Retest workflow after ca proposition 65 compliance failure: part 4

Key takeawaysA passing number alone is not verification; confirm the changed input is in the goods and the sample represents the corrected lot. · Structure the final file as one chain — failure, containment, cause, correction, resampling, passing retest — with matching lot identity throughout. · Most failed retests trace to unchanged conditions, unrepresentative resampling, or mismatched components rather than laboratory error. · Acceptance sampling logic governs retests: only a representative draw lets the second report speak for the lot.^[1] · Proposition 65 sits within wider technical-regulation disciplines, whose trade framework helps explain why evidence standards keep rising.^[2]

A passing retest report feels like the end of the story, but experienced quality teams treat it as the second-to-last chapter. The final chapter is verification in the full sense: confirming that the corrective action genuinely took effect in the shipped goods, that the evidence chain is complete enough to defend the lot months later, and that the program itself learned something worth keeping.

Skipping this stage is how organizations collect passing reports while repeating failures. The material gets swapped back quietly, the subcontractor drifts again, the testing rotation never updates, and next season's failure arrives with depressing familiarity. Verification closes the loop between the incident and the system.

This fourth part of the retest-workflow series covers that closing stage: what verification must confirm, how to structure the final file, why retests still fail, and how to convert each closed failure into a permanent program improvement.

Verify the fix, not just the number

When the passing report arrives, the verification routine has three questions. First, is the corrective action actually present in the tested goods — does batch traceability show the replacement material in the resampled lot, do certificates match the consumed batch, do production records cover the right dates? Second, did everything else stay constant, so the improvement is attributable to the fix rather than to an unrecorded simultaneous change? Third, does the resampling record show a representative draw across the corrected lot rather than a showcase selection? California Proposition 65 testing

Each question needs a document, not an assurance. Batch consumption records, incoming certificates keyed to batch numbers, production date logs, and the resampling record with carton coverage together prove that the passing number describes corrected goods. Where any link is missing, say so in the file and close the gap before release — a reviewer will find the hole eventually, and finding it yourself first is always cheaper.

This is also the moment to reconcile quantities. Confirm that the corrected lot's size matches what was contained, reworked, or rebuilt, and that no uncorrected units leaked into the shipping quantity. Count discrepancies at this stage signal the most dangerous failure mode of all: good paperwork describing goods that are not the ones in the container. Professional California Proposition 65 testing programs treat this reconciliation as mandatory, not optional.

Assemble the chain that defends the lot

The closed file should read as a single narrative in six documents: the original failed report, the containment record with hold photographs, the traceability and root-cause note, the corrective-action evidence, the resampling record with custody confirmations, and the passing retest report with matching lot identity. A one-paragraph closure note on top states the conclusion — lot released or rejected — and the basis in one breath.

Check identity alignment across every page before filing. Lot numbers, component descriptions, substance names, units, and applied limits must agree throughout; any divergence needs a written explanation attached at the point of divergence. Identity mismatches are the first thing auditors and marketplace reviewers probe, because a chain whose links describe different goods proves nothing.

Store the chain where the next incident team can find it, keyed to product family and supplier. Closed failures are training material and risk intelligence: they show new staff what rigor looks like, and they show planners which suppliers and materials deserve tightened controls. Files buried in personal inboxes might as well not exist.

Why retests still fail

The sampling discipline behind credible retests

Three rows of that table share one underlying cause: unrepresentative sampling. The logic is the classical acceptance sampling insight — inferences about a lot stand or fall on whether the sample represents it — applied to the emotionally charged moment after a failure, when everyone wants a quick clean number.^[1] Pressure to retest fast works directly against the breadth that makes retests believable, and managing that tension is a quality manager's core job at this stage.

Practical defenses help. Fix the resampling plan in writing before the draw, require carton coverage proportional to lot size, put an independent witness at the table, and photograph sealed samples against the sample list. Each measure costs little; together they make cherry-picking practically impossible and demonstrably absent. A retest conducted under these conditions may still fail — and that failure is valuable data — but it will never be dismissed as theater.

Keep retained samples from both rounds until the buyer formally accepts the lot. Retains are the physical anchor of the whole chain: if sample identity is ever disputed, re-examination settles it. Programs that discard retains to save storage space discover their value at the worst possible moment.

Convert the incident into system improvement

1. Lock the correctionName the approved input or process that replaces the failed one, lock it in purchasing documents, and require sign-off for future changes.
2. Update the testing rotationAdd the failed material-substance pairing to the testing rotation at an intensity matching its history.
3. Extend control boundariesIf the trail revealed an unknown subcontractor or process, bring it inside the quality agreement with defined requirements.
4. Retier the product familyAdjust the product family's risk tier and verification rule so the calendar reflects the newly learned risk.
5. Close the loop with the supplierBrief the supplier's management on the cause and the new controls, and record their acknowledgment with the file.

The wider regime behind the paperwork

It is worth remembering why evidence standards for chemical compliance keep rising. Product requirements function as technical regulations in trade-law terms, and the international framework balancing legitimate health protection against unnecessary trade obstacles expects conformity assessment — testing, inspection, documentation — to demonstrate compliance robustly.^[2] Each enforcement wave, in California or elsewhere, pushes buyer expectations one notch higher, and files that satisfied last year's review may not satisfy the next.

Suppliers sometimes experience this as ratcheting pressure; better to present it as accumulated protection. Every complete chain on file shortens the next dispute, every retiered product focuses the next budget, and every supplier that learns the system becomes a more reliable partner. Verification is where incidents stop being costs and start being assets — provided the file is complete, the fix is real, and the system actually changed.

Frequently asked questions