If you sell bulk material to an export buyer, three requests arrive with almost every order: which lot is this, when does it need retesting, and can we have the certificate of analysis. Soap noodles, oleochemicals, base oils, food ingredients, specialty chemicals: the products differ, the paperwork does not.
Plants usually answer those requests from a folder of PDFs and a spreadsheet that someone keeps by hand. It works until the person is on leave, or until a buyer asks why the certificate they were sent does not match the lot on the packing list. This post is about how lots, retest dates and certificates should be generated so that they cannot drift apart.
A lot is born when a batch completes, not when someone types it
The cleanest definition of a lot for a bulk product is the output of one completed production batch. That gives you a production date for free, because the batch has one, and it gives you a quantity, because the batch recorded what it made.
So the lot should be stamped automatically at the moment a batch is marked complete. No form to fill in, no lot number to invent. In Aourix, completing a bulk batch writes the lot onto it: production date, and a retest date derived from the product's retest interval. A product configured with a twelve-month retest interval produces lots that carry a retest date twelve months after production, every time, without anyone remembering to calculate it.
This matters for one reason: a lot that is created by a human step can be skipped. A lot that is created by the completion of the batch cannot.
Retest dates are a status, not a deadline
A retest date is not an expiry date. Material past retest is not unusable; it is material that needs its quality re-confirmed before it ships. Treating the two as the same thing leads to either of two mistakes: writing off good stock, or shipping stock nobody has re-checked.
The honest model has three states for a lot. Within retest, it ships normally. Past retest, it is flagged, and picking it prompts for a decision. Requalified, meaning someone has re-tested it and recorded that fact, it ships normally again with a record of who requalified it and when. Every one of those transitions should be visible on the lot itself, so the buyer's question "was this re-tested?" has an answer with a name and a date on it.
FEFO should advise, not block
First-expired, first-out is the right default for picking lots: ship the material closest to its retest date first. But a rule that hard-blocks any other choice fails in the real world within a week. A buyer specifies a particular lot. A particular lot is physically closer to the loading bay. A newer lot has a better analysis for a demanding customer.
The right design is advisory FEFO. When a dispatch line is being built, the picker orders the available lots oldest first and pre-selects the one FEFO would choose. If the operator picks a different lot, the system asks for a reason and records it on the line. Nothing is blocked; everything is explained. When the quality manager reviews the month, every FEFO override has a reason next to it, which is a far more useful audit trail than a list of times someone found a workaround.
The certificate comes from the lot, not from a template
A certificate of analysis should be a document generated from data that already exists: the product's published specification bands, the lot's recorded results, the dispatch it is travelling with. If the certificate is produced by opening last month's certificate and editing the numbers, it will eventually go out with last month's lot number on it.
In Aourix, a certificate is generated per dispatch line and batch. It carries the specification the product publishes, for instance total fatty matter, moisture, free alkali and chloride for a soap noodle grade, the lot's production and retest dates, and the quality status of the batch. Because it is generated at dispatch from the same records the packing list uses, it cannot disagree with the packing list.
Keep it to bulk
One deliberate limit: model lots for bulk products only. Packed finished goods carry a batch code on the pack, and the pack form, pallet configuration and carton count are their own concern. Trying to track a lot through a carton of retail units multiplies records without improving traceability, because the buyer's question at that level is "which batch is on the carton", which the pack already answers.
What to check in your own plant
- Is a lot created automatically when a batch completes, or does someone type it later?
- Does every bulk product have a retest interval, so the retest date is derived rather than guessed?
- Can a lot past retest be requalified, with the person and date recorded?
- When an operator picks a lot FEFO would not have chosen, is a reason captured?
- Is the certificate of analysis generated from the lot and dispatch records, or edited from a previous one?
Five yes answers mean the three questions buyers ask have answers that will still be right when the person who set it up is on leave.