EVIDENCE LOOP / 01
One finding, several different control questions.
Keep six questions separate: what failed, what is on hold, why it happened, what changed, whether the change was made and whether it worked.
Record what failed and the evidence
Name the supplier and site, product or SKU, purchase order and lot or process. Record the specification version, required result and actual result. Add the date, location and source files. Keep the supplier’s response separate from your original finding.
The ISO 9001 Auditing Practices Group library includes public papers on documenting, reviewing and closing nonconformities and on effectiveness. The library states that its guidance is informative, context-dependent and not a specified requirement or universal industry benchmark.
A label such as “bad quality” is not reviewable. Record the requirement, the actual evidence and the exact gap between them.
Find which goods and work are affected
List the affected quantities, lot or serial numbers and production dates. Include stock, work in progress, goods in transit and shipped units. Check related products and record gaps. Use this order’s contract and product risks to decide what needs to stop.
Keep confirmed affected work, potentially affected work and not-yet-assessed work distinct. The request should state which work is on hold, which may continue and who owns the extent review.
Anchor the finding to the approved quality-control baseline →
Contain the problem and correct the affected work
Label and separate affected or suspect goods. State what must stop and who must be told. Record who can approve rework or replacement, who will check it and when it is due.
Correction may repair, rework, replace or otherwise address detected nonconforming work. It does not by itself explain or remove the cause. The official ISO 9001 APG nonconformity review paper distinguishes correction from corrective action and describes objective evidence across correction, cause analysis, implementation and effectiveness before closure. It remains informative guidance, not a product-specific acceptance rule.
Test why it happened and why checks missed it
Ask why the failure happened and why checks missed it. Test each proposed cause against the evidence. Keep rejected explanations and open questions in the record. Check whether the same cause could affect other products or processes.
Ask for evidence behind “operator error” or “carelessness”. The APG paper asks reviewers to look beyond the first failure factor. Check whether the cause is a one-off event or a wider system problem. A 5 Whys or fishbone diagram can organise the work; the filled form does not prove the cause.
Approve actions tied to each supported cause
For each supported cause, name an action, owner and due date. List the documents, tools, materials, training and checks that must change. Include other affected sites or products. State how the action will prevent or detect a repeat, and what will prove it was carried out.
The U.S. Department of Energy’s public Project Management Lexicon describes a corrective action plan as documenting assumptions, constraints, responsibility, commitment dates, the action plan, verification steps and completion documentation. This is useful field-level government guidance, not a sourcing contract or a requirement imposed on every supplier.
ISO 10007:2017 provides lifecycle configuration-management guidance. Use controlled revisions and status records when the approved action changes a material, component, drawing, method, tooling, process, site or critical sub-supplier.
Check that the agreed action was carried out
Check when the action took effect and which lots it covers. Review the new document versions, staff training, tool settings and inspection records. Tie each record to the changed work. Log any late, partial or different action for review.
If an inspection is needed, name the goods or process to check, method, sample, acceptance criteria and reviewer. A photo shows only what was visible at that moment. It cannot prove that the fix covers all affected work or remains in place.
Use in-process evidence for the next continue, correct, reinspect or hold decision →
Keep final-lot inspection and shipment release as a separate decision →
Define and run an effectiveness check
Agree how to test the action before approving it. Name the lots or time period, what you will measure and what result is acceptable. Set an owner and review date. Include related complaints and the result that would trigger a wider review.
Implementation means the change was made. Effectiveness means it addressed the cause during the agreed review period. Keep the request open if there has not been enough production or time to test this.
What public records cannot prove
A penalty or historical anomaly entry does not prove that a supplier’s corrective action worked. Review your own finding, cause analysis, implementation and effectiveness evidence.
Our 21–22 August 2026 queries covered 45 company candidates from an NHTSA manufacturer list. These are provider-returned registry records, not quality audits.
| Record | Companies | Items where present |
|---|---|---|
| Administrative penalties | 9 of 45 | Seven had exactly one |
| Operating anomalies, current | 0 of 45 | — |
| Operating anomalies, historical | 3 of 45 | Historical entry returned |
Three candidates had historical anomaly entries; none returned a current anomaly entry in this query. The archive does not establish why an entry changed or whether a product defect was resolved.
Ask the supplier for evidence tied to the affected lot and agreed review period. Neither an empty public-record result nor a claimed repair closes your request.
The manufacturer study explains the selected cohort and lookup limits. It is not a representative supplier sample or a measure of corrective-action success.
Record the buyer’s decision to close or reopen
The authorised buyer reviewer records the decision and date. Link the finding, containment, correction, cause analysis and action taken. Include the effectiveness result, remaining risks and open issues. Name the person who signed off.
Close only the defined request and scope. Reopen or create a linked request when the same condition recurs or the action is not implemented as approved. Do the same when effectiveness criteria fail, the affected extent expands or new evidence changes the cause analysis. Closing a SCAR does not automatically release shipment, final payment, contractual acceptance or destination-market compliance.
