Skip to content

Serialisation: why identifying the item matters more than identifying the product

    A barcode tells you what something is. Serialisation tells you which one it is. That is the entire difference, and almost everything a brand wants to do after the sale depends on it.

    A product code is shared by every unit ever made to that specification. It answers “this is a size 10 walking boot, model X”. It cannot answer “this boot”, and questions about this boot are where the value sits: who owns it, when it was made, whether it is in a recalled batch, whether it is stolen, how often it has been back for repair.

    Two kinds of identity

    Most organisations have the first kind and assume it is the second.

    Question Product code answers Serialised identity answers
    What is it? Yes Yes
    How many did we make? Yes Yes
    Is this one in the affected batch? No Yes
    Who owns it now? No Yes
    Has it been repaired, refilled or serviced? No Yes
    Was it reported stolen? No Yes
    Where has it physically been? No Yes
    Is it genuine? No It can be made to

    Every “no” in that column is a business problem currently solved with paperwork, receipts, spreadsheets and phone calls, or not solved at all.

    Recall is where the money is

    Recall is the clearest commercial case, because the cost of not having serialisation is a number the finance director already recognises.

    Without serialised identity, a fault traced to three days of production at one plant means recalling the SKU. Once the units have left the warehouse you cannot prove which came from those three days, so you take back everything, including the large majority that was never affected: logistics, replacement stock, destroyed inventory, and press coverage of a recall that looks far larger than the fault.

    With serialised identity, the recall is scoped to the serial ranges that came off that line in that window. Units that were never at risk stay on the shelf and in customers’ homes, and the public message changes from “we are recalling this product” to “we are recalling these specific items, and here is how to check yours in five seconds”.

    That matters as much as the cost. A scoped recall demonstrates control. An unscoped one admits you do not know what you made.

    Warranty that belongs to the item

    Warranty is normally attached to a receipt: a piece of paper the customer has probably lost, which proves nothing about the object in front of you.

    Attach it to the item and the process simplifies. The item knows when it was made, when it was sold, what its warranty period is and what has already been done to it. The customer proves nothing, and the service agent stops adjudicating and starts repairing.

    The same record does a second job as the item’s service history. Every repair, part replacement and inspection sits against the individual object rather than in a job sheet in someone else’s system. For anything with a maintenance obligation, that record is the point of the exercise.

    Provenance, and why resale needs it

    A whisky cask is a useful illustration because it is a high value object whose entire worth is bound up in its history.

    Give the cask a serialised identity and its digital twin carries the things that make it worth what it is worth: fill cycles, which warehouse it has sat in, its movement history, when it moved and by whom. The buyer is not taking a broker’s word for it and the seller is not reconstructing a story from paperwork. The object carries its own account of itself.

    Generalise that and you have the resale market, which for almost everything durable runs on photographs and optimism. An item that states its own make, date, ownership history and service record changes that transaction, and to the brand’s benefit rather than at its expense: brands with strong resale values sell more new stock, not less. None of it is possible with a product code, because every unit would tell the same story.

    Theft, which only works at item level

    A stolen item registry is the cleanest demonstration of why serialisation is different in kind rather than in degree.

    An owner reports a bike stolen and flags it. The bike has an identity of its own, so the flag attaches to that bike and not to the model. Someone considering it second hand taps it, sees the warning and is asked to call 101.

    Try that with a product code and you have told the world that a model of bike may have been stolen, which is useless to everyone and actively harmful to the honest owners of the other fifty thousand.

    The interesting part is who it protects. Not the original customer, whose bike is already gone. The next person, saved from buying stolen goods by a brand they are not yet a customer of, and the theft itself becomes less worth doing because the resale market for it has been degraded.

    Knowing your installed base

    Most manufacturers can tell you how many units they shipped. Very few can say how many are still in service, where they are or what condition they are in.

    Serialisation closes that gap, and it is the part that quietly changes planning. The circular PPE tracking work we have done with Arco, the safety and workwear distributor, is a good example: what matters operationally is not how many garments were bought but how many are in circulation now, how many washes each has had, which are due for inspection and which should have been retired. No product code answers any of that.

    The same logic made our team’s earlier industrial work possible at all, tracking more than 900 million refillable containers through a circular supply chain. A returnable asset that cannot be individually identified is not an asset you are managing, it is a cost you are estimating.

    The honest cost side

    Serialisation is not free, and the costs are not where people expect.

    Data volume. One row per product becomes one row per item, plus a row per event on it. A different scale of system, though far less frightening than it was ten years ago.

    Variable data at production. Every item needs a different code printed or encoded, so the line has to produce unique output rather than identical output. Routine, but it is where the practical engineering sits.

    The point of assignment. Identity has to be created at the right moment and married reliably to the object. Too early and you have identities without objects. Too late and you have lost the production context that made the identity worth having.

    Uniqueness, forever. The scheme has to survive new ranges, acquisitions, migrations and twenty years. Cheap to design carefully, ruinous to fix afterwards.

    Notably absent from that list: the tag. Printed code, NFC or UHF label is a late decision and a reversible one. The granularity of identity is neither.

    Where to start

    Pick one product and ask what you would need to say about an individual unit of it in five years: who owns it, what has been done to it, whether it is genuine, whether it is still in service.

    If the honest answer is “nothing, we sell it and it is gone”, a product code is sufficient and you should save your money.

    If it is any of those things, the decision cannot be deferred until you need it, because identity is assigned at manufacture. You cannot serialise a population that is already out in the world.

    Veribli has been assigning and tracking item level identity for nearly thirty years, from industrial returnable assets to consumer products, and we won an Innovate UK competition in 2022 for the work. If you want to talk through what granularity your products actually need, call us on 01482 455 855.