Resale Margin After Everything

Built table of margins

A resale plan that starts from the price somebody hopes to charge is working backwards. The figure that decides the outcome is the unit cost of putting one item on a shelf, and that figure is the delivered total divided by the number of units in the parcel. Everything in the margin table follows from that division, and the columns beside it exist because a unit cost on its own cannot say whether a price is achievable. A plan assembled the other way round tends to look healthy until the first parcel has to go back.

Four columns carry the table. Unit delivered cost spreads the landed total across the units in the consignment. Break-even is the sale price at which the cost and the commission on a sale are both covered. Unit margin subtracts that cost, the commission on the chosen price, and an allowance for the category's recorded share of orders that were returned or closed at a reduced price. Reading the last column without the third is how a plausible plan becomes a loss, and reading the second without the first is how a comfortable margin turns out to be break-even.

Footwear and bags sit in the middle of the recorded range. Seven in a hundred footwear orders and nine in a hundred bag orders were returned or closed at a reduced price, and both travel at weights that keep a return leg affordable. Their margins are limited more often by the price a resale can realistically reach than by the cost of landing the goods, which makes them the categories where a local price check changes the plan most. Neither is cheap to send back either, so the price side carries more of the decision than the cost side does.

Watches, electronics and the light lines behave differently. Watches and electronics carry the higher recorded return shares, at eleven and thirteen in a hundred, and both need a stronger price to absorb that allowance. Accessories, jerseys, outerwear and fragrance come back less often, so their margins answer mainly to carriage and to the number of units sharing it. What moves margin across the whole table is the return share and the unit count rather than the headline price of the goods, and nothing in it rewards paying more for an item without knowing that.

Data provenance and method

Every return share here comes from a recorded set of orders rather than from a guess, and each carries its own sample size. Footwear rests on 58 recorded orders, bags on 41, watches on 33, jerseys on 37, outerwear on 26, accessories on 44, electronics on 29 and fragrance on 12. Those shares count orders that came back or closed with a partial refund; they describe events rather than items, and the check date behind them is 2026-09-24. Samples that small are printed beside every share rather than hidden behind an average, because a figure drawn from twelve orders and one drawn from fifty-eight do not deserve the same confidence.

Two inputs are deliberately not supplied by this desk. The commission a resale channel charges is entered by the reader, because it differs by channel and by category. The price an item can be sold for is also a reader input, since the desk records what buyers paid rather than what a seller might ask. A margin calculated without both would carry an authority it has not earned, so the tool asks for them instead of supplying them. That arrangement is uncomfortable and deliberate: a tool that invented a commission rate would be answering a question about somebody else's channel.

The returns allowance is constructed rather than assumed. A category's recorded share is applied to the value at risk on a return, and that value includes the freight already spent getting the item out plus the loss taken when a returned unit is sold locally instead of travelling back. The desk models that local loss at roughly a third of item value, an editorial figure rather than an observation, which is why the column is described as an allowance. Being the softest number in the table, it is also the one most likely to be argued with, and it stays editable for that reason.

What the method leaves out matters as much as what it includes. Storage, the holding time between paying and selling, and the cost of capital tied up in unsold goods are all absent, so a plan taking months to clear is understated by this table. Fees charged to a buyer rather than to a seller are absent for the same reason. Within those boundaries the three columns are worth more than one optimistic number, because they show which input would have to move for the plan to fail.

How to read the columns

Start at the left. Unit delivered cost is the landing figure carried in from the waterfall, divided by the number of units in the parcel. Where several items travel under one carriage charge, that division is what makes them comparable, and it is also where a consolidation decision shows up: adding a second unit raises the carriage line slightly and lowers the unit figure, provided the carriage band does not change. Where a consolidation has already been priced, enter the combined carriage rather than the two original rates, or the division will flatter the plan.

Break-even sits in the second column and is not the same number as unit cost. A sale attracts commission, so the price has to be high enough that what remains after commission still covers what was spent. Dividing the unit cost by one minus the commission rate produces that price, and the result is usually higher than buyers expect, because commission is charged on the sale rather than on the cost. A commission of one tenth, for example, lifts the required price by about a ninth on top of the unit cost, which is the step people most often perform in the wrong direction.

Unit margin is the third column and the only one worth comparing across categories. It subtracts the unit cost, the commission on the chosen price, and the returns allowance for that category. Because the allowance is a share rather than a fixed sum, a high-return category needs a visibly higher price to produce the same margin as a low-return one at identical cost, which is the effect most often omitted from a plan. Two candidates with matching unit costs can therefore finish with different margins, and the difference is usually the return share rather than anything about the goods.

The reading order is fixed for a reason. Check the category's return share before choosing a price, then choose the price, then compare margins across candidates. Working in the other direction, by fixing a target margin and searching for a price that delivers it, produces a plan that quietly assumes nothing ever comes back. The tool prints all three columns at once so the sequence can be followed rather than guessed at, and following it removes the commonest error in resale planning, which is a price chosen before the cost was known.

Open the listing index

Outlier explanation

An outlier here is a row whose sample is small or whose recorded share sits far from the rest, and both conditions are flagged rather than smoothed away. Fragrance rests on 12 recorded orders, the smallest set carried, so its ten-in-a-hundred share should be read as a direction rather than a rate. Electronics sits at the opposite end, holding the highest recorded share from a modest sample of 29. Both rows stay in the table for that reason: a thin sample is an argument for holding a figure loosely, not for deleting it. A row can be flagged for a second reason as well, which is a share that has changed direction between two readings, since a shrinking figure is as informative as a growing one.

Two explanations cover most of the variation. The first is physical handling: boxed watches and small devices are examined at the door more often than a folded jersey, so a marginal unit generates more queries before it is kept. The second is the cost of a failure, since a device arriving with a fault cannot be repaired locally at a sensible price, which pushes more of those orders towards a return or a partial settlement than orders where a small flaw is tolerable. Neither account can be tested with the data this desk holds, so both are offered as plausible rather than as established.

There are situations where the table should be ignored. A single item bought for personal use has no unit margin to calculate, and the returns allowance is irrelevant to somebody who intends to keep the goods whatever arrives. Gifts are the same case with less tolerance for delay. The table earns its place where several units are bought against an expected sale, because that is the only situation in which a share recorded across many orders predicts anything about one purchase. Applied to a single item, it describes a population the buyer is not sampling from.

Rechecking matters more here than elsewhere in this cluster, because the return shares and the loss share both carry medium review windows and the samples behind them are small. Re-read the parameters and their dates before pricing a batch, and treat any row whose sample has grown since the last reading as a new figure rather than as confirmation of the old one. Outliers stay recorded as outliers in the tool instead of being averaged into one comfortable number, and a figure that has moved twice is worth writing down again with the new date beside it.