REQUEST RECORD / 02
What to put in the request
| Control field | Supplier response must identify | Buyer review asks | Keep open when |
|---|---|---|---|
| Finding | Requirement, actual result, evidence, product/order/process and revision | Is the gap objective, traceable and understood without rewriting the original finding? | The requirement or evidence is ambiguous or disputed without a visible record |
| Affected extent | Confirmed, potential and unassessed quantities, locations, lots and related work | Does the scope cover inventory, work in process, transit and similar conditions where relevant? | Affected or suspect work cannot be bounded |
| Containment | Identification, segregation, stop/hold scope, notifications, owner and timing | Can uncontrolled work still move, mix, ship or reach a later stage? | Immediate exposure remains uncontrolled |
| Correction | Disposition, rework/replacement instruction, affected-unit identity and verification | Were corrected units checked against the defined requirement? | Corrected work is untraceable or unverified |
| Cause analysis | Occurrence and escape causes, considered factors, supporting evidence and extent | Does evidence support the cause beyond a label, tool or first failure factor? | The cause is assumed, incomplete or contradicted |
| Corrective action | Action per cause, owner, due date, changed controls and affected configurations | Does each action address recurrence or detection without creating an uncontrolled change? | Actions are generic, ownerless or disconnected from supported causes |
| Implementation | Effective boundary, approved revisions, records, training and verification results | Was the approved action implemented across the defined scope? | Evidence shows only an intention, sample or partial implementation |
| Effectiveness and closure | Evidence window, criteria, results, related recurrence review and accountable sign-off | Did the action work, and does the buyer own the close-or-reopen decision? | Effectiveness is untested, failed or awaiting enough evidence |
DECISION STATES / 03
Keep each request in one explicit state
| State | Minimum record | Permitted next action |
|---|---|---|
| Draft | Proposed finding, requirement, evidence and scope owner | Verify and issue the controlled request |
| Issued | Request ID/version, supplier, finding, due dates and buyer authority | Begin the required response and immediate controls |
| Containment open | Affected extent and immediate controls remain incomplete or unverified | Control exposure before relying on later analysis |
| Cause unsupported | Assumptions, missing evidence or contradictory factors are visible | Continue analysis; do not approve permanent action as cause-based |
| Action approved | Supported causes, approved actions, owners, dates and configuration impact | Implement only the approved plan and record exceptions |
| Implementation evidence due | Action is scheduled or claimed complete, but defined proof is outstanding | Collect and review implementation evidence |
| Effectiveness pending | Implementation accepted; evidence window or recurrence review remains open | Monitor the defined population, period or lots |
| Closed | All defined evidence, effectiveness result and authorised buyer decision recorded | Archive the request; maintain linked controls and separate release decisions |
| Reopened | Recurrence, failed criteria, expanded scope or changed cause evidence recorded | Re-control exposure and revise the linked analysis or action |
METHOD / 04
How this checklist was prepared
Currawong’s China-side verification desk reviewed the official and first-party sources below on 30 July 2026. The ISO APG library and paper support the distinctions among nonconformity, correction, corrective action, implementation and effectiveness. The DOE lexicon supports accountable action-plan fields. ISO 10007 supports controlled change records. Graco’s supplier page is used only as a real buyer-specific example of an 8D response workflow.
The eight-control workflow, request table and decision states are our evidence-control synthesis for global sourcing. We did not review any reader’s supplier, factory, product, process, lot, finding, contract, sampling plan or destination-market requirements. This is general evidence-organising guidance, not an imposed supplier requirement, investigation, inspection instruction, certification, legal advice, supplier approval, shipment release or quality guarantee.
PRIMARY SOURCES / 05
Official and first-party sources used for this guide
Links and page content were checked on 30 July 2026. Re-open the buyer’s contract, supplier manual, applicable standards and product or destination rules before a material decision because requirements and records can change.
- ISO 9001 Auditing Practices Group library Official ISO committee page linking nonconformity and effectiveness papers; the page explicitly limits the papers to informative, context-dependent guidance instead of specified requirements.
- ISO 9001 Auditing Practices Group: Review of nonconformity Official informative paper used for correction, corrective action, cause analysis, implementation, effectiveness and closure concepts; not treated as a product rule or universal buyer requirement.
- U.S. Department of Energy: Project Management Lexicon Official public lexicon used for corrective-action-plan fields including responsibility, commitment dates, verification steps and completion documentation.
- Graco: Supplier Corrective Action Request First-party supplier requirement showing one buyer’s 8D workflow, containment, cause verification and supplier-quality review. It is an operational example, not a universal sourcing standard.
- ISO 10007:2017 Official ISO record for configuration-management guidance across a product or service lifecycle; ISO lists the edition as current and confirmed in 2023.
COMMON QUESTIONS / 06
Questions about closing a request
- Is a correction enough to close a SCAR?
- Not if the request also requires you to remove the cause and prevent a repeat. Check the cause, the action taken and whether it worked. Repairing the affected goods is only one part.
- Must every supplier use an 8D report?
- No. 8D is one response format, not a universal requirement for every buyer. Use the format required by your contract or supplier manual. Keep the evidence and buyer decision for this finding.
- Is “operator error” a root cause?
- Treat it as a claim to test. Ask what allowed the error: the method, training, tools, materials, workload or checks. Look for evidence that supports the cause.
- Can photos close the request?
- Photos may show the defect or work done to fix it. They do not prove the cause, the full affected scope or that the action worked over the agreed period.
- When should a closed request be reopened?
- Reopen or create a linked request when the condition recurs or the approved action was not implemented. Do the same when effectiveness criteria fail, affected scope expands or new evidence changes the cause analysis.
- Does closing a SCAR release payment or shipment?
- No. Keep shipment, payment and product acceptance decisions separate unless your governing records explicitly link them to closure. Check who can approve each next action and what evidence they need.