keep one log per request. A table like this is enough to start:

Location Quantity checked Quantity suspect Method Checked by Date and time Disposition
Finished goods
Work in process
Raw material
In transit
Customer site
Other customers' stock (same part or process)

The last row is easy to miss. If the same process produces parts for another customer, ask whether that stock is also suspect. Customers care how widely you looked.

Also record how you will tell good product from suspect product going forward: a marking, a tag, a lot-code change, a certification mark on the packaging. The customer's receiving inspection often needs a way to identify the first clean shipment.

Send the initial response when the customer's window asks for it, and say what you know. If root cause is not known yet, say so plainly and give the date the customer can expect it. A truthful interim response with a clear next date is better than a polished guess.

Build the 8D analysis

With containment out, the work shifts to understanding. The eight-discipline (8D) method gives the response a structure customers recognize. Ford developed the 8D process and published it in its 1987 manual "Team Oriented Problem Solving (TOPS)", per Mind Tools, and added a planning step in the mid-1990s. Many customers use some version of it, sometimes with a D0 step for emergency response or preparation.

The eight disciplines in plain terms

  1. D1, team. Name the people who know the process, with an owner and a champion.
  2. D2, problem description. State what is wrong, where, how much, and how it was found, in measurable terms. An is/is-not comparison helps here.
  3. D3, interim containment. What you did to protect the customer, as above.
  4. D4, root cause. Why the problem occurred and why it was not detected.
  5. D5, permanent corrective action. What you will change to remove the cause.
  6. D6, implementation and verification. Proof that the change was made and that it worked.
  7. D7, prevention. What else should change so this does not recur on similar parts, processes or customers.
  8. D8, closure and recognition. Final sign-off, lessons learned and acknowledgment of the team.

Root cause: occurrence and escape

Many responses stall at D4, because the first cause someone names is usually a symptom. Two tools help. The 5 Whys method, which Taiichi Ohno of the Toyota Production System called the basis of Toyota's scientific approach (per Tulip), asks why repeatedly until a cause you can act on appears. The fishbone diagram, also called a cause-and-effect or Ishikawa diagram, sorts possible causes into categories, and ASQ lists it among the seven basic quality tools.

Neither tool tells you the answer. They help a team lay out its reasoning so a reviewer can follow it. For depth on D4, see the guide to 8D D4 root cause analysis.

Customers commonly expect two causes, not one: why the defect occurred and why it escaped to the customer. Here is an example 5 Whys chain with both branches, for the burr scenario from the opening. It is a worked example of the method only, and your own facts will differ.

Occurrence branch

  1. Why were burrs present on the flange edge? The trim die edge was worn past its limit.
  2. Why was it worn past the limit? Die maintenance was done by hit count, and the count was reset at the last changeover.
  3. Why was the count reset? The reset step is on the setup sheet, with no check that the hit count carries over.
  4. Why is there no check? The setup sheet was written before hit-count tracking was added.
  5. Why was the sheet not updated? No one owns the link between setup sheets and the maintenance system.

Escape branch

  1. Why did burred parts reach the customer? The final inspection did not catch them.
  2. Why not? The inspection plan checks flange width, not edge condition.
  3. Why? The edge condition was never a listed characteristic for this part.
  4. Why? The control plan was copied from a similar part where edge condition was not a risk.
  5. Why? There is no step in the control plan review that compares failure modes across similar parts.

Notice that the two branches produce different corrective actions. The occurrence branch points to the setup sheet and the maintenance link. The escape branch points to the inspection plan and the control plan review. Answering only one leaves half the problem open.

Corrective actions that a reviewer can check

A good D5 and D6 give each action an owner, a due date and a status, and attach the evidence of completion. Vague actions such as "retrain operators" invite follow-up questions. Better actions are specific and checkable: "Add hit-count carry-over check to setup sheet rev C, effective date, owner, with signed copy attached."

Verification should show the fix worked, not just that it was done. That might be a run of production with the revised inspection, a measurement study, or a period with no recurrence, whatever your customer's manual accepts. Ask the SQE what evidence they expect when the manual is silent.

Review, sign off and return the response

Before the response leaves the building, someone other than the author reads it. A second reader catches the gaps customers catch: an occurrence cause with no escape cause, a corrective action without an owner, a date that falls after the customer's window, evidence that is referenced but not attached.

A simple pre-release review covers these points:

  1. Does the response answer every field the customer's format requires?
  2. Are the occurrence and escape causes both stated, with the reasoning shown?
  3. Does each corrective action have an owner, a date and evidence?
  4. Does the containment record cover all the locations where suspect product could be?
  5. Do the dates meet the clocks you logged?
  6. Has a person with authority approved it?

Then return it in the way the customer requires. If that is a portal, you re-enter or paste the content into the customer's fields. If it is the customer's form, you transfer your content to it. Keeping your own record first, then re-keying, means your analysis does not depend on the customer's system staying available or on a single engineer's login.

After you send it

The request is not closed when you press submit. Customers may reject a response, ask for more evidence or request an update on implementation. Track each exchange in the record, with dates. Note when the customer accepts the response and when the request formally closes. Some customers also schedule a follow-up check on effectiveness, so write that date into your record too.

Final closure should leave a file that a new engineer could read a year from now and understand: the request, the clocks, the containment, the analysis, the actions, the evidence and the customer's acceptance. Record retention periods are set by the standard and your customers. IATF 16949 clause 7.5.3.2.1 sets minimum retention for certain quality records, and customers may specify their own periods. Check the clause and your customers' CSRs for the exact requirement.

Keep your own record across every customer

If you answer more than a handful of requests a year from different customers, the structure of your record starts to matter. Each customer has its own clock, its own format and its own portal. Your own record is the one place where all of them sit side by side.

Done well, that record lets you see which customers have open requests now, which responses are due this week, and whether the same root cause is turning up on a second customer's request. It also means history stays with the company when an engineer moves on. None of that requires special software. A well-kept spreadsheet with consistent fields can do it, as long as someone maintains the dates and formulas. A spreadsheet will not send a reminder, and it flags an overdue response only if someone has built that in.

Supplier CAR Response Desk is designed to be that record for suppliers who prefer a purpose-built one. It is built to log each request as its own record, run containment and plan clocks from editable per-customer rules, walk the response through D1–D8 with 5 Whys and fishbone entry, and export PDF or Excel in a customer's field order. The Response Desk records and calculates. It does not certify, advise, submit to portals or write the analysis. It is designed to work alongside the portals your customers require, and pricing is one flat price per plant rather than per seat.

A short routine to adopt this month

You can improve how your team handles a supplier corrective action request without changing any tools:

  1. Read the response section of each customer's current manual and CSRs, and complete the checklist above.
  2. Put the answers in one shared document, with revision dates.
  3. Use the same intake fields for every request.
  4. Keep a containment log per request, including other customers' stock.
  5. Require an occurrence cause and an escape cause on every D4.
  6. Add a second reader before anything goes out.
  7. Once a quarter, look across your closed requests for repeated causes, parts or customers.

If you would like to hear when the Response Desk opens, you can join the waitlist.

Put it to work

Compare plans on the pricing page.

Get the next guide

Practical guides for supplier quality teams, delivered straight to your inbox.