Answer in brief
A step-by-step walk through the unit economics of a marketplace product, from the realised selling price down to the contribution per unit, with the two lines sellers forget most often: returns and storage on slow stock.
What is left of the price: the order of subtraction on one page
Short answer: from the price the buyer actually paid you subtract the marketplace commission, fulfilment, delivery, returns, storage, advertising, tax and the cost of the goods. Profit is whatever survives all eight lines, not the gap between your purchase price and your price tag.
The order matters because some of those lines scale with the price and some do not. Commission and tax move with what the buyer paid. Handling, delivery and storage do not: they are the same for a cheap unit and an expensive one of identical size.
That one fact explains most of the unpleasant surprises. A low-priced item of the same weight and volume hands a far larger share of its price to the platform, not because the tariffs differ, but because fixed per-unit costs are divided into a smaller number.
The rest of this article walks the lines in the order they are subtracted, with extra attention on the two sellers skip most often: returns and storage on slow stock. Neither shows up in a sales report, and both take money after the stock has been paid for.
The price you actually receive, not the one printed on the card
Start from the price the unit genuinely sold at, not the price you typed into the listing. Between the two sit discounts, platform campaigns, promo codes and sale events, and every one of them shrinks the base that all the other lines come out of.
When you join a campaign the selling price falls while most costs stay exactly where they were: logistics, handling and storage know nothing about your discount. That is how a promotion pushes a listing into loss while the headline markup still looks perfectly healthy.
Commission is normally charged on the final amount the buyer paid, but the rules differ by platform and by category, including how a co-funded discount is treated. Read your own contract and your own category tariff rather than carrying an assumption over from another marketplace.
The practical step is to export real reports for a period and derive the average realised price per SKU. It is almost always lower than the card price, and it is the number that belongs at the very top of the calculation.
Cost of goods: the purchase price plus everything that got it to the warehouse
Cost of goods is not the number on the supplier invoice. Add inbound freight to your own warehouse, customs duties and fees where they apply, packaging, labelling, and the defective units you discovered when you accepted the delivery.
Defects are better spread across the sellable units of the batch. If part of a carton never reaches the shelf you still paid for it, and that money has to be carried by the pieces that will actually be sold to somebody.
If you buy in another currency, fix the rate at the date you paid for the batch rather than the date of sale. Otherwise your cost base drifts and you lose the ability to say which batch was profitable and which quietly was not.
What you end up with is the cost of one sellable unit sitting in your warehouse. Everything described after this comes out of the selling price, while the cost of goods waits patiently for its turn at the end of the calculation.
Marketplace commission: a percentage that depends on the category
Commission is a share of the sale, tied to the product category and sometimes also to the model you work under: the platform warehouse, your own warehouse, or shipping direct from your own stock. The same product can cost you a different share in two categories.
So a category should be checked against the tariff table, not only against search logic. A wrong category damages the visibility of the listing and changes the percentage you hand over on every paid order, both at the same time.
Tariffs get revised, and that is a normal part of trading on someone else's platform. A spreadsheet built once and left alone for six months stops describing reality, so tie the recalculation to the moment new terms are published.
Commission is the one line that grows with price almost linearly, which makes it predictable. For exactly that reason it is the least dangerous entry in the model: this is the line sellers get wrong least often.
Fulfilment: pick, pack and handling that ignore your price
Order handling — receiving, put-away, picking, packing — is charged per unit or per shipment and driven by dimensions, not by price. For a small, cheap item this line can easily turn out to be the heaviest one in the entire calculation.
Measure the dimensions in the packed state instead of copying the manufacturer specification. A couple of extra centimetres can move the item into the next tariff band, and you pay that difference on every single unit you ship out.
Bundles change the picture in both directions: one shipment instead of three saves handling, but adds weight and volume. Model the specific bundle with its real measured dimensions rather than adopting bundling as a general principle.
There are quiet savings hiding here as well. Reworking the packaging sometimes drops an item into a cheaper band without touching the product itself, one of the rare cases where physical work pays back faster than listing work does.
Logistics: delivery to the buyer and shipment into the warehouse
Logistics comes in two halves: delivery of an order to the customer, and shipment of a batch into the platform warehouse. The first is attached to a shipment, the second to a batch, and it has to be spread across every unit that arrived in it.
Inbound shipping is frequently left out altogether because it was paid up front as a single amount. Divide it by the number of sellable units and add it to the per-unit cost, or the model will overstate the remainder on every SKU you own.
Geography matters here. Delivery to distant clusters costs more, and the way you distribute stock across warehouses changes both cost and delivery speed. That decision moves unit economics more than most of the edits people make to a description.
If you ship from your own warehouse, part of this simply becomes your own cost instead: couriers, boxes and staff time. Those costs do not disappear merely because no separate invoice arrives from the platform for them.
Returns: the line sellers forget first
A return is not a cancelled sale, it is a separate operation with a cost of its own. You pay for the outbound trip, the trip back, a second round of handling, and sometimes the loss of sellable condition that stops the unit selling at full price.
That is why returns concentrate where the buyer cannot verify the product before it arrives: clothing and footwear with size charts, items where the exact shade matters, hardware with non-obvious compatibility. Trying something at home costs the buyer almost nothing, so they will.
Returns have to be counted per SKU, never per shop. A store-wide average hides the one listing that comes back far more often than everything else and single-handedly consumes the remainder of an entire category.
