Gross sales vs net sales: the number that tells you if you actually grew

There is a conversation that repeats itself every quarter inside marketplace businesses. Someone opens the Amazon dashboard, opens the MercadoLibre one, adds them together and announces the quarter closed twenty percent up. Everyone is pleased. Three weeks later the accountant sends the income statement and profit barely moved. Nobody understands what happened, and the most comfortable explanation — “costs got away from us” — is almost never the right one.

What usually happened is that growth was measured on gross sales. And gross sales is a number that includes orders cancelled before they left the warehouse, orders the customer returned and got fully refunded, and partial discounts you granted after the sale to avoid a return. None of those pesos ever existed. They are in the report because the report counts orders created, not money kept.

The gap between gross and net is not an accounting technicality. It is the difference between believing your business grew and knowing whether it did. And on Mexican marketplaces, where a chunk of returns resolves weeks after the original sale, the gap between the two figures is wider and slower to notice than on almost any other channel.

This post is about how the net number gets built, why the timing lag keeps the illusion alive for months, and which concrete decisions break if you keep measuring with the wrong figure.

what exactly counts as a gross sale

A gross sale is the value of an order at the moment it is created. The buyer clicked, the marketplace generated the order, the system logged it with its price. That is all it takes to enter the report.

The package does not have to ship. The customer does not have to receive it. The customer does not have to keep it. An order cancelled forty minutes after creation still existed as a gross sale for those forty minutes, and in many reports it is still there at the end of the day.

This is not a marketplace defect. For the operations team, gross sales is the correct metric: it measures demand, converted traffic, whether the listing is working. It is the right yardstick for an ad, a title or a photo. The problem starts when that same number is used to judge the business.

There is also a vocabulary trap. “Sales” in a marketplace dashboard almost always means gross. “Sales” in an income statement means net. When an owner compares the dashboard to the statement and they disagree, the instinct is to doubt the accountant. The accountant is usually right: they are looking at the number that survived.

what gets deducted to reach net sales

Going from gross to net means subtracting everything that reverses an already-recorded sale. There are four families, and they are worth keeping separate because each one is fixed differently:

• Cancellations before shipment. The buyer changed their mind, or you cancelled because the stock was not really there. The second cause is the dangerous one: a stockout cancellation also costs you account metrics.
• Accepted returns. The product shipped, the customer sent it back and got refunded. It is not only the revenue that leaves: the outbound shipping goes, usually the inbound too, and part of the fees does not come back.
• Partial refunds. The classic “keep the item, I will refund twenty percent” used to avoid a full return. That is money going out that almost never gets recorded as a return, so it disappears into the noise.
• Later adjustments. Claims resolved in the buyer’s favor, damage-in-transit charges, chargebacks. They land weeks afterwards and slip into the account statement without ever touching the sales report.

None of the four is optional. A company that deducts returns but ignores partial refunds has an optimistic net figure, and the distortion grows precisely when it is growing fastest, because partial refunds are the favorite tool for holding up a reputation under volume pressure.

Glossary: real net margin, with everything deducted →

the example: twenty percent on gross, four percent on net

The numbers below are invented to illustrate the mechanics, not a market data point. What matters is not the figures but that the arithmetic closes and you can see where the gap comes from.

Last year’s quarter:

• Gross sales: $1,000,000
• Cancellations before shipment: −$20,000 (2%)
• Accepted returns: −$70,000 (7%)
• Partial refunds and adjustments: −$10,000 (1%)
• Net sales: $900,000

Same quarter this year:

• Gross sales: $1,200,000 (+20%)
• Cancellations before shipment: −$48,000 (4%)
• Accepted returns: −$204,000 (17%)
• Partial refunds and adjustments: −$12,000 (1%)
• Net sales: $936,000 (+4.0%)

The gross growth is real: two hundred thousand pesos more in orders came in. The net growth is real too, and it is four percent. The total deduction rate went from ten percent to twenty-two.

Why would it double? In this example, because the growth came from a new line — apparel, say — pushed with aggressive promotion on a channel where that kind of product comes back far more often than the original catalog. The business did not grow evenly: it grew by pushing volume into the category with the worst post-sale behavior.

