Shelf Life Planning
Somewhere in your cooler, a pallet is aging out while the schedule makes more of the same SKU. Semia reads real lot dates from the systems you already run, learns each retailer's date-code rules from your planner, then writes tomorrow's plan so the oldest stock moves first and short-dated product stops getting picked for the wrong order. A named planner approves every plan before it reaches the floor.
The lot dates exist, in the ERP, in a spreadsheet, on a pallet tag, but nobody looks across all of them while building the schedule. So the plan quietly makes more of a SKU while pallets of it age in the cooler. Waste like this is one bucket of the $2-3M a year a typical plant bleeds, roughly $250K of it in product written off or dumped.
Retailers set minimum-remaining-shelf-life rules, so a load can be perfectly good by your standard and still refused at their dock. If the plan does not know each customer's date-code window, short-dated product gets picked for the wrong order. The driver finds out at the DC, not the planner at the desk.
A bounced load is never one loss: you pay the freight, the product comes back with even less life on it, and the deduction lands on the invoice. Chargebacks run roughly $500K a year at a typical plant. Most of the date-code portion is preventable at planning time, before the truck is loaded.
Even when Monday's schedule respects code dates, the week does not cooperate: orders move, a line goes down, a shift comes up short. Planners say the schedule is fiction by Wednesday, and every manual re-plan is another chance to lose track of which lot ships first. FEFO only works if it survives the re-plan.
Semia reads orders, inventory, lot dates, and shelf-life risk from the systems the plant already runs: SAP, Odoo, Microsoft Dynamics, Excel, Google Sheets. Access is read-only, nothing is migrated and nothing is installed on the line. If two sources disagree on a date, it asks your planner, the way a careful new hire would.
Every morning Semia writes tomorrow's plan with the oldest usable lots consumed first and runs ordered so finished goods still clear each retailer's minimum-remaining-shelf-life window at delivery. It balances that against finite capacity and allergen changeover cleanouts, not in isolation. For the definitions behind FEFO scheduling, see /glossary/fefo-shelf-life-scheduling.
When an order shifts or a line stops, Semia re-plans and recalculates which lots are now at risk of aging out, then re-sequences before the risk becomes a write-off. The date logic does not get dropped in the scramble, because it is not held in one person's head.
Every call comes with plain-language reasoning: which lot, which order, which date rule, with a full audit trail behind it. A named planner approves the plan before anything reaches the floor. When QA or a customer asks why a lot shipped where it did, the answer is on the page.
Setup starts with read access to the systems you already have, which typically takes about 30 minutes. There is no data migration project and no new software on the line.
Semia onboards itself: it sits next to your planner, asks questions, and learns by shadowing, including the unwritten rules like which customer bounces anything short-dated. It is useful in about two weeks, drafting plans your planner reviews and corrects.
Lot data and order data stay inside your company. The planning logic Semia captures, the date-code exceptions, the customer quirks, the sequencing habits, belongs to you, not to us.
Short-life products make FEFO planning unforgiving: bread and desserts, dairy, fresh meat. See how it plays out in /industries/bakery, /industries/dairy, and /industries/meat, and what it costs at /pricing.
FIFO consumes stock in the order it arrived; FEFO consumes it in the order it expires, which matters whenever a later delivery carries an earlier date code. In production planning, FEFO also shapes the run order, so finished goods leave with enough remaining shelf life to clear each customer's requirements. For full definitions, see /glossary/fefo-shelf-life-scheduling.
Most retailers enforce minimum-remaining-shelf-life rules: a load must arrive with an agreed share of its printed life still ahead of it, or it is refused at the dock. The product can be perfectly safe and still bounce, and the rejection usually brings freight costs, returned short-dated stock, and a chargeback. Planning run order and lot allocation against each retailer's window prevents it before the truck is loaded.
A typical food plant bleeds $2-3M a year across overproduction, idle capacity, and chargebacks; roughly $250K of that is waste and roughly $500K is chargebacks. The portion tied to dates, expired lots, short-dated picks, rejected loads, is largely a planning problem, because the dates were knowable when the schedule was written. Software that plans from real code dates addresses that portion directly.
No. Semia reads lot dates from wherever they live today: the ERP, Excel, Google Sheets, even a mix that does not quite agree. When sources conflict, it asks your planner instead of guessing, the same way a careful new hire would. That is a deliberate contrast with traditional APS tools, which tend to depend on clean structured data before they become useful.
A named human always approves. Semia drafts tomorrow's plan each morning with plain-language reasoning for every lot allocation and run-order call, plus a full audit trail. Nothing reaches the floor until your planner signs off, which also gives you a documented answer when a customer questions a shipment's dates.