Recall Quantity Reconciliation: Account for Every Unit in a Lot

A customer list is incomplete if the team cannot explain how the lot’s full quantity was distributed or disposed.

Why this matters

Product businesses make connected decisions. A choice that looks small in one spreadsheet can change purchasing, production, inventory, customer promises, and cash. The practical goal is not more paperwork. It is a result the team can explain, repeat, and review before a preventable exception becomes expensive.

A practical example

A lot produced 600 units: 410 shipped, 120 remain, 30 were sampled, 25 were scrapped, and 15 are unexplained. The trace is not complete until the 15-unit gap is resolved.

The example is intentionally simple. Real decisions should use current records, a representative order or batch, and the exact scope being approved. Write down what the calculation includes and excludes so the answer is not reused for a different question later.

Step-by-step method

1. Start with verified good output and every related lot transformation. Document the input, the person responsible, and the evidence produced. Do not mark the step complete merely because a number was entered; make sure another person can understand where it came from and what decision it supports.
2. Total on-hand, held, shipped, sampled, reworked, returned, and disposed quantities. Document the input, the person responsible, and the evidence produced. Do not mark the step complete merely because a number was entered; make sure another person can understand where it came from and what decision it supports.
3. Compare the accounted quantity with the batch record and investigate gaps. Document the input, the person responsible, and the evidence produced. Do not mark the step complete merely because a number was entered; make sure another person can understand where it came from and what decision it supports.
4. Record trace time, percentage reconciled, evidence, and corrective action. Document the input, the person responsible, and the evidence produced. Do not mark the step complete merely because a number was entered; make sure another person can understand where it came from and what decision it supports.

Common mistakes

• Using a perfect scenario. Normal loss, delay, rework, fees, and exceptions disappear, so the answer is optimistic.
• Mixing units or time periods. Per-order, per-unit, weekly, and monthly numbers are compared without conversion.
• Changing several assumptions at once. The team cannot tell which change created the result.
• Keeping only the final number. A result without sources, date, and owner cannot be reviewed confidently.
• Treating a tool as approval. A calculator organizes inputs; responsible people still evaluate safety, compliance, finance, and customer commitments.

What good looks like

A useful result is traceable to current inputs, understandable to someone other than its creator, connected to a real decision, and scheduled for review. The team knows what should happen when the result is outside the expected range. It can compare planned and actual outcomes without rewriting the original plan.

How to interpret the result

Do not reduce the answer to “good” or “bad.” Compare it with the business requirement, the previous result, and the range that normal operations can realistically produce. A result outside the target may point to a data problem, a one-time exception, or a process that needs attention. Confirm which explanation fits before changing a price, standard, supplier, schedule, or customer promise.

Look at dollars or units as well as percentages. A large percentage attached to a tiny order may matter less than a small recurring loss across the highest-volume product. Also check timing: an acceptable annual average can hide a shortage, workload peak, or cash gap in a particular week.

Decision worksheet

Before acting, answer these questions in writing:

1. What exact decision are we making, and when must it be made?
2. Which product, order, batch, supplier, channel, or time period does the answer cover?
3. Which source records were used, and how current are they?
4. What normal loss, delay, fee, or exception is included?
5. Which important factor is still an estimate?
6. What would change the answer enough to choose a different action?
7. Who approves the action, and who needs the result next?
8. When will actual performance be compared with the plan?

Questions people often ask

Should I use an average or the latest result?

Use the measure that fits the decision. A representative average may be useful for normal planning, while the latest confirmed supplier price or a known future fee may be necessary for a new quote. Preserve both the underlying observations and the chosen planning assumption.

How many examples should I review?

Start with enough normal cases to see ordinary variation, then include at least one difficult case. One perfect order or batch proves very little. As volume grows, segment results by the factors most likely to explain a difference rather than relying only on a company-wide average.

When should the calculation be updated?

Set a routine date and earlier triggers. A material price, formula, package, wage, yield, fee, supplier, process, customer term, or regulation change may make the old answer unsuitable before the next scheduled review.

Put it into practice

Use the relevant free Batch Scale tools to work through your own numbers, then connect the decision to the source records inside Batch Scale. Start with one representative product or order. Improve the method from actual results before expanding it across the entire catalog.

Authoritative starting point

Review FDA Food Traceability Final Rule for current official information. Applicability depends on the product, process, claims, location, and facts; this article is educational, not a substitute for qualified advice.