Lot Management for Recalls and Traceability

When a recall notice arrives, the clock starts. The question is not whether you can find the affected stock. The question is how fast you can prove where every unit of the suspect lot went, who received it, and what raw material fed it. For most Japanese manufacturers still running spreadsheets and paper logs, that answer is measured in days, not minutes.

Lot tracking solves this. It is the practice of grouping inventory into traceable batches so a single lot can be followed forward to the customers who bought it and backward to the suppliers who provided the inputs. Done well, it turns a recall from a fire drill into a structured, documentable response.

This guide walks through why lot tracking matters for recall readiness, what changes when it lives inside a core business system, and a step by step checklist you can run to test your own recall posture today.

The Problem

Recalls in Japan are not rare events for food, ingredient, and chemical makers. A lab result flags a contaminant. A customer reports a defect. A supplier quietly tells you a raw material batch was off spec. In every case the regulator, your customer, or your insurer will ask the same thing: show us exactly where this lot went.

Paper records make that question expensive. A typical small manufacturer keeps goods received notes in a binder, production logs on the shop floor, and shipment records in a sales spreadsheet. Tracing one lot means a person walking between three desks, cross referencing handwritten numbers against printed invoices, and hoping nothing was transposed. Every hour that trace takes is an hour a defective product stays on a shelf.

The deeper problem is not the trace itself. It is the assumption that the trace is even possible. Most SMEs discover, only when the notice arrives, that their lot numbers were never consistently recorded in the first place. A supplier invoice carries one batch code. The production log uses a shortened version. The finished goods label invents a third. Reconciliation becomes guesswork.

A recall handled badly costs more than the defect. It costs trust. Japanese B2B buyers, especially in food and chemicals, expect documentation within hours, not days. A maker that cannot produce a clean lot genealogy on demand will not keep that account for long.

A Real-World Scenario

Consider a food ingredient maker in Okayama, about 85 staff, producing flavor blends and functional powders for confectionery and beverage clients across western Japan. The company runs a respectable operation, with a small QA team and an experienced production floor.

Last year a customer flagged an off batch in a finished ingredient. The notice landed on a Friday afternoon. The maker knew the finished lot number. What it did not know, immediately, was which raw material lots had gone into that batch, which other finished batches shared those same raw lots, and which of its own customers had received the related shipments.

The team spent the better part of a week tracing paper. They pulled receiving logs for the period, matched raw material lot codes by hand to production batch sheets, then walked those batch sheets forward to packing lists and sales invoices. Several lots required re checking because a code had been written two different ways. By the time the trace was complete, one downstream customer had already shipped finished product to retail.

Nobody was harmed. But the maker absorbed the full cost of a wider recall than was strictly necessary, lost a week of senior management time, and faced a tense review with its largest account. The lesson, as the operations director put it plainly, was that the company could trace a lot. It simply could not trace a lot fast.

That gap, between can trace and can trace fast, is exactly what structured lot tracking closes.

What Changes

When lot tracking lives inside an ERP rather than in binders and spreadsheets, the recall exercise changes in three concrete ways.

First, the lot becomes a first class object on inventory, not a note scribbled on a line item. Each movement of stock, whether a receipt, a transfer between warehouses, a consumption in production, or a shipment to a customer, is recorded against the lot. There is one consistent code from goods in to goods out.

Second, the trace becomes a query, not an investigation. A suspect lot can be traced forward to every customer who received finished goods containing it, and backward to every supplier who provided a raw material lot feeding it, in a single lookup. The Okayama maker would have had its answer in minutes, not a week.

Third, the data is auditable. Every record change in the system carries who did it and when. When a regulator or customer asks how you know your trace is complete, you have a system of record, not a ream of photocopies.

This is the practical core of lot tracking for recalls. It is not pharma grade serialization, where every individual unit carries a unique identifier and a regulatory reporting pathway. It is batch level traceability, which is the right level for food, ingredient, chemical, and light manufacturing. It tells you, on demand, which lots of finished product are affected and where they went.

The surrounding inventory disciplines reinforce it. Inventory is held by location and warehouse, so a lot query also tells you physical placement. Inventory transfer orders move stock between locations with their own paper trail. Shipments carry a delivery method and can link to purchase orders, so the forward trace ends at a real dispatch record, not a guess. Inventory operations distinguish consumption from scrap, treating scrap as a distinct stock-out movement of its own, so destroyed recalled stock is logged as a clear inventory movement rather than lost in an unexplained write off. When you need that scrap to land in a specific general-ledger account, you post a manual journal entry and confirm the treatment with your accountant.

