Recall Readiness: Could You Trace One Batch to Every Customer by Lunchtime?

Somewhere in your order history is a batch number that could, one morning, arrive attached to a supplier’s recall notice. From that moment the law expects you to act immediately. The legal duties come from the legislation and guidance covered below; whether you can meet them at speed is decided by your records. This article covers what the law requires, the lunchtime drill — our operational test of whether those records can be retrieved quickly enough — a free one-page drill sheet to run it with, and the runbook for the day it is real.

What the law requires when a batch goes wrong

Recall duties in Great Britain sit in assimilated Regulation 178/2002 — the general food law — and they are blunt. Under Article 19, a food business that considers or has reason to believe a food it has distributed is not in compliance with the food safety requirements must immediately initiate procedures to withdraw it and inform the competent authorities; where the product may have reached consumers, it must effectively inform them and, if necessary, recall. Under Article 18, every food business must be able to trace one step back and one step forward — who supplied each food, and which businesses you supplied. The FSA’s guidance on food incidents, withdrawals and recalls, produced with Food Standards Scotland, sets out how those duties work in practice — including the expectation that details such as supplier and customer, products, quantities and dates are readily available on demand, and that root cause analysis follows an incident.

One further provision decides how expensive your recalls are: the regulation treats a whole batch, lot or consignment as unsafe unless it can be demonstrated otherwise. Where unsafe food forms part of a batch, lot or consignment, the rest of that batch is presumed unsafe unless a detailed assessment supports otherwise — and a wider presumed scope means more stock quarantined, more customers contacted, more credits issued, more product destroyed. Better traceability helps you establish the affected boundary more precisely. The cost of a recall is set less by the defect than by the precision of the records around it, which is exactly what the drill below measures. (None of this article is legal advice; the linked guidance and your own advisers govern.)

The lunchtime drill

To be clear about what follows: the four questions and the lunchtime deadline are not a legal standard. They are an operational drill — a self-test of whether the records the law does require can be retrieved fast enough to be useful in a live incident. Do not wait for a real product recall to discover your traceability. Run the drill cold, one morning, unannounced to the warehouse: pick a batch-tracked product, pick one batch received in the last month, and require four answers by lunchtime:

1. How much of the batch arrived, and when? Goods-in should hold the receipt, the quantity and the date against the batch number.
2. How much is still in the building or on the vans? A current count, by location, so the recall can quarantine what has not left.
3. Which customers received the rest, in what quantities, on which dates? The heart of the drill: the outbound list, per delivery point, that the recall letters and phone calls will work from.
4. Does it reconcile? Reconcile every recorded movement of the batch — deliveries, returns, transfers, wastage, disposal — against what remains on hand. As a simplified drill example: received minus on hand minus delivered minus recorded wastage should come to zero. The gap, if there is one, is stock in the world that no record can place, and in a real recall that gap is the sentence you least want to write down.

Score the drill on two axes: completeness and clock time. Retrieval speed matters in a live incident, and completeness matters just as much — a fast list with holes fails quietly.

Why batch definition and traceability affect recall scope

That whole-batch presumption is also the practical argument for how you receive and record stock. A distributor whose goods-in bench captures batch and date on every receipt, and whose deliveries record which batch went to which customer, can demonstrate the boundary of an incident: this batch, these customers, nothing else. A distributor who receives by product alone cannot demonstrate that boundary, and the presumption then applies to the wider batch. Well-defined, well-recorded batches can narrow the scope a recall has to cover, where the records support it — and that capability is built at the goods-in bench, one receipt at a time.

The capture chain that decides the drill

Nobody assembles traceability during a recall; they surface what the ordinary days captured. Which means the drill is really auditing four unglamorous habits:

• Batch at goods-in, every time. The batch number and its dates recorded at receipt, for every batch-tracked line, with no busy-day exceptions. Miss the batch here and everything downstream is inference.
• Batch through the pick. The pick names the batch and the picker confirms the batch actually taken, so a substitution under time pressure is a recorded fact rather than a silent divergence between the paperwork and the pallet.
• Batch to the visit. The delivered batch tied to the customer and the delivery point at handover, which is the record the outbound list is built from. Proof of delivery gives that record a signature and a timestamp to stand on.
• Batch through returns and transfers. Stock that comes back or moves between sites keeps its batch identity. Anonymous returns are how a recalled batch re-enters the sellable pool through the side door.

