A corrected lot should ship only after a defined re-inspection confirms the fix in both the goods and the records. The visit needs a written scope naming failed attributes, corrected quantities, and pass criteria before anyone travels. On site, identity comes first, then attribute checks with original methods, then proof that process controls changed. The lot releases when a passing retest, a confirming visit, and a written decision with rationale all stand together.
Key takeawaysRe-inspection needs a written scope naming the failure, the correction, the quantities, and the pass criteria. · On-site verification confirms stock identity first, then attributes, then the records behind the fix. · Lasting correction shows in process controls, documented inputs, and segregated nonconforming output.[1] · Retest reports carry weight when samples trace to the corrected batch and methods sit inside accredited scope.[2] · Release requires a passing retest, a confirming visit, and a written decision together.
The first part of this retest series covered quarantine, root cause analysis, and correction. This second part covers the other half of recovery: scoping the re-inspection, verifying the fix on site and in the laboratory, and making a defensible shipment decision. Re-verification must be as disciplined as the original check, because a casual revisit that assumes the fix worked proves nothing.
Vague scope is where re-inspection most often fails. A booking that says re-check the goods invites a general visit that may or may not examine the corrected attribute. A booking that names failed attributes, corrected quantities, acceptance criteria, and required records produces an answer the buyer can act on the same day.
This article gives buyers a working method for that second half, from the one-page scope through the release decision. Quality system thinking explains why records and live controls matter as much as goods, and laboratory recognition explains why some reports travel while others do not. Together they turn a failure episode into evidence of diligence.
Scope the revisit before anyone travels
A re-inspection without a written scope is a second chance to disagree about what was supposed to be checked. The scope document should fit on one page and name the original failure with its report reference, the correction the factory says it made, the exact quantities and locations of corrected stock versus untouched stock, the attributes to verify with their pass criteria, and the records the inspector is entitled to review. Every line removes an ambiguity that could dilute the visit, and the factory should countersign the scope before anyone travels.
See the inspection scope and check that the failed attributes, the corrected quantities, and the acceptance criteria are all named explicitly before the visit is scheduled. Attribute selection should follow the correction footprint: verify corrected attributes in full across every affected carton, apply risk-based checks to neighboring attributes that share materials or processes, and leave previously passed untouched attributes alone. This focus keeps the visit affordable while preventing the familiar embarrassment of fixing one defect and introducing another during rework.
Confirm goods and records on site
Verification starts with identity, not with testing. Batch tags, material labels, and artwork revision marks should trace the presented goods back to the rework record, and presented quantities should reconcile with the logged corrected population. When the numbers do not match, the visit stops, because examining the wrong stock wastes the day and invalidates any samples drawn from it.
Product checks then repeat the original methods so results stay comparable: markings against approved artwork at actual size, dimensions and materials against specification, quantities against the packing list. Alongside the goods, the inspector reviews whether process controls changed, since lasting correction means documented inputs, monitored steps, and segregated nonconforming output rather than a hurried sort of finished cartons.[1] A report that covers both layers, product and system, tells the buyer what actually improved.
People evidence matters as much as paper. Operators should be able to describe the revised instruction in their own words, and the segregation area should be visibly in use rather than freshly painted for the visit. Ask for a live demonstration of the new incoming check instead of accepting a meeting-room description. Controls that cannot be shown operating are claims, and claims do not prevent repeat failures.
| Verification item | Evidence reviewed | Release signal |
|---|---|---|
| Stock identity | Tags, batches and quantities against the rework log | Presented goods match the corrected population |
| Corrected attributes | Same methods as the original inspection | Results within criteria across the sample |
| Neighboring attributes | Risk-based checks on shared materials | No new defects introduced by rework |
| Correction records | Rework log, incoming records, instructions | Documents trace the cause to the fix |
| Retest linkage | Sample identity matched to the report | Report tied to the corrected batch |
Read the retest report like an auditor
A retest report answers one narrow question: does the corrected material meet the failed requirement. Read it through that lens by checking sample identity against the corrected batch, the method against the requirement, and the cited limit against the current rule version. A favorable result on the wrong sample or the wrong method is not good news; it is a new problem wearing a passing grade.
The weight of the report depends on the competence behind it. Results issued within recognized accreditation frameworks travel across borders because customers and authorities can verify the method, scope, and signatory behind the numbers.[2] Keep the accreditation reference with the report and confirm the laboratory scope covers the retest method. For high-stakes substance failures this single check prevents later disputes about whether the retest counts.
Timing and linkage complete the reading. Retest samples must postdate the correction, come from the corrected population, and carry batch identifiers that match both the sampling record and the rework log. Gaps in this chain, such as samples drawn before rework finished or batch numbers that match nothing on file, undermine the report whatever its numbers say. Treat a report with broken linkage as pending, not passing, until the linkage is repaired.
Release rule of three: a passing retest on the failed attributes, an on-site visit confirming the corrected stock and its records, and a written decision with rationale in the file. Ship only when all three stand; storage fees are the cheapest line in any failure budget.
Sources
- Wikipedia: ISO 9000 —— Overview of the ISO 9000 family of quality management standards.
- ILAC: International Laboratory Accreditation Cooperation —— ILAC homepage describing the international cooperation on laboratory and inspection body accreditation.
Frequently asked questions
What belongs in a re-inspection scope?
Name the original failure with its report reference, the claimed correction, the corrected quantities and locations, the attributes to verify with pass criteria, and the records the inspector may review. Have the factory countersign the scope so both sides expect the same visit. A scope that fits on one page and names all five elements produces a decision; anything vaguer produces another round of questions.
How do quality system ideas change the visit?
They shift attention from goods alone to the controls that produced them: documented incoming checks, segregation of nonconforming output, and updated work instructions with trained operators. The inspector reviews records and watches the new control operate live. Goods can be sorted in an afternoon, but only system fixes stop the next lot from failing the same way.
What makes a retest report reliable?
Three things together: samples drawn after correction from the corrected population with batch traceability, a method that matches the failed requirement with the current limit cited, and issuance within the laboratory accredited scope so the result is verifiable across borders. Missing any one of the three leaves the report open to challenge. File the sampling record beside the report so the chain from lot to number is complete.
When can the lot ship after a failure?
Ship when the retest passes the failed attributes, the visit confirms the corrected stock and records, and the file holds a written decision with its rationale. Low-risk marking defects with full-lot correction and updated artwork files can sometimes move under defined safeguards. Substance failures should clear completely before loading, since a finding after shipment costs far more than days of delay.