La respuesta corta
A good trace answers two questions without a spreadsheet: which lots went into this one, and which lots and customers received it. Pick a finished lot, run the trace both ways, and time every step. Fifteen minutes is a target you set and measure on your own data, not a number we claim for your plant. What matters is that you know your real time before a regulator asks.
Why run it as a drill
Most plants only learn how slow their trace is during a real recall, when the clock is already running. A mock recall moves that discovery to a quiet Tuesday.
The drill has one rule: it must cross at least one blend, one rework and one co-product. Those are where traces break. A lot that went through a simple one-in, one-out step is easy. A lot made from three ingredient lots, topped up with reworked product and splitting into a co-product is the real test.
We suggest 15 minutes as a starting target. Treat it as yours: set it, time it, and change it once you know your baseline.
Set up the drill
- Choose the lotPick a finished lot that went through a blend and a rework, and write down what you expect to find.
- Start the clockOne person runs the trace; one person times each stage on paper.
- Trace backwardsFrom the finished lot to every ingredient lot and supplier.
- Trace forwardsFrom one ingredient lot to every lot that contains it, and on to customers.
- CompararDo the two directions agree? Did the result match what you expected?
Trace backwards: what went in
Start at the finished lot and work back. A good backward trace lists every ingredient lot used, including those that entered through rework, and the supplier of each. It should also show the lot attributes that matter to the question, such as potency, moisture or fat, and the lot's current status.
If the finished lot came out of a co-product split, the trace should pass back through the parent batch: whey traces to the milk it was made with, not to a dead end.
Watch for the places a backward trace usually stalls: a bulk tank that mixed several deliveries into one lot, an ingredient received without a lot number, and a rework batch whose source lot was never recorded. Write each stall down with the time it cost you. That list is the real output of the drill.
Trace forwards: where it went
Now take one suspect ingredient lot and go the other way. The forward trace must find every lot that received it, directly or through a blend, through reworked product, and through each co-product made in the same batch. This is the list that decides how wide the recall is.
Then match those lots to customers. A recall workflow that lists affected lots and customers gives you the outbound list as one document. Check what is still on hand too: stock that has not shipped can be held by setting its lot status, from quarantine to released.
What a good trace looks like
- Complete: every branch is followed, including rework and co-products.
- Consistent: backward and forward traces agree on the same lots.
- Specific: lots, quantities and dates, not just item numbers.
- Actionable: affected lots and customers appear as a list you can act on.
- Reproducible: a second person gets the same answer from the same starting lot.
Failing any of these tells you where the data is weak: a lot not recorded at receipt, a rework batch that lost its parent lot, a co-product that never got its own lot.
Timing it
Record the clock at each stage. Leave the last column blank until you have run the drill; the numbers will be yours.
| Stage | What counts as done | Your time |
|---|---|---|
| Backward trace | All ingredient lots and suppliers listed | |
| Forward trace | All receiving lots found, through blends, rework and co-products | |
| Customer list | Affected customers and quantities listed | |
| Stock on hand | Remaining affected stock identified and held | |
| Total |
Repeat the drill every quarter with a different lot, and note what slowed you down. The gaps are usually data entry gaps, not software gaps.
If you miss your target, resist the urge to shorten the drill. Narrow the cause instead. Was the time lost finding the lots, reading the genealogy, or building the customer list? Each has a different owner: receiving, production or sales administration. Fix the slowest stage first, then run the drill again with a new lot and compare the two totals.
In Business Central
The features sit in ManufacturingFlow 365, in the Lots and traceability module.
- Lot attributes, such as potency, moisture and fat.
- Lot status, from quarantine to released.
- Expiry derived from the earliest-expiring ingredient.
- Genealogy forwards and backwards through blends, rework and co-products.
- A recall workflow that lists affected lots and customers.
The dairy scenario in the sandbox gives you a batch with a co-product, a by-product and waste to practise on. This supports traceability and audit trails; it is not a statement of compliance with any scheme. See the ManufacturingFlow 365 page for the full module list.
Limits
Genealogy and recall have been run in our user acceptance sandbox with demo data, and we have not measured recall times in a live plant. The electronic batch record, e-signature and change approval have not been tested with a second user. Exports and some planning pages have not been run. A trace is only as good as the lots recorded at receipt and at each step.
See it on your process
Tell us the product you would pick for your own drill. Book the dairy demo scenario and we will trace a cheese lot back to its milk and forward to customers in a sandbox, so you can time it yourself.
Keep reading: questions to ask an MES vendor and the weighbridge as your first quality gate, where good lot data begins.