Fashion from warehouse to shop floor: keeping one identity across the handover

01 / FIELD NOTE

Keep the decision tied to the operating context.

A fashion item passes through more hands than most products. It is made, packed, shipped to a distribution centre, allocated, picked, dispatched to a store, received, put on a rail, possibly moved between stores, and eventually sold or returned. Each of those is a handover, and each is an opportunity for the record to stop following the goods.

The value of tagging at the point of manufacture is that one identity can serve the whole journey. The item that was packed is the item that was received, counted, picked and sold, and every one of those steps can refer to the same thing. Where tags are applied later, the earlier part of the journey is invisible and the record begins partway through.

The warehouse stage is where the identity is most directly useful, because that is where quantity is handled. Receiving against a document, put-away to a location, picking against an order and dispatch verification are all steps where knowing which specific items moved improves a decision. The store stage is where identity becomes buyer-facing, because that is where a specific size matters to a specific person.

The handover between them is the weak point, and it is organisational rather than technical. A dispatch recorded at the warehouse but never confirmed at the store leaves the goods in transit indefinitely — not lost, not present, and unattributable. Unit identity makes the gap locatable, which converts an open question into a reconciliation task with a defined scope.

Allocation decisions add a layer that general warehousing does not have. Goods may be allocated to stores after arrival, or reallocated when one store sells faster than another. Those decisions are made against the record, so a record that is stale produces allocations that send stock where it is not needed and hold it where it is. The frequency of the count sets how current the allocation inputs are.

The seasonal rhythm compresses everything. Goods arrive in a window and ship in a window, so both the warehouse and the stores experience a short period of high volume and high density, which is exactly when reads are hardest and when errors are most likely to accumulate unnoticed. A plan validated in a quiet period has been validated against the easiest conditions it will see.

Returns run in both directions and behave differently. A store returning unsold stock to the warehouse is a transfer that can be reconciled against a shipment. A shopper returning a garment to a store is a sale being reversed, and the item needs both a refund and a disposition. Treating them as the same flow is what leads to stock that is physically present and recorded nowhere useful.

Markdown and outlet flows are the disposition for goods that did not sell, and they depend on knowing age and movement. A garment that has stopped selling at one store may sell at another, and moving it is cheaper than discounting it. That decision needs the movement history of the specific item, which a periodic count cannot supply because it does not record when an item went quiet.

The shop floor adds physical constraints the warehouse does not have. Fixtures, fittings and display furniture shield tags; garments on a dense rail read differently from folded stock; security fittings and metallic trims change how an individual item responds. Each of those is a property of the store, and a measurement taken there rather than in a demonstration is what establishes what the store can rely on.

The checkout connects the two halves and is the moment the identity stops being an inventory fact. Deactivating a tag at the till removes the item from the security system, and posting the sale against the same identity removes it from available stock in the same transaction. That is the point at which the warehouse’s record and the store’s record converge, and it only works if the identity read at the till is the one the record holds.

What holds the whole chain together is that each handover is recorded as an event rather than inferred from a change in position. A transfer that is sent and received as two events can be reconciled; a transfer inferred from two stock figures that differ cannot. The design decision is small and its consequence is that a discrepancy can be attributed to a handover rather than to the system as a whole.

The realistic outcome is a garment whose journey is reconstructable: where it was made, when it arrived, where it was stored, when it went to which store, and what became of it. That is what makes a supply-chain question answerable without a phone call, and it is also what makes a loss attributable to a stage — which is the difference between a problem someone can fix and a variance nobody can explain.