Lot and Batch Traceability for Recalls and Quality Control

It is 9:42 on a Monday morning. A hospital pharmacy in Osaka calls your quality desk. A vial from lot 2403-B-117 caused a reaction. They need to know, in the next hour, every shipment that left your warehouse carrying that lot. Where it went. How many units. When.

If your team opens a spreadsheet and starts scrolling, you have already lost. Health authorities want a complete chain of custody, fast. Customers want confidence. Your board wants the cost contained. The difference between a controlled event and a public crisis is whether your records answer one question instantly: which lots, in which locations, tied to which orders.

This is the moment a core business system earns its keep. Not in the demo. Not in the brochure. On the morning something goes wrong.

Why traceability is a Japan boardroom issue right now

Japan manufacturers face a traceability squeeze from several directions at once. The 2025 legacy cliff has forced companies off systems vendors stopped supporting, and the replacements must do more than the old ones ever did. The qualified-invoice system and the broader digital transformation push mean finance, inventory, and sales now share one ledger of truth. And consumer trust, hard won and quickly lost, depends on a company being able to prove where a product came from and where it went.

The recall math is unforgiving. Marsh reports that average damages for a single product recall exceed roughly USD 1.5 million, and routinely climb far higher, while general liability insurance limits frequently fall short of recall costs. The gap between what a recall costs and what insurance covers often lands on the balance sheet. Speed shrinks that gap. A recall resolved in hours, against the right lot only, costs a fraction of one that drags across days and sweeps up unaffected stock.

Japanese regulators have also tightened expectations. MAFF publishes an official handbook for introducing food traceability systems, and under the Food Sanitation Act, HACCP is now mandatory for food manufacturers. The expectation is no longer "keep good records." It is "produce them on demand, lot by lot, with proof." A core business system that models lots natively turns that demand from a fire drill into a query.

What lot traceability actually means inside the system

Traceability is not a label you add. It is a data structure. In a properly built ERP, every quantity of a lot-managed product is tied to a lot record that carries its own identity, lifecycle, and dates. Let us be concrete about what that record holds, based on what is built.

Each lot carries a unique lot number within the company, an expiry date and a manufactured or received date, and a lifecycle status. The status is not decorative. It controls pickability. An ACTIVE lot can be picked and shipped. A QUARANTINED lot is on quality hold. A BLOCKED lot cannot move. An EXPIRED lot is past its date and frozen. When your warehouse team picks stock, the system can drive First-Expired-First-Out, so the lot closest to its expiry leaves first. That alone reduces write-offs and the quiet margin erosion that comes from stock dying on the shelf.

Lot numbers do not depend on someone remembering a format. The system auto-generates them when staff omit one, following a readable pattern of product code, date, and sequence, with an optional prefix per product. The number must be unique within the company, and the system rejects duplicates. That matters more than it sounds. Duplicate or sloppy lot numbers are the single most common reason a recall expands beyond its true scope, because no one can tell which physical batch a record refers to.

Each lot also carries a free-form notes field. Quality findings, a supplier reference, an inspector initials, a nonconformance ticket number. When an auditor asks why lot 117 was released, the answer lives on the lot, not in someone's inbox.

The recall scenario, played out in software

Picture a precision components and reagents maker in Higashi-Osaka, about seventy staff, supplying parts and test kits to medical device assemblers across Kansai and into Tokyo. They run roughly 1,800 active SKUs, of which about 600 are lot-managed because they carry expiry or carry a regulated material.

A customer flags a defective unit. The quality lead opens the lot record in their core business system, types the lot number, and in seconds sees the manufactured date, the expiry, the status, and the notes. They quarantine the lot with one status change. From that moment, the warehouse cannot ship a single unit of it. The error surface for human follow-through shrinks to near zero.

Now the harder question: where did it go? Because inventory balances reference the lot, the team can trace which warehouse bins still hold stock, and which shipments carried the lot out the door. They pull the affected sales orders, contact the right customers, and arrange returns. They do not recall lot 118. They do not recall the whole product line. They recall exactly the units tied to 117.

This is the gap between a targeted recall and a blanket one. A blanket recall pulls good stock off shelves, infuriates customers who did nothing wrong, and destroys revenue that was never at risk. A targeted recall, enabled by lot-level records, costs a fraction as much and preserves the relationships that took years to build.

When the health authority calls, the team exports the lot data to CSV, lot number, product, dates, status, filtered by date range or status, and sends a clean, complete record. No rekeying. No reconciliation. The audit trail is the system.

Why built-in controls beat a spreadsheet

Spreadsheets die in exactly the conditions where traceability matters. Here is what a core business system does that a spreadsheet cannot.

It blocks bad data at entry. If a product is flagged as requiring lot management, the system will not let staff create inventory for it without a lot. If a product requires expiry management, the expiry date is mandatory, and the system rejects an expiry that lands before the manufactured date or sits in the past. These are not suggestions. They are hard stops. A spreadsheet lets you type anything. The system refuses the wrong thing.

It protects against silent overwrite. Every lot record carries a version number, and updates are checked against it using optimistic locking. If two people edit the same lot, the second save is rejected with a clear conflict. You always know which change won, and when. In a recall investigation, where the integrity of a record can decide a lawsuit, that matters.

It stops destructive cleanup. A lot that still has on-hand stock cannot be deleted. The system checks the balances first and refuses. You cannot accidentally erase the evidence trail for stock that is still in play.

