After a sale, most owners look at the revenue chart: a spike, more orders, a sense of success. That chart answers a different question from the one you need to decide. What matters is not how much sold during the promotion, but how much more money remained compared with what would have sold without it. Every buyer receives the discount, not only the ones the promotion brought in, and part of the spike simply borrows sales from other weeks or other products.
Why a discount needs so much extra volume
The number that decides a promotion is contribution margin: revenue minus the costs that rise with every additional sale. Contribution first covers fixed costs and only then becomes profit. A discount thins it on every unit, so the store has to sell more to cover the same fixed costs. A teaching note from the Yale School of Management accordingly advises resisting the temptation to use discounts as bait for new customers.
How much extra volume a discount needs is simple to calculate: required volume increase = discount / (margin − discount), both as a percentage of the original price excluding VAT. This illustrative table shows how much unit sales must rise for contribution to stay the same as without the discount:
| Product gross margin | 10% discount | 20% discount | 30% discount |
|---|---|---|---|
| 30% | +50% | +200% | impossible, margin gone |
| 40% | +33% | +100% | +300% |
| 50% | +25% | +67% | +150% |
The table is a lower bound. It ignores free shipping, payment fees and promotion advertising, all of which reduce contribution further. Above all, it assumes every extra unit is genuinely extra. That is usually the weakest assumption, as the next section shows. The wider breakdown of revenue growth into volume, price and mix is covered in Revenue is growing. Is the store earning more too?; this article deals with one part of it, the decision about discounts.
What hides inside the revenue spike
Promotion research has spent decades asking where the extra sales come from. An analysis of store scanner data split the sales spike into three parts: switching from other brands, sales borrowed from other periods and genuine growth in purchases. On average, each accounted for roughly a third.
A study of every promotion run by the US drugstore chain CVS in one year drew a similar picture. About 45% of the gross sales lift was genuinely incremental; the rest was switching between products in the store and purchases brought forward. Even so, more than half of the promotions were not profitable. The reason matters for any online store: the lower margin applies to all units sold on promotion, not just the extra ones. Purchases brought forward also show up as a dip after the promotion. It tends to be stronger for frequently promoted products and for promotions whose timing customers cannot predict. A dip can also appear before a promotion that customers are waiting for.
This research comes from US physical stores and fast-moving consumer goods, so do not transfer the specific shares to a Czech online store. The mechanisms hold online too: a customer who would have bought anyway receives the discount for nothing, a discounted product can take sales from another product in your store, and a purchase brought forward by two weeks is missing from the weeks that follow.
An illustrative example with invented numbers: a store normally sells 1,000 units a week at CZK 1,000 excluding VAT with a purchase cost of CZK 600, a 40% margin.
| Three normal weeks | Week with 20% off and the two weeks after | |
|---|---|---|
| Units sold | 3,000 | 1,800 + 850 + 850 |
| Revenue excl. VAT | CZK 3,000,000 | CZK 3,140,000 |
| Contribution after cost of goods | CZK 1,200,000 | CZK 1,040,000 |
The promotion week looks excellent: 80% more units and 44% more revenue. Over the full three-week window, however, revenue rose by less than 5% and contribution fell by CZK 160,000. Anyone looking only at the promotion week would run it again.
The long-term bill: expected prices and deal-driven customers
Frequent discounts also carry a cost that no single promotion reveals. In a controlled experiment, both the frequency and depth of discounts lowered the price customers expected to pay for a brand. Long-running panel data showed customers becoming more sensitive to price and promotions as promotions increased. A store can end up training customers not to buy at full price.
For new customers the evidence is mixed. Data from a newspaper and an online grocer showed that the deeper the discount used to acquire a customer, the less often that customer bought again. Customers acquired with a 35% discount had about half the long-term value of those acquired without one. Field experiments by a mail-order catalogue found the opposite for first-time buyers: a deeper discount increased their future purchases, while it reduced future purchases by established customers. Both results point to the same practice: the same discount works differently on new and existing customers, so evaluate them separately. A discount that merely subsidises purchases by loyal customers can still pay off with new ones, provided they demonstrably come back.
Judge promotions on margin, not revenue
Compare revenue excluding VAT after discounts, gross margin and new and returning customers during and after a promotion.
How to evaluate a promotion without being fooled by the chart
Choose a comparable baseline. The best one is a group that never saw the discount, for example a randomly selected part of your list that did not receive the discount email. Without one, compare a period of the same length and season without a promotion, and the same products outside the promotion. Be honest that such a comparison shows association, not proven cause: weather, season or a competitor’s campaign could also have lifted sales.
Count the full window. Include two to four weeks after the promotion, and for announced promotions the week before it. Only then can you see how much of the sales the promotion merely borrowed.
Measure contribution, not revenue. Revenue excluding VAT after all discounts and coupons, minus the cost of goods, free shipping, payment fees and promotion advertising. Also add up the margin given to customers who would have bought without the discount.
Split customers. How much of the discount went to new customers, to existing ones and to regular buyers? And how many new customers from the last promotion have since bought again at full price?
Then decide:
- Repeat when contribution over the full window rose and new customers are coming back.
- Narrow when only part of it paid off: higher-margin products, stock that is not selling, or new customers only. Consider a shallower discount too.
- Stop when contribution over the full window fell and the promotion mainly subsidised loyal customers or pulled their purchases forward.
The rule that also shapes how you announce a discount
Since January 2023, Czech law has required any information about a price reduction to state the lowest price at which the trader offered and sold the product in the 30 days before the reduction. The Czech Trade Inspection Authority treats general claims such as “sale” or “Black Friday” as discount information and states that the advertised reduction must be calculated from that lowest price. The rule applies to online stores and to discount codes offered to everyone, but not to offers such as “3 for 2” or to genuinely personalised discounts. The Court of Justice of the EU confirmed that the reduction must be calculated from the lowest 30-day price. Shoptet therefore calculates a discount from the difference between the Standard price and the Price, and states that the Standard price should contain the lowest price of the previous 30 days.
This has a practical consequence for strategy. If you run promotions often, the lowest price of the last 30 days falls, and with it the discount you may advertise. The inspection authority’s own example shows that after a short promotion at CZK 80 and a return to CZK 100, a further cut to CZK 50 may be advertised only as 37.5% off, not 50%. This is not legal advice; check your price settings before a major promotion.
What to do now
- Pick your last major promotion and define its full window: the week before it (if announced), the promotion itself and two to four weeks after.
- Find a comparable baseline. A period of the same length and season without a promotion, or a group of customers who did not receive the discount.
- Calculate contribution, not revenue. Revenue excluding VAT after discounts, minus cost of goods and the promotion’s variable costs, over the full window.
- Split the result into new and existing customers, and check whether new customers from the previous promotion bought again.
- Decide and record why: repeat, narrow or stop. Before the next promotion, use the table to work out how much extra volume it needs.
For any chosen period, Korzaro shows revenue excluding VAT as the actual order total after discounts and coupons, gross margin from purchase prices, margin after advertising, and orders and revenue against the previous period of the same length. The Customers section shows how many customers bought for the first time and what share of revenue comes from returning customers. You can compare the promotion window and the weeks after it without assembling numbers by hand. Korzaro does not calculate a promotion’s incrementality for you: the store itself has to define a group that did not see the discount.