Bottom line first: food warehouse batch expiry management works only when three things happen at the same time. Receiving identifies the batch and the expiry date correctly on the first pass. Putaway carries the batch and the expiry date with the location, never separately. Picking returns a FEFO result from the system and leaves the picker no room to improvise. If any one of those three relies on memory or a spreadsheet workaround, expired write-offs and wrong-batch shipments are a matter of time, not luck. I have seen this pitfall three times, and all three times the script was nearly identical: the team decided to go live first and fix the details later, and the busy season blew the whole thing up.
Let me start with a client in Hangzhou, a baking-ingredients distributor with roughly 800 pallet positions, four dock doors, and more than 200 orders a day in peak season. Their WMS had a batch field that sat empty for months. Expiry dates lived on paper delivery notes, and pickers grabbed whichever pallet was closest. One year before the Mid-Autumn rush, a key account ordered cream, and the warehouse shipped fresh stock while three pallets from the previous winter sat untouched in the back rows. By the time the count caught it, those pallets had barely twenty days left. The customer refused them, and the write-off plus markdown cost more than half a year of system fees. Worse, the same SKU had four batches in the building but the system held a single receiving record, so nobody could say which pallet shipped first or who picked it. When they hired me for the cleanup, my first sentence was simple: you do not have a headcount problem, you have a batch problem that never entered the system.
The second story runs in the opposite direction. An imported-food warehouse in Ningbo, frozen and chilled under one roof, with twenty-odd batches crammed into a single freezer. Their system did have batch and expiry fields. The failure was in master data. One supplier changed packaging twice without changing the barcode, and the date format on the label switched from year-month-day to month-year. The system kept parsing everything as year-month-day, so every expiry date it calculated was wrong. Picking ran FEFO faithfully, and faithfully produced the wrong sequence. Recovery took a two-day shutdown and a pallet-by-pallet relabeling check. Since then my pre-go-live checklist has one rule: pull twenty live batches at random, hold the goods next to the screen, and match batch number, production date, expiry date, and location four for four. If the sample fails, fix the mapping that night.
Why food warehouse batch expiry management breaks at picking
Most teams assume food warehouse batch expiry management is a receiving job, done once the goods are booked in. Receiving is only the starting point. The real break happens at picking. A batch misidentified at receiving poisons everything downstream. Different batches mixed onto one pallet leave the picker no honest way to tell them apart. A picking strategy written as FEFO but open to manual override is the same as having no strategy at all. A general warehouse recovers from a wrong shipment with an exchange. A food warehouse that ships the wrong batch may face a refused delivery, and in the worst case it cannot even scope a recall, because the batch trail is broken.
Tracking batch and expiry separately is the most common structural mistake. Some systems keep the batch number in one field and the expiry date in another, with no validation between them. The fix is to treat the batch as the identity of the stock: one batch number must map to exactly one production date, one expiry date, and one supplier lot, and receiving blocks goods in quarantine until all three are present. Mixed pallets get a hard rule too: one pallet, one batch. It sounds rigid, but it pays off at count time.
Putting food warehouse batch expiry management into daily practice
In practice I hand clients one checklist per stage.
At receiving: match the delivery note against the physical goods first, and photograph the batch number, production date, and expiry date. When the system scans goods in, the batch number should pull the expiry date automatically, with no manual editing. Different batches of the same SKU go onto separate pallets with separate labels, even if they differ by a single day. Goods arriving with less shelf life than the contract allows go to quarantine until purchasing confirms acceptance.
At putaway and picking, the rule is one sentence: the system decides, the person executes. Putaway separates short-dated stock into its own zone, and short-dated goods never get buried in deep rack rows. Picking tasks arrive pre-sequenced by FEFO, and the picker scans the batch code to confirm; a wrong batch locks the task. Split-case picking must scan the batch code, not just the SKU code. Some clients ask to drop that scan to save seconds per pick. I talk them out of it, because those seconds get repaid many times over in complaints.
Alerts are where most implementations quietly fail. An alert is not a text message the day before expiry; it is a staged process tied to how fast each product moves. Fast-moving ambient goods can work with a thirty-day alert, while slow-moving imports and seasonal stock need a promotion or transfer process ninety days out. Every alert must name a person and an action, with automatic escalation if nothing happens. The table below is my starting staging; tune the thresholds to your contracts and product mix, and check the official source for the latest policy wherever a regulation or label rule is involved.
| Days remaining | Action | Owner | Escalation |
|---|---|---|---|
| Over 90 days | Sell normally, flagged in system | Warehouse clerk | None |
| 60 to 90 days | Add to promotion or transfer candidate list | Sales contact | Escalate after 7 days idle |
| 30 to 60 days | Freeze full-pallet issue, prioritize split-case issue | Warehouse supervisor | Escalate after 3 days idle |
| Under 30 days | Ship only to customers who accept short-dated goods | Operations lead | Escalate same day |
Three failure points that keep coming back
Master data with no owner is the quietest one. A supplier changes packaging, a barcode, or a date format, and nobody updates the parsing rule, so expiry dates are wrong from the source. The cure is procedural: the first delivery from a new supplier or specification must build and test the master record, including a live scan parse, before bulk receiving. I have seen this pitfall three times, and every retrospective ended the same way: we should have sampled twenty pallets first.
Short-dated stock with no exit is the most expensive one. Alerts fire, nobody on the sales side picks them up, and the goods sit until a bulk write-off. Here is the counterintuitive judgment: expiry management is not won by warning earlier, it is won by agreeing the exit in advance. A client with no short-dated clause gains nothing from a six-month warning, while a client with three agreed exits, markdown, transfer, or return to supplier, handles a thirty-day warning comfortably. Negotiating those exits is purchasing and sales work, but the warehouse must deliver the short-dated list on time. A late list is the same as no list.
Counting quantity without counting batches is the most self-deceiving one. The books balance on quantity while the batch structure underneath is wrong, and the system keeps shipping the wrong goods with confidence. A food warehouse count must sample batches: quantity, batch number, and expiry date must all match before a location passes, and one mismatch expands the count to every location holding that SKU. A report listing only quantity variances has not really counted anything.
One closing note, the same one I leave on every pre-go-live checklist: receiving must identify, putaway must separate, picking must lock, and alerts must land on a named person. If any box cannot be ticked the night before go-live, delay the start by a day. A day of delay costs far less than one wrong batch shipment.
By Lin Xiao, WMS Implementation Consultant







