A seller who says the business is profitable has made a claim about one report, covering one period, at a level of aggregation they chose, using a cost basis they may not have checked. Each of those choices can hide a catalog where forty products lose money and nobody has looked.
The word should be retired from operating conversations and replaced with three specific questions.
Q: Isn’t profitable simply what the profit and loss statement says?
It is what the P&L says, which is a different thing from what the business is doing.
A P&L is an aggregate. It sums every SKU, every channel and every customer into one line. Aggregation is the entire problem: a catalog of two hundred products where forty lose money and one hundred and sixty make money reports as profitable, and the report gives you no reason to find the forty.
Worse, the aggregate hides the direction of travel. If your losing products are your growing products, the P&L improves every month while the underlying mix gets worse, and the reversal arrives without warning.
Q: What is the first question that should replace it?
Profitable at what level of aggregation?
The only unit of analysis that means anything in ecommerce is the SKU, because pricing, fees, freight and returns all attach to the SKU rather than to the company. A business is a portfolio of products, and portfolios are evaluated position by position.
Ask any seller for the net margin on their fifth-best-selling product last month. Most cannot answer without building a spreadsheet, and the spreadsheet usually turns up at least one top-ten seller sitting below breakeven.
Q: And the second?
Profitable after which fees, exactly?
Marketplace fee structures are less intuitive than sellers assume. eBay’s published selling fees page puts most categories at 13.6% of the total sale amount up to $7,500 per item plus a per order fee of $0.40, or $0.30 on orders at or under $10.00. The detail that matters is eBay’s own definition of the base: the total sale amount “includes the item price, any handling charges, any shipping costs collected from the buyer, sales tax, and any other applicable fees.”
The percentage applies to the shipping you collected and to the sales tax you never kept. A seller modeling 13.6% against the item price alone understates fees on every order, and the error grows with how much shipping they charge, which means the SKUs with the heaviest boxes carry the biggest hidden gap.
Amazon has its own version of this, with referral fees varying by category and separate fulfillment and storage charges that never attach to an order in the settlement data. The published schedule sits on Amazon’s selling fees page and the percentages change, so any figure written down needs a date attached.
Q: What is the third question?
Profitable before or after returns?
This is where most product-level margin math falls apart. The National Retail Federation, in its 2025 returns report produced with Happy Returns, estimated that 19.3% of online sales would be returned in 2025, against 15.8% across retail overall.
And the refund is not the whole cost. Appriss Retail’s 2026 Total Retail Loss Benchmark Report found that returns cost retailers an average of 30% of an item’s value to process. So a product with a 12% net margin and a 25% return rate is not a 12% margin product. It is close to nothing, and possibly negative, depending on how much of the returned stock comes back sellable.
A seller reporting that product as profitable is quoting a number computed before the largest variable cost in the business.
Q: Are you saying most ecommerce businesses are actually unprofitable?
No. Plenty of these businesses are profitable in aggregate, and the aggregate figure is the right one to report to a lender or a tax authority.
My claim is narrower: the aggregate figure is not actionable, and treating it as actionable is what produces the pattern where a business grows revenue for three years and never grows cash. If you cannot see which products carry the company, you cannot allocate inventory dollars toward them, and inventory allocation is the main operating decision an ecommerce business makes.
Q: Why does this go undiagnosed for so long?
Because the data required to answer these questions sits in three systems that do not talk.
Fees live in settlement reports. Cost of goods lives in a purchasing spreadsheet or an inventory system. Returns live in the marketplace’s returns dashboard. Producing per-SKU net margin means joining all three at the unit level, and most sellers try it once, discover it takes a weekend, and go back to the aggregate.
This is the entire reason a category of software exists. Products in it, ConnectBooks among them, post marketplace settlement detail into QuickBooks or Xero at the item level and attach cost of goods and inventory movement to the same units, which turns SKU-level net margin from a weekend project into a report. Whether you buy a tool or build the join yourself matters less than accepting that the number does not compute itself.
Q: What should a seller actually do on Monday?
Pick your top twenty SKUs by revenue. For each one, compute four numbers: gross revenue, total marketplace fees using the actual fee base rather than the item price, landed cost of goods including inbound freight and duty, and return rate.
Then compute net margin per unit and sort ascending. You are looking for the products near the top of the revenue list and near the bottom of the margin list, because those are the ones consuming your working capital while contributing nothing.
Twenty SKUs takes an afternoon. It is not elegant and it is not automated, and it will change what you buy next quarter more than any dashboard will.
Q: What replaces the word?
Say which SKUs, after which fees, after returns. Three qualifiers, and the sentence becomes checkable.
“We are profitable” is a claim nobody can falsify, which is what makes it comfortable to say and useless to act on. “Forty of our two hundred SKUs are below breakeven after returns, and three of them are in our top twenty by revenue” is a sentence that produces a decision about next quarter’s purchase orders by Friday.