The good news is that this line responds to work. An accurate size chart, honest photography, a plain description of the material and explicitly stated compatibility limits reduce the share of orders that were placed as a guess in the first place.
Storage: the second forgotten line, and it lives on slow stock
Storage is charged per unit per day, which makes it a function of time rather than of sales. A unit sitting in the warehouse for three months pays for all ninety of those days whether or not anybody ever buys it at all.
The mechanics are unforgiving: the worse an item sells, the more it costs you. Slow stock does not merely fail to earn its keep — it subtracts money daily from the results of the SKUs that are actually working.
Sellers miss this line because it arrives as one lump sum for a period and is attached to no individual order. It is invisible in the sale report and appears in the monthly statement, long after the purchasing decision has been made.
Allocate storage to SKUs in proportion to the volume they occupy and the days they occupy it. Once you do, it becomes obvious which items deserve a markdown and removal and which genuinely earn the shelf space they are taking.
Advertising: a cost per unit sold, not a cost per click
Advertising enters unit economics as a cost per unit sold, not as a cost per click. Divide the budget spent on a SKU by the number of its sales in the same period, and you have an honest deduction line to work with.
Count it against organic sales as well: some orders arrive with no direct click on the ad, because the buyer saw the listing and came back later. Attributing only the clicked orders flatters your acquisition cost and hides the real number.
The bid at which a listing stays profitable comes out of the remainder on that SKU, not out of a market average. If little is left after everything else, your ceiling is low, and rewriting the ad creative will not change that arithmetic.
Advertising a loss-making listing only accelerates the loss. Fix the economics of the position first and open the budget second; the reverse order costs more and delivers exactly the same conclusion, just later and with less money.
Tax and settlement: the last subtraction before the remainder
Tax depends on your regime and your jurisdiction: in some it is calculated on turnover, in others on the difference between income and costs. The common part is that it leaves before the remainder can honestly be called profit and reinvested.
Platforms usually withhold their fees from revenue and transfer the difference to you. The amount landing in your account therefore looks like income when it is already a net figure, while your taxable base may still be the full sale price.
Payment processing and other service deductions deserve a line of their own too. Small percentages are invisible on a single order and extremely visible across a month of orders stacked one after another.
There is a neighbouring problem worth naming: money for goods sold arrives with a delay, while stock and advertising are paid earlier. A profitable SKU can still demand working capital that you do not currently have available.
The remainder: what it has to cover and what it tells you to do
What survives every deduction is one unit's contribution towards your fixed costs: people, software, rent, photography and the tax on profit itself. Only once all of those are covered does anything that deserves the name profit begin.
Track the amount per unit, not only the percentage. A high-margin SKU that sells three times a month can deliver less actual money than a dull one with a modest remainder and a steady, predictable turnover behind it.
Decisions after the calculation usually reduce to four: raise the price, cut a per-unit cost, fix the listing so volume grows, or take the item out of the range. Selling more of something that loses money is not a fifth option.
Rebuild the table whenever tariffs, purchase prices, exchange rates or packaging change. Unit economics is a recurring procedure tied to events, not a file you built once and have been trusting ever since without looking.
When to hire help, and what the VITON13 packages actually include
Outside help makes sense once the calculation has already shown the problem and you have neither the time nor the hands to fix it. VITON13 keeps four packages and one express format for exactly this on /services/marketplaces.
The Demand Audit costs $50 and takes 1-2 working days: a findings document plus one round of revisions on it, with access to your seller account not included. Listing Refresh is $70 over 2-3 working days with one round, excluding photography, ad spend and account management.
The Marketplace + SEO System is $110 and runs 3-5 working days with two rounds on listings and copy, excluding photography and paid placement. Marketplace + SEO Express costs $190 and is delivered in 3 working days with one round, also without photography or ad spend.
Managed Demand Growth is $200 per month on a monthly cycle with 30 days notice to stop, and advertising spend on the marketplace is not included. Which one fits is decided by your table: a one-off correction, or continuous work on the range.
Practical checklist
- Take the average realised selling price per SKU from your reports, not the price on the listing.
- Build the landed cost: purchase price, inbound freight, packaging, labelling and defects found at intake.
- Check the listing category against the commission tariff, not only against search logic.
- Spread inbound shipping across the sellable units of the batch and add it to the per-unit cost.
- Count returns and storage as separate lines for every SKU rather than as a store-wide average.
- Rebuild the table whenever tariffs, purchase prices, exchange rates or packaging change.
Questions and answers
Where do I start when calculating the unit economics of a marketplace product?
Start from the average realised selling price per SKU taken from your own reports. Then subtract commission, fulfilment, delivery, returns, storage, advertising and tax in that order, and place the cost of goods last.
Which lines do sellers forget most often?
Returns and storage. A return pays for both trips plus a second round of handling, while storage accrues per unit per day, so slow stock keeps taking money in a month with no sales.
What does the Demand Audit cost and what is in it?
The Demand Audit is $50 and takes 1-2 working days. You get a findings document with one round of revisions on it. Access to your seller account is not included in the package.
Why is a cheap product often unprofitable at a normal markup?
Because handling, delivery and storage are driven by size and time rather than by price. The same physical cost is divided into a smaller sale, so it consumes a far larger share of it.
How should advertising be counted if sales also arrive organically?
Divide the spend on a SKU by all of that SKU's sales in the same period. Ads also produce orders that arrive without a direct click, so counting only clicked orders distorts your cost per unit.