One jar leaks. One candle wick sits off-center. One pouch seal opens too easily. The uncomfortable question is immediate: Do you hold that unit, the rest of the case, everything packed since lunch, or the entire batch?
Good defect containment means drawing the smallest affected-inventory boundary the evidence can support, then keeping it from moving until someone makes a clear decision. Product-specific rules may require a broader hold or defined sampling plan.
First, stop movement and preserve the evidence
Move the failed unit to a clearly identified hold area, preserve its batch or run identity, and record when and where it was found. If work is still running, pause it and block nearby inventory from sale, picking, or shipment while facts are available.
Record the observable defect without guessing at the cause. “Cap loose; product visible around neck” is evidence. “Operator forgot to tighten cap” is a theory until the process is checked.
Find the possible start and end of the defect
Treat production as a sequence rather than one undivided batch. Ask:
- What is the earliest point when this defect could have started?
- What is the latest point when it could have stopped?
Use timestamps, material lots, equipment settings, operator changes, adjustments, packaging cases, and in-process quality checks to locate those points.
Suppose a leaking bottle is found in case 7. The closure tool was checked before case 1, adjusted after case 4, and checked again after case 8. Cases 5 through 8 may be the first defensible boundary. If nobody recorded the adjustment or case sequence, the business may have to hold more inventory because it cannot show where the risk begins or ends.
Dependable production records help locate the affected window and reduce the inventory that must be treated as unknown.
Test which defect pattern fits the evidence
Isolated defect
An isolated defect has no reasonable shared cause with surrounding units. A chipped jar may fit if adjacent units and source packaging show no similar damage. Check the nearby sequence and shared inputs before narrowing the hold.
Sequence-based defect
A sequence-based defect affects a time window: labels drift after a guide loosens, fill weights change as a hopper empties, or seals weaken after an adjustment. The boundary may run from the last documented good check to the first verified good check after correction.
Common-cause defect
A common-cause defect can reach several batches because the same material, formula version, component, tool, or method applies throughout. The boundary should follow the cause, not a convenient batch label.
Build an affected-inventory boundary
Use a short worksheet rather than relying on a tense conversation at the workbench:
| Boundary question | Evidence to review | Possible hold scope |
|---|---|---|
| Which units share the same material or packaging lot? | Receiving record, lot use, case or roll changes | Units made with that lot |
| When was the last verified good check? | Time-stamped check, first-piece approval, retained sample | Units after that check |
| Did equipment, setup, or method change? | Settings, adjustment log, cleaning, changeover notes | Units since the change |
| Is the defect visible, measurable, or hidden? | Inspection method and detection limit | Sample, sequence, or full shared-cause group |
| Can affected units be identified physically? | Case IDs, pallet locations, batch labels, timestamps | Identifiable subset or broader hold |
Choose the boundary from recorded evidence. If the trail breaks, expand the hold to the next point you can defend.
Use a red-yellow-green hold decision
Red: hold the full shared-cause group. Use this when the defect may involve product identity, safety, required labeling, an incorrect formula or component, or a condition that cannot be reliably detected after the fact. Follow the applicable product rules before release.
Yellow: hold a defined sequence or location. Use this when records identify a time window, equipment setting, packaging case, or process stage that could contain the defect. Keep the boundary visible while the team inspects or evaluates it.
Green: isolate the unit and verify nearby units. Use this only when evidence supports a truly isolated event, the inspection method can detect the same problem, and surrounding units pass the written check.
Every status still needs the product, quantity, location, reason, date, owner, and next action.
Inspect deliberately instead of “checking a few”
Before opening cases, decide what will be checked, how units will be selected, what counts as a failure, and what result will expand the hold. Otherwise, the inspection can quietly stop as soon as the team feels reassured.
The quantity depends on the product, defect, history, detectability, and applicable rules. A cosmetic scuff, incorrect allergen statement, leaking closure, and unstable formula do not deserve the same response.
Record both passes and failures. “Looked fine” does not show which units were reviewed or what standard applied. The guide to documenting simple quality checks can help define the next check.
Make one of four clear disposition decisions
Every held group should end with a named decision:
- Release: evidence shows the defined units meet the requirement.
- Rework: an approved correction can bring identified units to the requirement, followed by verification.
- Reject: the units cannot or should not be made acceptable.
- Continue the hold: more evidence, testing, customer direction, or qualified review is required.
Record the decision owner, evidence, quantity, and inventory destination. Reconcile every disposition with the original hold total.
After the immediate decision, calculate what the event consumed. The guide to calculating the true cost of rework helps capture materials, labor, packaging, delay, and missed contribution without confusing containment with blame.
Verify the change on the next run
Correction handles today’s inventory. Prevention changes what happens next time.
Update the instruction, fixture, setting, supplier requirement, inspection point, or training after the likely cause has evidence behind it. Add a temporary check to the next run. The resource on using production records to improve batch consistency explains how to compare several runs.
Frequently asked questions
Does one failed unit mean the whole batch is bad?
No. It means the affected boundary is not yet known. Hold enough inventory to protect the decision, then use shared materials, time, equipment, process changes, and inspection evidence to narrow or expand it.
How should held inventory be identified?
Use a visible status with product, batch or run ID, quantity, location, reason, date, owner, and next action. Prevent held units from appearing available to fulfillment or sales.
Can you release part of a batch?
Sometimes, if records and inspection evidence clearly separate acceptable units from the affected group and applicable rules allow it. Document the boundary, decision, quantities, and approver.
How many units should you inspect after finding a defect?
There is no universal number. Set the method from the defect’s risk, detectability, process history, affected boundary, and product-specific requirements before inspection begins.
What belongs in a defect-containment record?
Record the failed result, where and when it was found, initial hold scope, affected-boundary evidence, units inspected, results, final disposition, quantities, decision owner, corrective action, and next-run verification.
Practical takeaway
When one unit fails, freeze movement before the evidence disappears. Find the last verified good point, identify shared causes, draw the smallest defensible boundary, and give every held unit a clear status and owner.
That discipline protects product consistency and customer trust without turning every defect into an automatic full-batch loss. Use the worksheet on the next real exception, then strengthen the records and checks that made the decision difficult.
