Sales are slow, so the first idea on the table is a discount. It is quick, it is easy to announce, and it almost always moves something this week. That is exactly why it is used so often, and why so many businesses end up with customers who only buy on sale and margins that never recover.
A discount is not good or bad. It is a tool with a cost, and it is worth paying only when it does a specific job that nothing cheaper can do. This article gives the maths that most people skip, a table of when a discount helps and when something else works better, and a worked example of a store that was paying a lot to sell to people who would have bought anyway.
In short: A discount helps when it has one clear job, such as getting a first purchase from someone who would not otherwise try you, clearing specific stock, or filling quiet hours, with an end date and a number to check. It hurts when it is used to fix a problem that is not price, when it becomes regular enough that customers wait for it, or when the margin maths does not work. Do the maths first, give the discount a job, measure what happens after it ends, and reach for alternatives that protect the price.
The maths most people skip
A discount comes straight out of the margin, not out of the price. If a product sells at 100 with a gross margin of 40 (40%), a 20% discount leaves a margin of 20 per sale. To earn the same gross profit as before, the business has to sell twice as many units. The general rule is simple: the extra volume needed equals the discount divided by what is left of the margin after the discount.
| Gross margin before discount | 10% off | 20% off | 30% off |
|---|---|---|---|
| 30% | +50% units | +200% units | No profit left |
| 40% | +33% units | +100% units | +300% units |
| 50% | +25% units | +67% units | +150% units |
| 60% | +20% units | +50% units | +100% units |
These figures are arithmetic, not research: the extra units needed just to stand still on gross profit, before counting the cost of promoting the offer. Most owners who look at this table for the first time are surprised by the middle row. A "small" 20% discount on a 40% margin has to double sales to break even.
Why discounts often disappoint
Three well-known findings explain why the extra volume often does not come, or does not last.
- Most promotional sales go to existing buyers. Andrew Ehrenberg, Kathy Hammond and Gerald Goodhardt's study "The After-Effects of Price-Related Consumer Promotions" (Journal of Advertising Research, 1994), later summarised in Byron Sharp's How Brands Grow (2010), found that price promotions mainly sell to people who already buy the brand, and that sales return to normal afterwards with no lasting gain.
- Customers judge a price against the one they expect. Richard Thaler's "Mental Accounting and Consumer Choice" (Marketing Science, 1985) describes how part of a purchase's appeal is the "deal" compared with a reference price. When the sale price is seen often enough, it becomes the reference, and the full price starts to feel like a loss.
- Short bursts and long growth are different jobs. Les Binet and Peter Field's The Long and the Short of It (IPA, 2013) separates short-term sales activation, which produces quick spikes that fade, from brand building, which grows sales slowly and makes buyers less sensitive to price over time. A business that only discounts is doing the first job and paying for it in the second.
The reverse lesson matters too. When the US retailer J.C. Penney tried to drop its constant sales almost overnight in 2012, the change was widely reported as a failure. Customers trained to expect discounts do not unlearn it quickly. Moving away from heavy discounting has to be gradual and has to give customers another reason to buy.
A decision table: discount or something else?
| Situation | Does a discount help? | Often better |
|---|---|---|
| New customers will not try you | Yes, if first order only | A trial size or a no-risk first order |
| Old or seasonal stock | Yes, limited to that stock | Bundles with best sellers |
| Quiet hours or days | Yes, tied to those times | An add-on only available then |
| Sales slow, cause unknown | No, it hides the cause | Diagnose first |
| Competitors cut prices | Rarely | Show what your price includes |
| Reward loyal customers | Rarely | Early access, a gift, priority |
| Customers ask for "the last price" | No | Clearer value and tiers |
Running a discount that earns its cost
- Name the job. One sentence: who should do what, that they would not do without it. "Bring first orders from people who follow us but have never bought" is a job. "Boost sales" is not. Output: the job sentence.
- Do the margin maths. Use the table above with your real gross margin. Output: the number of extra sales needed to break even, written next to the job.
- Design the limits. Who qualifies (new customers only, a product line, a time slot), how long it runs, and a cap. A discount without limits is a price cut. Output: a one-paragraph offer with an end date.