Traceability is configured, not assumed. Batch tracking is switched on per product, and the chain only exists for the lines it covers. Part of recall readiness is a deliberate review of which products carry batch tracking and whether that set matches your actual risk, not just your highest-wastage lines.

Hour zero: scope the incident before moving

When a real notice arrives, the first minutes are for reading, not running. Supplier notices normally define the affected scope, for example through batch codes, pack sizes and date ranges. Acting on a misread scope fails in both directions. Too narrow, and affected stock keeps shipping. Too wide, and you are uplifting sound goods and alarming customers about product they never had. So the first three actions are deliberate. Restate the scope in your own terms: which of your product codes, which batches, which dates. Check goods-in for whether you ever received the affected batches at all. "We never handled it" is a legitimate answer, and one you want to be able to evidence rather than assert. That is its own argument for clean receipt records. And classify what you are dealing with. A trade-level withdrawal (stock pulled back before it reaches consumers) and a consumer-level recall are different exercises, with different urgency and communication shapes. The definitions and your formal obligations sit with the FSA guidance linked above, not with this article. Then start the log. A recall is also an evidence-producing event from its first minute. A plain timeline, notice received at 9:40, scope confirmed at 10:05, picking blocked at 10:20, costs nothing to keep and is very expensive to reconstruct.

The product recall runbook: roles before tasks

Recalls go wrong in the gaps between people, so the runbook assigns roles before it lists tasks. One incident owner, a single name with the authority to decide, who runs the clock and the log. A records lead, who produces the drill’s four answers for the live batch and keeps producing them as stock moves. A warehouse lead, who owns quarantine. A transport lead, who owns uplift and what is currently on vehicles. A commercial lead, who owns customer contact and the credit tail. And a nominated contact for the supplier and, where the incident requires it, the authorities, so external communication has one voice. In a small operation several hats sit on one head, which is fine. The failure mode is not few people but unassigned jobs: the quarantine everyone assumed someone else had actioned. Keep the first incident meeting short and focused, with four agenda items: confirmed scope, expected time for the outbound list, confirmation that picking is blocked, and the customer-contact plan. Everything else follows from those.

Quarantine that actually holds

Quarantine has a physical half and a system half, and it holds only if both do. Physically, affected stock moves to a marked, segregated location, and the check extends beyond the racking. Vans out on rounds are carrying today’s stock: any affected quantity on a vehicle needs the driver told, and the stock either returned or set aside on board. Goods in transit from the supplier arrive into quarantine, not into the pick face. In the system, the batch must be unpickable, blocked, reserved or otherwise excluded, so that the 6am picker who never heard the news physically cannot ship it. That last clause is the point. Quarantine by announcement fails on shift changes, and the honest test of yours is to attempt a pick of the quarantined batch and watch it refuse. Two further habits keep the quarantine clean over the days a recall runs. One person, the warehouse lead, controls movements in and out of the quarantine area, and every movement is recorded. The reconciliation at the end should account for every recorded movement of the batch against what remains on hand — and it has to survive the recall period, not just precede it.

Telling customers: the outbound communication

The outbound list turns into a contact exercise, and the contact exercise deserves the same discipline as the trace. Prioritise deliberately, and log every contact. Which customers are contacted first, and by which channel, depends on the incident, the product, the risk and any supplier or regulatory instructions — the discipline is that the sequence is decided, worked through completely, and recorded as it goes. The message itself is short and complete. It names the product, the batch identifiers and how to read them off the pack, and the dates affected. It says what the customer should do (stop using, set aside), how uplift will happen and when, and that credit follows, so the conversation is not quietly resisted on commercial grounds. Record every contact as it happens: who was reached, when, what they said. Chase the unreached rather than assuming a voicemail did the job. The contact log is both the management view of how covered you are and a core page of the evidence pack. Where your customers are retailers with consumer-facing obligations of their own, point-of-sale notices and their own recall duties, your role is to inform them fast and accurately. Their obligations are theirs, and the FSA guidance is the reference for where those lines sit. And if the reconciliation left a gap, unplaceable quantity, say so honestly in the plan. Stating a known gap honestly is far better than a recall later found to have understated its reach.