The Steps

A recall readiness check does not require a live incident. Run this checklist against your current setup, and the gaps will surface on their own.

1. List every product that should be lot tracked. Food, beverage, ingredient, chemical, pharmaceutical input, and any finished good with a shelf life or safety implication belong on this list. If you cannot name them, you cannot track them.

2. Confirm that every receipt records a supplier lot or batch code. Open your last 10 purchase receipts. Is there a lot number on each line, captured at goods in, before the stock is moved to storage? Missing codes at receipt are the single most common reason a backward trace fails.

3. Confirm that every shipment records the lot shipped. Open your last 10 sales shipments. Does each line tie to a specific lot, not just a product? A forward trace is only as good as the codes written when goods leave the building.

4. Map your lot numbering convention. Suppliers use their own codes. You may add your own internal code. Decide one rule for how a lot is recorded in your system, and apply it without exception. Two codes for the same physical lot is how a week long trace begins.

5. Define how production links raw material lots to finished lots. For manufacturers, the finished lot is a blend of input lots. Your system must connect them, so tracing a finished lot backward reaches its raw lots, and tracing a raw lot forward reaches every finished lot that used it.

6. Run a mock recall. Pick one lot currently in stock or recently shipped. Give yourself a target of one hour to produce the full trace: which customers received it, which suppliers fed it, and where any remaining stock sits. If you cannot do it in an hour, you have work to do.

7. Assign responsibility for lot data quality. Lot tracking fails when nobody owns it. Name one person accountable for the completeness and accuracy of lot codes at receipt, in production, and at shipment.

8. Review how scrap is handled. When recalled stock is destroyed, the stock-out must be recorded as a clear inventory movement with a clear reason. Confirm your system treats scrap as a distinct operation separate from normal consumption, so the operational impact of a recall is documented. To land that write off in a specific general-ledger account, post a manual journal entry and confirm the treatment with your accountant.

9. Document the recall procedure itself. Once the data is clean, write down who is notified, who authorizes the customer outreach, and how the trace report is produced and shared. The data is only useful if the people and steps around it are clear.

10. Test again in six months. Lots drift. New products appear. Staff turn over. A recall system that worked once can quietly degrade. Treat the mock recall as a recurring drill, not a one time project.

Frequently Asked Questions

Does lot tracking mean full unit level serialization?

No. Lot tracking operates at the batch level, which is the right granularity for food, ingredient, chemical, and light manufacturing. You trace a production batch or supplier lot as a unit, rather than marking every individual item with a unique serial. Serialization, with per unit identifiers and regulatory reporting, is a heavier discipline suited to pharmaceuticals. Most SMEs need batch traceability to answer a recall notice, not serialization.

How fast should a lot trace realistically take?

With lot data captured consistently inside an ERP, a forward and backward trace should resolve in minutes. The target in step six of the checklist, one hour including assembling the report, is achievable for a well maintained system. If your current setup requires days, the bottleneck is almost always inconsistent lot coding at receipt or shipment, not the trace itself.

What does a recall cost beyond the defective stock?

The stock itself is often the smallest cost. The larger costs are management time, lost customer trust, and a wider recall scope than necessary when the trace is imprecise. A maker that can prove a tight lot genealogy can often recall only the affected batch, rather than a broad window of production. Clean lot tracking is a margin protection tool as much as a compliance one.

Key Takeaway

Recall readiness is a data problem before it is a product problem. The maker that can name every lot, follow it to every customer, and tie it back to every supplier in a single query, is the maker that survives a notice with its accounts and margins intact. Paper does not give you that. A core ERP with lot tracking built into inventory does.

See How Kikan System Handles Lots

Kikan System gives a Japanese maker the inventory foundations this guide describes: lot and batch tracking tied to movements, locations, transfers, and shipments, so a recall trace runs forward and backward in one query. Inventory is structured the same way across goods in, production, and dispatch, and every record change is logged for J-SOX internal controls. The system is authored natively in Japanese and English, and each company data is isolated with role based access and passwordless passkey login.

You can run Kikan System with up to 2 users, no credit card required. (-> Start free)(/#get-started)

If you are earlier in your core business system evaluation, two related guides are worth reading: how to choose a cloud ERP core business system, and the case for a bilingual core business system in Japan.