- Measure before, during and after. Sales and orders for the four weeks before, the weeks of the offer and four weeks after, split by new and returning customers. Output: one small table. The "after" column is the one most people never fill in, and the most honest.
- Decide. Did it do the job, and did profit over the whole period beat the period without it? Output: repeat, change, or drop, written down for the next time someone suggests a sale.
A worked example
This example is hypothetical, and the numbers are illustrative. An online abaya store in Riyadh runs 30% off every month-end. The owner believes the sale is what keeps the business going, because month-end weeks are always the busiest.
The maths. With an illustrative gross margin of 50%, the table says a 30% discount needs about 150% more units just to match normal gross profit. Month-end weeks are busier, but nowhere near two and a half times busier.
The before, during and after table shows the rest. Illustratively, about seven in ten sale orders come from returning customers, and orders in the week before each sale drop noticeably: regular customers have learned to wait. New customers are a small share of the sale, so the discount is mostly a reward paid to people who would have bought anyway, which is exactly the pattern the Ehrenberg study describes.
The change. The monthly sale is reduced gradually, not cut overnight. In its place: a first-order offer only for new customers, early access to new collections for returning ones instead of a discount, and two clearly limited sales a year, before Ramadan and at the end of season, on specific older stock. The number to watch over the next three months is gross profit per month, not order count, alongside the share of first-time buyers.
Five mistakes we see most
- Discounting before diagnosing. It happens because a discount is the fastest lever available. If the real brake is slow replies or a weak offer, the discount spends margin and leaves the brake in place. Find the cause first.
- The permanent sale. A "limited" offer that returns every month stops being an offer. It becomes the real price, and the full price becomes a fiction customers ignore.
- Measuring only the busy week. The sale week always looks good. The weeks before and after tell you whether it created sales or just moved them.
- The same discount for everyone. Loyal customers who would pay full price get the same cut as new customers who needed a push. Target the discount at the behaviour you want.
- Inflating the "was" price. Showing a crossed-out price that was never really charged may work once, but customers notice, and trust is far more expensive to rebuild than margin.
The strongest objection
"Everyone in my market discounts. If I don't, I lose customers." Some, yes. The question is which ones, and whether they were profitable. In many categories there are buyers who choose on price alone and buyers who choose on fit, trust or convenience. Matching every competitor cut means competing for the first group with the second group's margin. Taking part in the moments your market expects, such as White Friday or end of season, can be sensible; following every cut is not.
Honest limits
In some categories, such as groceries or fast fashion, regular promotions are part of how the market works, and the question is how to run them well, not whether to run them. Clearing dead stock is a legitimate use of a discount, even at a loss. If the price really is too high for the value customers see, the fix is a new price or a better offer, not a temporary cut. And if the business needs cash this week to survive, that is a finance decision first; we can help shape the offer, but not make that call.
How OG helps with discounts and offers
We start with your numbers: margins by product, the history of past offers, and what happened before, during and after each one. Through pricing strategy we work out where the price is right, where the offer around it needs to change, and which situations really call for a discount. Where we need to know why customers buy or wait, customer research asks them. Targeted offers to new or returning customers usually run through email or WhatsApp marketing rather than a site-wide banner, and analytics sets up the before-and-after tracking so the next decision is based on profit, not on a busy week. Scope follows the need; there are no fixed packages.
For how the shape of an offer steers choice without cutting price, see pricing a new product. For restaurants, menu pricing covers the same questions on a menu.
Built on: our pricing work with clients since 2013; A. Ehrenberg, K. Hammond and G. Goodhardt, “The After-Effects of Price-Related Consumer Promotions”, Journal of Advertising Research (1994); Byron Sharp, How Brands Grow (2010); Richard Thaler, “Mental Accounting and Consumer Choice”, Marketing Science (1985); Les Binet and Peter Field, The Long and the Short of It (IPA, 2013). The margin table is arithmetic. The abaya store is a hypothetical example.
Before closing this page
- What is the gross margin on the products most often discounted?
- How many extra sales did the last discount need just to break even, and did it get them?
- What happened to sales in the four weeks after the last offer ended?
- How many of the discounted orders came from customers who buy anyway?
- What one-sentence job would the next discount have?
