A manifest, not an ASN.
The load was announced by your own system, so both sides read the same expected quantity. That is exactly why the two counts have to be captured separately.
Reference · Returns intake
When the shipper is another site in your own operation, there is no supplier to invoice and no vendor label to scan. The load arrives on an internal manifest, in totes and gaylords packed by hand, and the argument that follows is between two parts of the same company. The evidence still has to be built at the door.
The difference
The receiving motion looks familiar. Four things underneath it are not.
The load was announced by your own system, so both sides read the same expected quantity. That is exactly why the two counts have to be captured separately.
Returns and transfers carry an internal tote tag, asset tag, or route label. The barcode reads fine, but it resolves against your numbering rather than a published standard.
The counted thing is a tote, gaylord, bin, or mixed pallet built by hand at the sending site. Its contents are not implied by its label.
Whether an item arrived intact is often the whole reason the unit is moving, so the picture matters as much as the count.
The symbologies are mostly the same ones on inbound freight. See the barcode and label cheat sheet for what each encodes.
The anchors
A returns intake has no purchase order and no advance ship notice. These identifiers do that work instead.
The internal document the load moves under. Later questions get answered against this number, not a memory of which week it arrived.
Which plant, store, kiosk, or processing center released it. A shortfall that started at origin is a different conversation than one that started at your door.
The seal read and photographed before it is cut, which door, and when the unload started and finished. Door and time are what tie one read to one load.
The tote, gaylord, or pallet tag naming one physical unit, read from the label rather than typed in from a clipboard.
Record the received count before it is reconciled. The original, the correction, and the reason are three separate facts.
The intake
A gaylord of returned material is a container of unknowns until someone opens it. The order of operations is what keeps the record standing up.
Read every handling unit off the trailer before anything is opened. A missing unit is cheapest to catch here.
A mixed tote is counted at the item level against what the manifest says it holds. Expected, counted, and the difference are three values, not one net figure.
Photograph the unit as it arrived, then the damage, then the label. Name the affected quantity rather than writing off the whole unit.
An unreadable internal tag is its own discrepancy. Document it at the door with a picture, keyed in by hand and flagged.
Camera inspection of material condition as it comes off the trailer is what Trainsight does.
The record
Weeks later the question is rarely how many arrived. It is which unit, from where, and in what condition.
The same argument, on the supplier side, is laid out in the receiving record.
Receiving material that is moving back through your network rather than forward from a supplier: customer returns, retailer returns, kiosk pickups, and transfers between your own sites. The load arrives on an internal manifest instead of an advance ship notice, and the counterparty is another part of the same operation.
Against the tag your own network printed. A tote tag, asset tag, or route label identifies one handling unit the way an SSCC identifies a pallet, and it resolves against your numbering rather than a published standard. What matters for the record is that the identifier is read from the label rather than typed in from a clipboard.
The transfer or return number, the sending site, the trailer, seal, dock door, and time, and the tag on each handling unit. Then the manifest count and the received count with the difference stated plainly, pictures of the unit, label, and any damage with their capture times, the condition call, and who counted and verified.
In two passes. Count the handling units against the manifest before anything is opened, since that is the cheapest place to catch a missing unit. Then count the contents at the item level against what the manifest says that unit holds, recording expected, counted, and the difference as three separate values.
Photograph the unit as it arrived, then the specific damage, then the label, before the unit moves anywhere in the building. Name the affected quantity rather than writing off the whole unit, and record whether it was refused, accepted with an exception, or accepted clean.
A per unit record rather than a total. One total can only be answered with another total, while a record showing which unit was short, at what time, at which door, next to an intact sequence of every other unit received, can be checked against the sending site pack and load records.
Every scan is saved with its photo and the exact time it crossed, so the unit tag, the count, and the picture are already together when a sending site reads the number differently than you do.