A label reconciliation at the end of packaging does not balance. How would you investigate?

Short answer

Hold the batch, then account for every label issued, used, destroyed and returned. A shortfall can mean a label is on the wrong product or loose somewhere, so search the line and nearby product before accepting a counting error, and record the investigation before release.

Why

US regulation requires procedures to reconcile labeling issued, used and returned, and discrepancies outside narrow limits based on historical data must be investigated as a production record discrepancy. EU Chapter 5 likewise says any significant or unusual reconciliation discrepancy must be investigated and accounted for before release. The reason is that a missing label is a potential mislabeled unit.

If 10 labels are unaccounted for, they might be in a reject bin, stuck inside the labeler, or on containers of another product packed nearby. The investigation starts with arithmetic: check counter readings, the number issued, the number applied according to the line counter, rejects, samples, damaged labels destroyed and labels returned. Then look physically, because a recount on paper cannot find a label on the wrong bottle.

Search the line, the area and the previous and next batches if they were adjacent. A surplus is also a signal, since it can mean units left the line without a label. Note that reconciliation is waived for cut or roll labeling only when 100 percent electronic or electromechanical verification of correct labeling is performed.

Reconciliation works only if the limits are tight and based on real data, and only if counts are captured at the time rather than estimated at the end.

How to work it through

Quarantine the batch and stop further use of the label stock.
Recount issued, applied, rejected, sampled, destroyed and returned labels from the original records and counters.
Search the line, label magazine, reject bins and the surrounding area for loose labels.
Check product packed on adjacent lines or immediately before and after for wrong labels.
Decide whether the gap is explained by evidence or remains open, and assess the risk to every lot that could be affected.
Record conclusions and actions under the deviation before any release decision.

An example

A run issues 50,000 labels. The line counter shows 48,620 applied, 1,180 are returned, 150 were destroyed as damaged and 20 were taken as samples, leaving 30 unaccounted for against a historical limit of 10. A recount of the returned roll finds 1,195, closing 15 of the gap.

A search of the labeler finds 9 labels stuck behind a guide. The last 6 cannot be found, so the team inspects 100 percent of a different product packed on the next line that shift and finds none, then documents the residual 6 as unexplained with the risk assessment supporting release.

Where the answer stops holding

A balanced reconciliation does not prove no mix-up occurred, because a wrong label and a missing label can cancel out. Reconciliation is less useful when labels are printed on demand, where verification of the printed content matters more. Searches can never fully exclude a label on distributed product, which is why vision systems and code readers are stronger controls.

A strong answer covers

Quarantine before investigation
Account for issued, used, destroyed, returned and sampled labels
Physical search, not just recounting
Check adjacent products for wrong labels
Investigate surpluses as well as shortages

Common mistakes

Writing off a small label shortfall as a counting error without a physical search.
Ignoring a surplus because it does not look like a mislabel.
Releasing the batch before the discrepancy has been explained or risk assessed.

The interviewer might ask next

How would you set the reconciliation limit for a new labeling line with no history?