Stated per order, the story is even clearer: out of every hundred gross orders in the new quarter, four cancel and seventeen come back. Seventy-nine survive. If each one is worth a thousand pesos of base revenue, you collected seventy-nine thousand pesos, not a hundred thousand. And against that seventy-nine thousand you paid commissions, fulfillment and advertising on all one hundred.

There is the part that hurts most: costs are incurred on gross and revenue is collected on net. The commission on the returned order does not come back in full, the outbound shipping was already paid, the ad click that generated it was already charged. That is how a business can be up twenty percent on gross, four on net, and down on profit.

why the illusion lasts for months

If returns happened on the same day as the sale, nobody would fall for it. The problem is the lag.

An October sale can be returned in November and settled in December. Meanwhile, the October report closed with the full gross figure. When the return finally lands, the system records it in the month it occurred, not in the month of the original sale. The result is that October looks spectacular for weeks and November looks worse than it was, because it carries returns that do not belong to it.

The effect is amplified in peak season. A high-volume November drags returns into December and January. If you measure December on gross sales and subtract the returns that arrived in December, you are mixing two different periods and the number means nothing.

The correct fix is to attribute the return to the date of the original sale, not to the date of the refund. That requires your system to keep the link between the refund and the order that caused it. Almost no spreadsheet does, because the spreadsheet is built from two separate exports that do not share the order identifier.

which decisions break when you measure on gross

This is not cosmetic. Four decisions get made badly, and all four cost money:

• Inventory replenishment. If your forecast is fed by gross units, you are buying units that will come back. On a product with a seventeen percent return rate, one in six pesos of your purchase order finances a round trip.
• Ad budget. An ACoS computed on gross sales looks better than it is. If seventeen percent of that attributed revenue comes back, the real ACoS on money kept is materially higher, and campaigns that looked healthy are losing.
• Price. A product that “sells a lot” at a low price may be selling a lot because the low price attracts the buyer who returns. Dropping the price further makes both problems worse at once.
• Sales team commission. Paying on gross means paying for orders that reversed. It is the wrong incentive written straight into payroll.

Glossary: forecast, the projection that drives your purchasing →

how to build the net report without losing your mind

You do not need a full accounting system to have the number. You need three things, in this order.

First, an order table with its final status, not its status at the moment of export. An order created on October 3 and returned on the 20th is not “delivered”: it is returned, and its row has to say so. That means re-exporting history, not just the new month.

Second, the link between each refund and its order. Amazon and MercadoLibre both give you the order identifier on refund movements. If your process stops there and only totals the month’s refunds, you lost the ability to attribute and you will never close the timing gap.

Third, one written definition of net sales, shared by operations, marketing and accounting. The most practical version: net sales is the pre-VAT selling price of orders the customer kept, minus partial refunds and adjustments, all attributed to the date of the original order. Write it down, post it, and let nobody report growth on any other basis.

With those three pieces, the report builds itself and net stops being a quarter-end estimate.

how this reads in iqseller

The Profitability module starts from the order, not from the channel aggregate. Each order carries its price, its final status and the movements that belong to it: Amazon settlement commissions, MercadoLibre order commissions, FBA or Full fees, shipping, advertising, and the VAT and withholding breakdown. When a return comes in, it attaches to the order that caused it, so the period in which the sale happened corrects itself instead of polluting the following period.

COGS is loaded by you per SKU, and it is what lets net go past “revenue kept” all the way to net margin per SKU. That is where the pattern that matters shows up: which models have a net figure far below their gross, and whether those models are exactly the ones you are pushing with advertising.

The Parent → Model → SKU tree lets you see the effect where it starts. One specific size of a t-shirt can drag the whole line, and at parent level that is invisible. Dropping one level is usually enough to find the culprit.

The Forecast module consumes the same data. If the replenishment projection runs on units kept instead of units ordered, purchase orders stop financing returns and tied-up capital falls without anyone having to tighten anything.

the number worth tracking every week

If you keep one practice out of all this, make it this one: every week, next to gross sales growth, report the deduction rate — cancellations plus returns plus partial refunds, over gross — and net growth.

Two lines. If both rise together, you grew. If gross rises and net does not, the growth is coming from catalog, category or traffic quality, and the diagnosis is in which of the three components of the rate moved. If gross falls but net holds, you probably cleaned out bad sales and the quarter is better than it looks.

That last case is the hardest to defend in a meeting and the most valuable to the business. Having the net number is the only thing that lets you defend it with evidence instead of intuition.