Uplift, returns and the commercial tail

Recalled goods coming back are a delivery operation in reverse, and they reward being run as one. Collection on existing routes may be suitable where it is compatible with the incident instructions, segregation and recovery requirements; a separate uplift run is the alternative where it is not. Brief drivers so uplifted goods travel as recorded, condition-tagged returns against the customer and the batch, not as loose crates in the bulkhead. Back at the depot, returned recall stock goes into the quarantine area, never the sellable pool. It leaves only by a recorded route: disposal with a record, or return to the supplier with quantities both sides sign. The commercial tail then runs off those records rather than off memory. Credits are issued promptly from the recorded returns, so customers are made whole while the goodwill still matters. And the costs — uplifted quantities, credits, disposal, special runs — are assembled against the incident. Recovering those costs from the supplier is a negotiation that depends directly on the quality of your records. Aim to resolve the credits with the same discipline as the physical stock, rather than letting them run on for weeks afterwards.

The evidence pack

When the dust settles, several parties may reasonably ask you to show your workings: the supplier, particularly where cost recovery is being discussed, your insurer, the technical teams of larger customers, and, in a food incident, the authorities. The evidence pack is what the recall should have produced as a by-product, if the habits above held. It contains the timeline log from hour zero; the goods-in records for the affected batches; the outbound list; and the customer-contact log. It adds the uplift and return records with conditions and quantities, the quarantine movement record, disposal or return-to-supplier documentation, and the credits issued. It closes with the reconciliation, ideally to zero, of everything received against everything accounted for. The test worth applying in the post-drill calm: could you print that pack, today, for the batch you drilled on? Each missing page names a habit to fix. Assemble the pack during the incident, as each step completes, rather than reconstructing it afterwards under someone else’s deadline.

Afterwards: gaps, corrective actions, re-test

End every drill, and every live incident, with a short debrief while memories are fresh, reviewed against the clock times and the gaps. The findings are specific and fixable. Batch capture skipped on the busiest inbound day; a line that should be batch-tracked and is not; customer contacts a year stale; a driver briefing that never reached the agency cover. Each finding becomes a corrective action with an owner and a date, not a nodding agreement. The loop closes with a re-test: run the same drill after the fixes, and see whether the reconciliation now comes to zero and the clock time has moved. Then put the next drill on the calendar as a recurring fixture rather than a one-off virtue. The capture chain decays quietly: staff change, products change, habits slip. The drill is how you find out in a controlled morning what you would otherwise find out during a live incident. A drill that finds a gap has done its job.

Carrying the batch with the goods

The drill sheet fixes what you check. It cannot create the records it checks — that is decided long before any notice, by how goods-in, picking, delivery and returns are captured.

RouteMagic carries batch identity along the whole chain: once batch tracking is enabled for a product, capture is required at the defined points across the back office and the mobile apps, from goods-in through warehouse picking to the site visit, so the delivered batch is tied to the customer with the signature and timestamp of the visit itself. The Product Batch Inventory report answers the drill’s on-hand question by batch, and because deliveries, returns and transfers preserve batch identity, the trail for a named batch runs from goods-in through visit records to the customer — and the Product Batch Inventory and Product Batch Variant Inventory reports provide batch-level inventory visibility, while visit records preserve the customer-delivery trail. The rotation side of the same batch data, using expiry to decide what ships first, is its own discipline with its own article.

Conclusion

How a recall goes depends heavily on records. Either the batch’s journey was captured as it happened — at goods-in, through the pick, to the visit, and back through returns — and the outbound list is a lookup, or it was not, and the list is three weeks of delivery notes read by hand against the clock. The lunchtime drill tells you which business you are, and the lifecycle above turns the answer into procedure: scope before moving, roles before tasks, quarantine that survives an attempted pick, communication that is prioritised and logged, a commercial tail that runs off return records, and an evidence pack that exists as a by-product. Run the drill before a real notice runs it for you, without warning and against the clock.

Recall readiness questions