Discounts, vouchers and what they really cost

Discounting is the most-used lever in Malaysian ecommerce and the least measured. Nearly every store runs promotions continuously — festive campaigns, first-order codes, bundle pricing, free shipping thresholds — and very few can say what any of them earned.

The reason is that the cost of a discount is easy to calculate and easy to calculate against the wrong baseline.

A discount comes off margin, not off revenue

The arithmetic that surprises people.

A ten per cent discount on a product carrying a thirty per cent gross margin does not reduce your profit by ten per cent. It reduces it by a third, because the discount is deducted entirely from the margin. The cost of goods does not fall.

So on thin-margin lines a modest-looking discount can consume most of the profit, and on some products a discount that looks routine sells the item at or below cost. The percentage that matters is the discount as a share of margin, not as a share of price — and it is almost never presented that way.

This is why cost data has to exist before promotions can be evaluated at all — see cost of goods sold for Shopify stores.

The baseline problem

The genuinely hard part, and where most promotion analysis goes wrong.

A campaign runs, sales rise, and the campaign is judged a success. But some of those sales would have happened anyway, at full price, to customers who were going to buy. The discount on those is pure cost with no benefit.

So the honest question is not what did we sell during the promotion, but what did we sell that we would not otherwise have sold, and did that exceed the discount given on everything.

That is harder to answer and it is not unanswerable. Comparable prior periods, products excluded from the promotion as a rough control, and the pattern of your own repeat customers all give you something. The important discipline is asking the question at all, because measuring gross sales during a discount period will always look encouraging.

Two specific traps.

Pull-forward. A promotion moves purchases that were coming anyway into the discount window. Sales rise, then fall after, and the net effect over both periods is a discount given for nothing.

Discount dependence. Run promotions frequently enough and customers learn to wait. The discounted price becomes the real price and full-price sales stop happening — see how to grow Shopee sales without killing your margin for the marketplace version of the same dynamic.

The types, and how each behaves

Percentage off. Scales with basket, so the cost is unbounded on large orders. Simple, effective, and expensive on your best customers.

Fixed amount off. Cost is capped, and it is punishing on small baskets — a fixed discount on a low-value order can exceed its entire margin. Usually needs a minimum spend.

Free shipping. Costs you the courier charge rather than a share of margin, so its cost varies by destination and weight rather than by order value. Frequently the cheapest effective promotion on light goods and the most expensive on heavy ones — see shipping revenue versus shipping cost.

Bundles. Margin depends on the mix actually bought, so a bundle can be profitable on one combination and not on another. Needs modelling per combination rather than in aggregate.

First-order codes. Cost is an acquisition cost and should be judged against what the customer goes on to spend, not against the first order. This is the one discount type routinely judged too harshly.

Each of these needs recording distinguishably. Lumped into one discounts figure, none of them can be evaluated, and the mix between them shifts constantly.

Where they land in your books

Discounts reduce revenue. They are not a marketing expense, and recording them as one overstates both revenue and expenses while leaving margin unchanged. Gross sales less discounts is your actual revenue.

Allocate order-level discounts to lines using the allocation Shopify recorded, so product-level margin reflects what was actually earned — see Shopify discounts, taxes and line-level allocation.

Keep free shipping separate, because its cost is a courier charge rather than a revenue reduction.

Tag the campaign. A discount recorded without knowing which promotion produced it can never be evaluated. This is the single most useful thing to get right and it costs nothing at the point of setting the promotion up.

The four questions worth answering per campaign

Simple enough to run after every promotion, and almost nobody does.

Total discount given. The gross cost.

Incremental sales, against a comparable prior period. The gross benefit.

Net margin effect. Incremental margin earned less total discount given. This is the answer.

What happened in the two weeks after. Whether the promotion created demand or moved it.

Four figures, and they turn promotion from an article of faith into something with a track record. Most stores find that a minority of their campaigns generate the majority of the benefit, and that at least one recurring promotion has been losing money for a long time — see measuring whether AI accounting worked.

Common questions

How much does a discount really cost?

Far more as a share of profit than as a share of price, because the discount comes entirely off margin while the cost of goods is unchanged. A ten per cent discount on a product carrying thirty per cent gross margin removes a third of the profit, and on thin-margin lines a routine-looking discount can sell the item at or below cost.

Why is measuring a promotion difficult?

Because some of the sales during it would have happened anyway at full price, so the discount on those is pure cost. The honest question is what you sold that you would not otherwise have sold, and whether that exceeded the discount given on everything. Measuring gross sales during a discount period will always look encouraging.

Should discounts be recorded as a marketing expense?

No. Discounts reduce revenue, so gross sales less discounts is your actual revenue. Recording them as an expense overstates both revenue and costs while leaving margin unchanged, and it separates the discount from the sale it was given on, which makes campaign-level evaluation impossible.

What is the most important thing to record about a discount?

Which campaign produced it. A discount recorded without that attribution can never be evaluated, and tagging costs nothing at the point the promotion is set up. Without it, the total discounts figure mixes percentage offers, fixed amounts, free shipping, bundles and acquisition codes, which behave completely differently.

Related: cost of goods sold for Shopify stores · Shopify discounts, taxes and line-level allocation · running profitable Shopee promotions