Insights · Pricing

The discount nobody costed: where margin actually goes

A rep asks for two points of discount. The sheet says gross margin is 45 percent, so two points looks like nothing. The sheet is not wrong about arithmetic. It is wrong about which number it is doing arithmetic on.

Gross margin is the single most misleading term in any conversation about price. Not because it is wrong, but because it is incomplete while still functioning as the basis for decisions. It takes a price, subtracts cost of goods, reports a number. Everything that happens after the invoice - the cash discount for paying early, the freight you absorb, the returns, the expedited shipments, the POS material sent to that one account because they asked - never enters it. Those items come out of profit, and they appear in no file that the discount conversation is based on.

List price is a fiction

Between the list price and the money that actually stays in the business sits a waterfall of deductions. Trade discount, volume rebate, listing fees, an annual rebate calculated on turnover, a cash discount for payment terms, the cost of delivering, the cost of accepting a return, the cost of servicing the account. Each was agreed separately, by a different person, in a different year, for a different reason. Almost nobody has ever added them up for a single customer.

The chart above is not an extreme case, it is a normal one. From a list price of 100, what reaches the pocket is 16.5. The sequence matters: the first four deductions are visible on the invoice, so they enter the system and somebody polices them. The next four appear in no pricing report at all, because they post as operating costs in an entirely different part of the P&L. Together they take 8 points - half of what finally remains.

Three levels of the same illusion

The same transaction carries three different margins, depending on where you start counting. Margin on list: list price minus cost of goods, 45 percent in this example. That is how most boards think about "margin on the product". Margin on invoice: price after the visible deductions minus cost, just under 31 percent. That is what most ERP systems report. Pocket margin: the same thing, but after everything that happens off the invoice - 16.5 percent. That one is real.

A discount does not take a percentage off the price. It takes a percentage out of what is left. And because what is left is three times smaller than the first view suggests, every point hurts three times harder.

Discount	Of margin on list 45%	Of margin on invoice 31%	Of pocket margin 16.5%	Volume increase needed
1 point	2.2%	4.1%	6.1%	+6.5%
2 points	4.4%	8.2%	12.1%	+13.8%
5 points	11.1%	20.4%	30.3%	+43.5%
8 points	17.8%	32.7%	48.5%	+94.1%

Last column: how many more units you have to sell to end up with the same absolute profit after the discount. Calculated on pocket margin, because that is the only one that is real. Waterfall structure as charted above.

Follow one row across. Eight points of discount costs 18 percent of profit in the first column - acceptable, if the customer promises a bit more volume. In the last column it costs almost half, and needs sales to nearly double. Nobody ever promises to double volume for eight points. Yet those discounts get granted every month, because the file in front of the person deciding shows the first column.

A discount is not a pricing tool. It is a loan a business takes against its own margin - and almost nobody calculates the interest.

Your biggest customer is often your worst

Customer rankings get built on revenue, because revenue is easy to pull. It is one of the most expensive habits in distribution. The large account usually has the deepest discounts, the longest terms, the most returns, the highest service expectations and the strongest negotiating position in every subsequent conversation. The smaller account pays list, collects on its own truck and asks for nothing beyond an invoice.

Recalculate customer profitability at pocket level and the order frequently inverts, with a distribution that is more uncomfortable still: roughly 20 percent of accounts typically consume most of the profit generated by the rest. The point is not to fire them. The point is to know they exist before somebody grants them another point for loyalty.

Discount discipline, not a discount ban

Banning discounts is as naive as leaving them unmanaged. A discount earns its keep when it buys something you would not otherwise get: a first order from an account that is testing you, entry for a new product nobody knows, volume taken off a competitor, stock cleared before year end. It stops earning its keep the moment it subsidises a purchase that was going to happen anyway.

In practice, every point of discount needs three things. A reason recorded in the system, not agreed on a phone call. A threshold, meaning a condition that, unmet, voids the discount entirely. And an expiry date, because a discount without one becomes the new list price within two quarters, only nobody announced it. An unconditional, open-ended discount is not a discount, it is a price cut - and price cuts, as we have argued before, do not reverse.

How we count it

This is work that sits across two of our lenses. Pricemore rebuilds the price waterfall all the way down - invoices, credit notes, logistics cost and the real cost to serve - and shows pocket margin per customer, per product, per order, instead of an average that describes nobody. Cashstream, our portfolio P&L lens, shows what that structure does to profit across the whole range: where a mix shift beats a price move, and which volume is worth defending versus which merely looks large. The output is not a new price list. The output is a list of discounts with a reason, a threshold and a date - plus a short list of the ones that survive none of the three.