A product tracking system is set up through four decisions: the level at which the product is numbered, how the number gets onto the label, the points where the product is scanned, and what is written to the record at each scan. These records make it possible to read backwards and forwards where a product came from and where it went; for food and feed, Turkish Law No. 5996 makes this an obligation of the business. What shows the system works is a drill in which a randomly chosen batch can be traced end to end from the record alone.
Which question should the set-up start from?
Projects that try to record everything from day one usually stall at the end of the line. A sounder start is to pick the question the system must answer first, because that question sets both the scan points and the numbering level.
- Narrowing recall scope: a batch number is enough; the first scan points are the end of the line and shipping.
- Seeing channel diversion: distributor and dealer deliveries are scanned from the first phase.
- Catching copied codes: unit-level numbers and end-user scans are needed.
Scope is widened once the records are seen to hold for the first product group.
Batch level or unit level?
A batch number gathers thousands of products made under the same conditions into a single record. The variable data on the label changes once per batch, there is no need to scan every unit separately, and for most product groups it is enough for targeted recalls and expiry date tracking. That is also where its limit lies: products in the same batch carry the same number, so the location of a single pack is unknown and a copied batch code cannot be told apart from the genuine one.
At unit level every label carries its own number. On the line this means a print set-up that can print variable data on each label separately, and a scan of every unit at the end of the line. In return, the same number being read in two places becomes a meaningful signal; how copied codes are caught is covered in the article on serial number verification.
The decision does not have to be frozen. If the number structure is built from the start so that it can extend to unit level, you can begin at batch level and move to unit level only for higher-risk product groups. What makes a later move hard is a batch number designed in a format that leaves no room for an extra field.
How does the number get onto the label?
When the number is placed in the area of the existing label layout where variable data is already printed, the packaging design is not reopened. Printing the human-readable number next to the code lets the record be completed by hand at a point without a reader, or when a code is damaged. Code size, surface curvature and exposure to abrasion are checked on the first trial print; a code that cannot be read is a gap in the record.
Once labels carry numbers, the difference between labels printed and labels applied to products also has to be tracked. How to keep that account is the subject of the label accounting article.
At which points should the product be scanned?
Scan points are the places where the product changes hands. In a typical chain they are:
- End of line: the number is read for the first time and activated as "produced". Without this scan every later record lacks a basis.
- Case and pallet building: the record links which units went into which case and which cases onto which pallet. At later points, scanning the pallet alone then records the units inside.
- Warehouse dispatch and shipping: the record notes which customer the batch went to.
- Distributor and dealer receipt: the record shows the product reached the next link in the chain.
- Returns receipt: a returned product re-enters the record under the same number.
If scanning is optional at any point, the record there thins out over time. Making the distributor and dealer scan part of the handover itself, so that a delivery cannot be completed without it, is the most effective arrangement for keeping the record continuous.
What is written to the record at each scan?
A custody record carries five pieces of information: which batch or unit, which site or point, which date and time, which party, and which operation (production, packing, shipping, receipt, return, destruction). A product code alone is not enough; the batch or serial number must be part of every record.
The GS1 Global Traceability Standard (GTS) describes the identification, data capture, record-keeping and sharing requirements for interoperable systems within this frame. Its practical benefit is that a distributor and a customer can read the same record, with the same fields, in their own systems.
Exceptions need separate handling. If reworked, scrapped or returned products stay out of the record, the chain breaks exactly where the problem arose.
How do you know the system is ready?
Through a drill. A random batch is chosen and traced from the record alone, backwards to its raw materials and forwards to the last delivery point. Two things are noted: how long the trace took and where the record broke. The break is usually where a scan point was skipped or an exception was not recorded, and that is where the fix starts. Repeating the drill regularly is what allows the scope of a real recall to be kept narrow; we cover that side in the article on recall time, and the legal steps in the article on the recall process.
How does set-up work with Tagvex?
Tagvex gives products a unique number at batch or unit level as needed and places the number in the existing label layout; moving from batch to unit level later is possible. The label is read and recorded at the end of the line, and changes of hands at warehouse, distributor and dealer points are gathered in the same record. Records can be connected to ERP and warehouse systems through an API. The record infrastructure is a tool for meeting the traceability obligation that Law No. 5996 places on the business; the obligation itself stays with the business.
