Jak przeprowadzić próbne wycofanie produktu
in a food factory, step by step.
A mock recall is a fire drill for your traceability system: pick a real batch, pretend it's contaminated, and prove, with documents, not confidence, that you can find every kilogram of it before it reaches a consumer. Here is the full procedure we use with Irish producers, including the timing targets auditors actually apply.
What a mock recall is, and what it is not
A mock recall is a documented simulation of a product recall. You select one batch (or one raw-material lot), assume it is unsafe, and then execute your traceability and recall procedures for real: identify affected production, identify every customer who received it, reconcile the quantities, and produce the paperwork, all against the clock.
It is not the same thing as a traceability test, although the two overlap and are often run together. A traceability test proves your records can connect a lot to its inputs and outputs. A mock recall additionally tests the decision-making and communication side: who convenes the recall team, who would call the customers, who would notify the FSAI, who drafts the notice. BRCGS Food Safety BRCGS traceability requirements treats these as related but distinct requirements, traceability is tested under clause 3.9, and the withdrawal/recall procedure is tested under clause 3.11.
Rozróżnienie ma znaczenie podczas audytu. Producent, który potrafi zlokalizować partię w dwadzieścia minut, ale nie ma dowodu, że kiedykolwiek ćwiczył łańcuch decyzyjny przy wycofaniu, nadal otrzyma niezgodność.
Jak często należy je przeprowadzać
BRCGS expects the traceability system to be tested at least annually, and the product withdrawal and recall procedure to be tested at least annually as well. Many retailers' own supplier standards ask for more, twice a year is a common contractual requirement for own-label suppliers, and some specify that at least one exercise per year must include a raw material or packaging lot, not just finished product.
In practice, producers who run a short mock recall quarterly find audits dramatically less stressful than those who scramble once a year. The exercise takes an afternoon when your records are in order, and exposes exactly where they are not when they aren't.
Procedura krok po kroku
- 1Nominate the lot, and log the start timeHave someone other than the QA manager choose the batch: a production date and product picked semi-randomly, ideally one with multiple customers and at least one export or private-label dispatch. Write the start time down immediately, the clock is part of the evidence.
- 2Określ scenariuszGive the exercise a realistic trigger: "supplier notified us that milk delivered on the 14th failed an antibiotic residue test" or "a customer complaint of foreign material in lot B-247". The scenario decides whether you trace forward from a raw material or backward from finished product, do both across the year.
- 3Śledzenie wstecz (o jeden poziom)Zidentyfikuj wszystkie partie surowców i opakowań użyte w dotkniętej partii: nazwy dostawców, daty dostaw, kody partii dostawcy, ilości oraz dokumenty przyjęcia towaru (w tym certyfikaty analizy — COA) dla każdej z nich.
- 4Śledź naprzód (jeden poziom w dół)Identify every dispatch that contained the affected batch: customer, delivery address or depot, dispatch date, dispatch-note number and quantity per line. Include transfers to other sites, samples sent out, and product sold at farm gate or markets, these are the ones spreadsheets miss.
- 5Uzgodnij bilans masyQuantity produced must equal quantity dispatched + quantity in stock + waste/rework/samples. BRCGS expects the reconciliation to land within a tolerance your procedure defines, many producers work to ±1–2%. An unexplained gap is a failed exercise, even if every dispatch was found.
- 6Zasymuluj komunikacjęDraft (do not send) the customer notification and the FSAI notification. List the phone numbers you would call, in order. If a real recall touched retail, the FSAI expects notification without delay, Article 19 of EU food safety regulations places that duty on the food business operator, not on the retailer.
- 7Zanotuj czas zakończenia i sporządź raportThe report should state: scenario, lot, start/finish times, quantities reconciled, percentage recovery, every document referenced, the people involved, and, critically, what did not work and the corrective action for it. A mock recall that finds no weaknesses is usually a mock recall that was not honest.
Co musi zawierać raport końcowy
Pięć najczęstszych błędów podczas symulacji wycofania
W arkuszach kalkulacyjnych kontra w systemie
Everything above is achievable in Excel, producers pass BRCGS audits on spreadsheets every year. The cost is time and fragility: a typical spreadsheet-based mock recall takes half a day of cross-referencing goods-in books, make sheets and dispatch folders, and it depends on the person who built the spreadsheets being available and the VLOOKUPs being intact.
In Keystone, the same exercise is a single action: pick the batch, and the full chain, supplier deliveries upstream, dispatches and customer branches downstream, with quantities, renders in under 30 seconds, exportable as a signed PDF. The mass balance is computed from the same records, so the reconciliation is arithmetic, not archaeology. Producers who moved off spreadsheets typically compress the whole documented exercise from an afternoon to under 90 minutes, most of which is writing the narrative report.
- Run at least one traceability test and one recall-procedure test per year, quarterly if you supply own-label retail.
- Log start and finish times: BRCGS expects full traceability with mass balance within 4 hours.
- Mass balance must close within your stated tolerance, unexplained quantity gaps fail the exercise.
- Testuj oba kierunki w ciągu roku: surowiec — śledzenie do przodu; produkt gotowy — śledzenie wstecz.
- Document the gaps honestly, the corrective-action section is what auditors respect most.
Najczęściej zadawane pytania
Czy symulacja wycofania jest tym samym, co test śledzalności?
Jak długo powinna trwać symulacja wycofania?
Czy muszę powiadomić FSAI w przypadku symulacji wycofania?
Jaką partię wybrać do ćwiczenia?
Co oznacza nieudana symulacja wycofania dla mojego audytu?
Zobacz to w oprogramowaniu, na rzeczywistych danych produkcyjnych.
Keystone runs the full chain, supplier delivery → batch → dispatch → customer, in under 30 seconds. 20-minute discovery call, no sales pitch.