It keeps each company's data fully isolated. A multi-company group running shared infrastructure never sees one company's lots bleed into another's. The traceability you query is the traceability for your operation, and nothing else.

It speaks the company's timezone. Manufactured and expiry dates resolve in the Japanese company timezone, not a server default. When an auditor asks when lot 117 was made, the answer is the local date your staff lived, not a UTC artifact.

The honest limits, and why they still leave you ahead

A trustworthy system tells you what it does not do yet. Honesty sells, and it is the only thing that holds up under scrutiny.

Lot records are not yet an immutable audit-log table. What you have instead is the approval-workflow history and the version-controlled lot record, which together capture who changed what and when. For most recall and quality-control purposes that is sufficient, and it is far better than a paper trail. If your industry demands a dedicated, append-only audit log, treat that as a gap to confirm during selection.

Lot data does not yet auto-post to the accounting ledger. Inventory valuation still requires a manual journal entry. Only sales invoices, purchase bills, and expense reimbursements auto-generate journal entries today. So when a lot is scrapped or written off, the inventory side is tracked in the system, but the ledger side is a deliberate, reviewed entry. That keeps a human in the loop on write-offs, which is conservative and defensible, even if it is not yet one click.

Capacity planning and shop-floor scheduling are not built. The system tracks lots through inventory, not through a manufacturing execution system. If your recall risk lives on the shop floor rather than in finished-goods inventory, you will supplement with a manufacturing execution layer.

None of these limits changes the core point. When the phone rings on a Monday morning, the question is never "do you have MES." The question is "where is the lot." A core business system answers that, today, without a spreadsheet.

How this connects to the rest of your operation

Lot traceability does not stand alone. It is one muscle in a connected body, and its value compounds when the other muscles work with it.

Inventory with lot traceability is the obvious pairing. Movements, transfers, shipments, and shipment batches all carry lot context, so the trace you need at recall time is the same trace your warehouse used to pick and ship. You are not maintaining a parallel record for emergencies. The operational record is the emergency record.

Approval workflows layer on top. When a lot moves from ACTIVE to QUARANTINED, that status change can sit inside a controlled approval, with a requester, an approver, and a timestamp, rather than a free edit. For regulated industries, that chain is the difference between a defensible action and an unexplained one.

The order-to-cash flow ties lots to customers. Quotations become sales orders, sales orders drive shipments, shipments carry lots, and invoices follow. So when you trace lot 117 forward, you land on a customer and an invoice, not just a warehouse bin. That closes the loop from raw material to paid revenue.

And because it is one core business system, the same product master that drives lot management drives purchasing, manufacturing bills of material, and sales. A change to a product's lot requirements ripples correctly everywhere. There is no second list to reconcile.

Frequently Asked Questions

We survived this long without lot traceability. Why change now?

You survived because the calls were rare, and because someone's memory held. The 2025 legacy cliff removes both cushions, because the old system is going away and the staff who carried the knowledge in their heads are retiring as part of the succession-planning wave. The risk is not theoretical. It is demographic and technical at the same time.

Will a migration disrupt our current lot records?

A clean migration maps your existing lot numbers into the new records, preserving dates and statuses. Because the system enforces uniqueness and validates dates, the migration itself becomes a data-quality audit. Bad or duplicate lot numbers surface during import, when you can fix them calmly, rather than during a recall, when you cannot.

Is this affordable for a seventy-person company?

The answer is more favorable than most owners assume. A cloud core business system spreads cost as a predictable subscription, avoids the capital shock of on-premise hardware, and replaces several disconnected tools with one. Kikan System offers a free plan for up to 2 users, no credit card, so you can prove the lot workflow on your own data before you commit.

How does the system stop a recall from expanding beyond the bad lot?

Each lot carries a lifecycle status that controls pickability, so quarantining the lot with one status change immediately stops the warehouse from shipping a single unit of it. Because inventory balances reference the lot, the team can trace which bins still hold stock and which shipments carried the lot out the door, then recall exactly the affected units instead of the whole product line. That targeted recall costs a fraction of a blanket one.

Can the system handle our specific industry rules for lots and expiry?

Food, pharma, chemicals, electronics, and medical devices each carry their own lot and expiry rules. The flexible status model, the notes field, and the per-product lot and expiry toggles let you model most of them without custom code. Where you need a certified storage standard or industry-specific filing, confirm that scope directly, and treat anything beyond finished-goods lot tracking as a roadmap conversation.

Key takeaway: a recall tests your data, not your courage. The companies that survive it cheaply are the ones whose core business system already knows where every lot is, where it went, and who touched it, before the phone ever rings.

Make traceability your default, not your emergency

The cheapest recall is the one confined to the exact lot that is actually defective. The fastest response is the one that does not require building a record from scratch. The most defensible audit is the one your system produced in the ordinary course of business.

None of that happens by accident. It happens when lot traceability is a native capability of your core business system, woven into inventory, approvals, and order-to-cash, with controls that stop bad data before it forms and integrity that holds under scrutiny.

Kikan System builds lot and batch traceability into the core, with auto-generated unique lot numbers, FEFO-driven picking, quarantine and block statuses, date and uniqueness validation, version-controlled records, and clean CSV export for the moment an auditor or a customer asks. Each company's data stays fully isolated, dates resolve in your local timezone, and a lot with on-hand stock cannot be deleted.

Start with the free plan, up to 2 users and no credit card, and prove the lot workflow on your own inventory. Begin at /#get-started, and see plan options at /#pricing. When the Monday morning call comes, you will already have the answer.