Mock Recall Program

A mock recall is a planned, time-bounded drill that simulates the full recall workflow without actually executing the recall: select a lot, define the simulated trigger (e.g. "laboratory finding of Salmonella in retained sample of Lot 24-3815"), and require the recall coordinator to execute the recall plan to 100% lot reconciliation. The drill measures the time taken, the percentage of units reconciled, and the gaps identified. Mock recalls are required by GFSI-benchmarked schemes annually (SQF) or more frequently (BRCGS often quarterly), and many major retailers' supplier scorecards require quarterly mock-recall results as a condition of continued supply.

• GFSI-benchmarked schemes (SQF / BRCGS / FSSC 22000): ≥ 99.5% reconciliation in ≤ 4 hours. Below this is a non-conformance.
• Walmart Supplier Quality Excellence Program (SQEP): ≤ 2 hours for high-risk products, ≤ 4 hours otherwise; mock-recall results audited quarterly.
• Costco supplier standards: ≤ 2 hours; quarterly drill required; failure flags supplier scorecard.
• Target supplier requirements: ≤ 4 hours; annual mock-recall report required.
• FSMA 204 implicit expectation: KDE / CTE records must be retrievable in 24 hours; firms should be able to assemble Sortable Spreadsheet on demand. Mock recalls demonstrate this capability.
• USDA-FSIS expectation under Directive 8080.1: full distribution list in ≤ 24 hours, but for Class I recalls (high-risk) USDA expects firms to be able to assemble immediately.
• Pharmaceutical / supplement firms under 21 CFR Part 7: Class I recall expectation is ≤ 24 hours to initiate communications, ≤ 48 hours to file with FDA, ≤ 30 days to complete recall execution.

The weakest mock recalls are the ones designed for the firm to pass — a single recent lot of a top-velocity SKU with full bin-level traceability. A credible mock-recall program rotates scenario design to exercise the harder cases:

• Oldest distributed lot within recall window — exercises long-tail distribution records and customer-list staleness.
• Discontinued SKU — exercises ability to reconstruct distribution for a product no longer in active production.
• Contract-manufactured product — exercises supplier-side traceability and quality-agreement records.
• Co-manufactured product — exercises the brand's ability to recall when the brand does not control the manufacturing site.
• Multi-lot recall — exercise scaling: 1 lot, 3 lots, 10 lots. Reconciliation time should scale sub-linearly.
• Cross-border distribution — exercise multi-jurisdiction recall (US + Canada + EU + Mexico) coordination.
• Foodservice + retail dual-channel — exercise distinct distributor / retailer / foodservice paths.
• Component-level recall — supplier notifies us a raw-material lot is recalled; exercise downstream traceability across multiple finished-product lots.
• Equipment-mediated contamination — exercise traceability via equipment-usage records (every batch that used the same blender for the 48 hours before / after the suspect window).
• Inadvertent label mix-up — exercise label-reconciliation records and reverse-logistics for relabelling.

• Distribution data lives in a system disconnected from quality — sales / shipping data in ERP, lot data in MES, no automated join. Reconciliation requires manual spreadsheet work.
• Customer contact list is stale — names / emails / phones of customer quality contacts are 2-3 years old; reaching them in real-time is impossible.
• Lot-to-pallet-to-customer is not fully captured — pallet build records show pallet contents but not which pallet went on which truck to which customer.
• Bin-level granularity missing — finished product is tracked at batch level but split across multiple containers / bins / pallets each of which may have different downstream destinations.
• Co-manufactured product traceability gaps — brand does not have copy of CM batch record; CM not contractually obligated to provide within recall-response window.
• Components shared across many finished-product lots — single contaminated component lot affects 50+ finished lots; reconciliation requires recursive graph traversal that takes hours by hand.
• Mock-recall scope artificially narrow — same scenario every year; firm passes; no learning.
• Mock-recall report not reviewed by senior management — gaps identified but not closed; same gaps appear in subsequent drills.
• Mock recall counts as recall preparation but no parallel recall pre-flight engine — when a real event happens, the firm has tested the data side but not the decision side (lot scope, distribution scope, regulatory notification, media response).
• Effectiveness checks on corrective actions never close — gap identified in 2024 drill still open in 2026 drill.

• Scenario library: rotating library of mock-recall scenarios across the dimensions above (lot age, SKU lifecycle, contract / co-mfg, multi-lot, cross-border, equipment-mediated, component-level, label-mix-up).
• Live data: traceability runs against the same genealogy graph that real recall execution uses; results are representative of real-world performance.
• Bidirectional graph traversal: forward (lot → customer / consumer) and backward (lot → component / supplier) using Postgres recursive CTE; typical traversal of 50M+ row event tables in ≤ 2 seconds.
• Reconciliation report: units produced, shipped, on-hand, returned, scrapped, unaccounted for; percentage reconciled; time to reconciliation.
• Customer / distributor list freshness: V5 tracks last-contact date for each customer quality contact; flags contacts > 12 months stale before drill.
• Communication drafting: mock-recall communications drafted (not sent) using current customer / distributor / retailer / regulator contact lists; sample-of-five live-call verification ("can we reach this person right now?") for the critical customers.
• Gap-to-CAPA: every gap identified in the drill auto-opens a CAPA workflow with assigned owner + due date + effectiveness check.
• Senior officer dashboard: mock-recall results visible to executive team; quarterly trend reports.
• GFSI-audit-ready report: mock-recall report generated in the format SQF / BRCGS / FSSC 22000 auditors expect; signed by senior officer; archived.
• Retailer-scorecard export: results exported in the format Walmart / Costco / Target supplier-portals consume.
• Real-recall continuity: when a real event happens, the same engine + same scenarios + same data backbone drives recall pre-flight and execution.