Discount-Adjusted Profitability

Discounts are approved in percentage-off terms and paid for in margin. Because those two percentages are different, and nobody calculates the second one, promotions get authorised on a number that understates their cost by roughly half.

The correction is a single line of arithmetic, and it changes how the conversation goes.

Why a discount costs more than its percentage

A discount reduces price. It does not reduce cost of goods, shipping, packaging or fulfilment, and it barely reduces payment fees.

So the entire discount comes out of the margin. If a product carries a sixty-one percent contribution margin and you take twenty percent off the price, you have not given away a fifth of something — you have given away roughly a third of the only part that was yours.

The general form: the share of contribution margin surrendered equals the discount rate divided by the contribution margin rate. Twenty percent off a sixty-one percent margin gives up thirty-three percent of profit. The same twenty percent off a forty percent margin gives up half of it.

This is why thin-margin categories cannot run the promotions that thick-margin ones can, and why copying a competitor’s discount depth without knowing their cost structure is a reliable way to lose money.

Stacking, and the discount you did not intend

The advertised rate and the effective rate diverge more often than most stores realise.

Codes that stack with an automatic discount, a sale price, or each other. A twenty percent code on a product already reduced fifteen percent is not a twenty percent promotion.

Free shipping bundled in. Adding free delivery to a discount code adds the full label cost, which on a heavy item can exceed the discount itself.

Loyalty and referral credits applied on top.

Post-purchase goodwill refunds, which are discounts issued after the margin was already booked.

The check is to compare the effective discount rate — total discounts divided by gross sales — against the rates you believe you are offering. When the effective rate is higher, something is stacking, and the gap is usually concentrated in a handful of codes.

Why promotions sell your worst products

Discount codes do not distribute evenly across a catalogue. Price-sensitive buyers move toward the cheapest items, and the cheapest items usually carry the thinnest margins.

The result is a mix shift running in the worst possible direction: a sitewide promotion lands hardest on the products least able to absorb it. Store-level margin falls further than the discount rate alone would predict, and the cause is invisible unless margin is tracked per code and per product.

The fix is rarely to stop discounting. It is to scope the discount — excluding the thinnest-margin products, or applying it only to categories that can carry it. Same conversion mechanism, materially different cost.

Judging whether a promotion paid

The comparison that matters is not revenue during the promotion versus revenue before it. Revenue almost always rises, which is why promotions are almost always declared successful.

The right comparison is total contribution margin in the promotional period against a matched baseline period — same weekday pattern, similar advertising, adjusted for seasonality where you can.

Three outcomes are possible. Margin rose: the extra volume more than covered the discount, and the promotion worked. Margin held flat: you sold more, worked harder, and ended in the same place. Margin fell: you paid customers to buy things they would have bought anyway.

The third outcome is far more common than most stores believe, and it is entirely invisible in a revenue chart.

Setting a discount floor

Most stores approve promotions case by case, which is how depth creeps upward over time — each individual decision is defensible and the trend is not.

A floor fixes it in one line: no discount may take contribution margin on an order below a stated percentage. Not a cap on discount depth, which ignores that different products can afford different things, but a floor on what has to survive.

The advantage is that it scales automatically. A high-margin product can carry thirty percent off and stay above the floor. A thin-margin one cannot carry ten. Nobody has to calculate anything per promotion; the rule does the work, and it stops the specific failure where a sitewide percentage is applied to a catalogue that varies enormously in what it can absorb.

Two practical additions make it hold. Set the floor above break-even, not at it, so a promotion that overruns by a few days does not immediately cost money. And check effective rather than advertised depth when applying it, because stacking is what usually breaches a floor that everyone believed was being respected.

The customer-acquisition case

There is a legitimate version of running a promotion at a margin loss: acquiring customers who buy again.

A welcome discount is acquisition cost by another name, and it belongs in CAC alongside media spend. Judged that way, the question becomes the ordinary one — does the twelve-month value of the acquired customer exceed what it cost to acquire them, including the margin given up on their first order.

That framing also disciplines the decision. It rules out the sitewide promotion that discounts existing customers who were going to buy at full price, because those are not acquisitions and the margin surrendered on them buys nothing at all. Scoping a discount to first-time buyers is usually the single highest-return change available to a store that promotes frequently.