Return Logistics Reality

Return request filing

Opening a return is a filing exercise, and the first version is the one that gets judged. The request asks for a reason chosen from a fixed list, a short description, and usually a photograph. Reason codes are not interchangeable: a flaw that arrived with the item, a wrong item sent, and a change of mind sit in different categories, and the category decides which remedy the system offers next. Reading the code list before writing anything is the cheapest three minutes in the whole process.

The window runs from a date the system chooses, normally the delivery scan rather than the afternoon the parcel was opened, and it is short. Reading the window before writing anything prevents the most common loss in this cluster, which is a well-documented case filed after the option has closed. Note the closing date in the same file as the evidence, not in a device calendar that may be replaced. A case filed on the final day also has the least room for a second attempt, which is an argument for filing early rather than perfectly.

Description length is capped, so the useful text is dense: part, measurement, promise made on the page, difference observed. Photographs go up as originals, because the filing screen compresses what it accepts, and a compressed copy of a defect can lose exactly the detail the chosen reason code is about. Keep the untransmitted originals in the folder whether or not every one of them is used. Where the form accepts only one image, send the frame that carries the scale.

Reverse logistics quote

A return becomes real only when somebody prices the journey back. Small parcels travel the lane in reverse at rates set by the carrier rather than by the seller, and the figure is quoted per parcel, so a five-kilogram coat and a two-hundred-gram scarf land in different worlds. Two quotes taken in the same week are enough to show whether the return is a remedy or a second loss. Ask both carriers for the same declared value and the same dimensions, or the comparison is worthless.

Volumetric rules follow the parcel home as well. A bulky item charged by volume outbound is charged by volume inbound, which is where outerwear and boxed electronics turn an apparently cheap return into one costing more than the item. Ask for the quote in writing with the weight and dimensions used, and keep the reply; a figure that changes at the counter cannot be argued with afterwards. Where two quotes differ by more than a third, one of them is usually pricing a different service tier.

Net recovery is the number that decides. Item value, plus outbound freight already spent, minus return freight, minus packing and the trip, minus the value of the hours involved. Where that arithmetic lands near zero or below it, the return route expresses dissatisfaction rather than recovering money, and the desk's costing tool places all three dispositions side by side for exactly that reason. Writing the subtraction down before deciding also settles the argument with yourself.

One line in the quote deserves separate reading: the declared value used for the reverse leg. A parcel declared at the full item value attracts a higher rate and, on some lanes, a customs event on the way back into the origin country. Parcels declared at a nominal figure travel cheaper and carry almost no cover if they disappear on the way. The desk records which figure was used rather than recommending one, because that choice belongs to whoever would carry the loss.

Label issuance

Three arrangements cover most returns. The seller issues a prepaid label, the platform issues one from its own account, or the buyer pays at the counter and claims the cost back. Which arrangement applies follows from the reason code and from the seller's settings, and it is visible before the parcel is packed; a return that starts down the wrong arrangement usually has to be refiled from the beginning. Capture the arrangement as it appears on the case page.

A prepaid label carries the seller's reference, which makes the parcel traceable to the case even if the carton is separated from its paperwork. Where the buyer pays first, the counter receipt is the only proof of the amount, and photographs of receipts fade, so the desk keeps both the printed slip and a photograph of it in the same folder. A receipt shot on a dark counter with a flash is worth less than one scanned flat on a pale surface.

Labels also carry the address the parcel must reach, and that address is not always the one the item came from. Consolidators, forwarders and return centres all appear. Before handing anything over, compare the label address with the address held in the case file and photograph both; a parcel sent to the previous address in a case is the most expensive filing error recorded in this cluster. Address mismatches are also the error that takes longest to unpick.

Handover scan

The return begins at the first scan, not at the moment the parcel leaves the hand. Until a carrier records acceptance, the system shows the return as not shipped, and the clock governing the refund has not started. Handing a parcel to a courier who collects without scanning leaves the whole case resting on a memory. The gap between handover and first scan is where most return disputes are actually born.

Get acceptance in a form that can be re-read: a counter receipt with a tracking number, or a collection note the driver generates. Photograph it beside the parcel before the parcel goes. Where collection goes unrecorded, ask for the tracking number to be activated at the point of pickup, which turns a vague handover into a dated event with an owner. Collection notes fade faster than counter receipts, so the photograph is the copy that lasts.

Pauses happen on the way back as well. A return lane can sit in export processing, cross a border, or wait at a return centre for a batch to be opened, and none of those stages produce a scan a buyer can see. Reading the pause as failure is the fastest way to spend a fortnight on messages that change nothing; the useful move is to record the last scan and its date, then compare that silence with the band the same lane produced outbound.

Where a return scan never appears at all, the case changes shape rather than ending. The receipt becomes the claim, the carrier becomes the party being asked, and the platform route is usually put on hold while that question is answered. Filing a carrier enquiry with the tracking number, the receipt image and the date of handover gives the case something to move on, and it takes less time than a fifth message to the seller.

Refund settlement

Refunds arrive in the currency of the original payment, through the original rail, and on the platform's schedule rather than the seller's promise. That schedule starts from a scan, which is why the handover receipt matters: without a first scan the refund clock is not running, and the case sits in a state that looks like a refusal but is only an unmet condition. Reading the case page for the condition is faster than asking about it.

Partial settlements come back faster than full ones in most recorded cases, because they need no reverse journey. The trade is arithmetic: a partial refund that keeps an unusable item leaves the buyer with the loss minus the credit, while a full refund requiring a return leaves the freight line to be settled somewhere. Pricing both before choosing is what the return-cost tool is for, and the figure usually makes the decision obvious.

Watch the currency spread on the way back. A refund converted from the settlement currency into the account currency passes through a rate set on the day of settlement rather than the day of payment, and the difference is not recoverable through any route in this cluster. Recording both figures in the case file turns an unexplained shortfall into a dated, checkable line. Two refunds a month apart on the same rail usually show the spread moving.

Ledger entry

A closed case earns its value only if it is written down. The desk logs six fields for every return it follows: date opened, reason code, route taken, money spent on the reverse leg, days elapsed to settlement, and the net figure recovered. Six fields fit on one line, and one line is enough to compare two cases a year apart. Fields that cannot be established are marked blank rather than estimated.

The ledger changes decisions rather than describing them. Where three previous cases on one lane each cost more in return freight than they recovered, the fourth case begins with the partial-refund route as its default. That pattern is invisible in a folder of screenshots and obvious in six numeric columns, which is the whole argument for keeping the line. Patterns take time to appear, which is why the ledger is reviewed monthly rather than after every case.

Keep entries factual and keep them dated. A ledger written in the language of grievance becomes a diary; a ledger written as amounts and days becomes a dataset, and datasets can be argued with, updated, and abandoned once they stop paying for themselves. Entries that turn out to be wrong are corrected in place with the correction dated, never deleted and rewritten. A ledger nobody trusts costs more to keep than it returns.

Retire lanes from the ledger when the numbers stop changing. A route that has produced the same net loss across four consecutive cases is settled knowledge, and further entries only add work. The desk keeps those rows and stops updating them, with the retirement date on the line, so a future lane change can be compared against a baseline that is stated rather than remembered.