- Every dollar of discount comes straight out of contribution margin, so a 20% code on a product with a 40% margin gives away half the profit on every order it touches, including orders that would have happened anyway.
- Before any promotion, calculate the extra volume you need just to stand still: Required sales lift = discount % ÷ (contribution margin % − discount %). On thin margins, the number is usually out of reach.
- Pick the cheapest offer that changes behavior (thresholds, gifts, bundles, early access), give it only to segments that need it, and judge promotions with a holdout group and contribution margin, not revenue.
What does a discount actually cost?
A discount doesn’t reduce your product cost, shipping, packaging or payment fees. It only reduces the price, so the full amount comes out of contribution margin, the money left after variable costs. A store with a healthy gross margin can still have a thin contribution margin once fulfillment and fees are counted, and that thin layer is what the discount eats.
The four hidden costs:
- Subsidy. Many customers who use a code would have bought at full price. Every one of those orders is pure margin given away.
- Pull-forward. A sale moves purchases earlier. Customers who would have bought next month buy now at a lower price, and next month looks weak.
- Price training. Promote often enough and customers learn to wait. Full-price conversion drops between sales, and shoppers ignore your abandoned cart emails while holding out for a code.
- Customer quality. Customers acquired on a deep discount often reorder less. If so, their customer lifetime value is lower and they cost more to acquire than the first order suggests.
Add leakage on top: codes on coupon sites, codes that stack, and welcome codes reused by existing customers.
How much more do you need to sell to break even on a discount?
If you cut the price, each order earns less margin, so you need more orders to earn the same total profit. Few teams calculate how many.
Required sales lift = discount % ÷ (contribution margin % − discount %)
Both percentages are measured against the full selling price, and contribution margin here is before ad spend. The result is the percentage increase in units you need just to make the same contribution margin as without the promotion.
| Contribution margin | 10% off | 15% off | 20% off | 25% off |
|---|---|---|---|---|
| 30% | +50% | +100% | +200% | +500% |
| 40% | +33% | +60% | +100% | +167% |
| 50% | +25% | +43% | +67% | +100% |
| 60% | +20% | +33% | +50% | +71% |
Read the 40% row. A 20% sitewide code needs to double your unit sales for the week just to break even. A 25% code needs nearly triple. And that lift has to be incremental, meaning extra orders that would not have happened without the promotion, not orders pulled forward from next month or subsidized for customers who were already buying.
Worked example (illustrative numbers)
A store normally takes 1,000 orders a week at an $80 average order value with a 45% contribution margin. That is $36 of margin per order and $36,000 a week.
The team runs a 20% sitewide sale. Orders jump to 1,500 for the week.
| Line | Normal week | Sale week |
|---|---|---|
| Orders | 1,000 | 1,500 |
| Average order value | $80 | $64 |
| Revenue | $80,000 | $96,000 |
| Contribution margin per order | $36 | $20 |
| Total contribution margin | $36,000 | $30,000 |
Revenue is up 20% and orders are up 50%, but profit is down $6,000. The formula says this promotion needed 0.20 ÷ (0.45 − 0.20) = 80% more orders to break even.
Then count pull-forward. If 200 of those orders would have happened the next week at full price, that week loses about 200 × $36 = $7,200 in margin. The real cost is closer to $13,000, and none of it shows up in a revenue dashboard.
Which discount types cost the least?
Percentage-off codes are the most expensive way to change behavior because they apply to every item and every customer. Most of the alternatives cost less per order or give something back.
| Offer type | What it really costs you | Best use | Main risk |
|---|---|---|---|
| Sitewide % off | The full % of every order’s price, from margin | Clearing stock, rare tentpole events | Subsidy, pull-forward, price training |
| $ off above a threshold | Fixed amount, only on larger orders | Raising average order value | Threshold set too low to change anything |
| Free shipping threshold | Your actual shipping cost, on qualifying orders | Removing a common objection | Cost rises with heavy or remote orders |
| Gift with purchase | The gift’s product cost, not its retail price | Introducing a second product line | Gifts nobody values |
| Bundles and volume pricing | Discount only on extra units | Consumables, sets, gifting | Bundles that cannibalize single sales |
| Store credit or points | Cost only when redeemed, on a later order | Earning the next visit | Liability building up unredeemed |
| Early access or exclusives | No price cut at all | Rewarding loyal full-price buyers | Needs a product or drop worth waiting for |
Offers tied to a threshold or a second purchase protect margin on orders that would have happened anyway. Offers paid in your own product cost you product cost, while the customer values them at retail.
Who should get a discount, and who shouldn’t?
The cheapest discount is the one you don’t send to someone who was going to buy. That makes discount strategy a segmentation problem.
Our Value × Intent (RFMB) model scores customers on recency, frequency and monetary value, then adds a Buying Stage layer from the last 30 days: B1 visit, B2 product view, B3 add to cart, B4 checkout, B5 purchase. It tells you how valuable someone is and how close they are to buying. A practical rule set:
| Segment | Default offer | Why |
|---|---|---|
| Loyal full-price buyers | None. Early access, new products, a thank-you | They already buy; a code is pure subsidy |
| High-value customer at B3 or B4 | Personal reminder, no code | Intent is high; friction is the issue, not price |
| New subscriber, no purchase | A modest first-order offer, or a non-price incentive | Earn the first order without setting a low anchor |
| One-time buyer past their reorder window | Product-led reminder first, small incentive later | Save the offer for people who don’t respond |
| Lapsed customer | Escalating win-back campaign offer | Price can matter more once a customer has drifted |
| Repeat code-only buyers | Limit exposure to sitewide codes | Their margin is already lowest |
For new subscribers, consider asking instead of discounting. A short quiz or preference question collects zero-party data you can use to personalize the next message, and it gives a reason to sign up that is not a price cut.
How do you know whether a promotion made money?
Revenue during the promotion tells you almost nothing. You need to know what would have happened without it.
Holdout groups. For an email or SMS promotion, randomly withhold the offer from a slice of the audience, for example 10% to 20%. Compare contribution margin per recipient between the two groups over the promotion and the weeks after. This is incrementality testing in its simplest form.
The post-promotion dip. Compare the two to four weeks after a sale with the same period in a normal month. A deep dip after each promotion is pull-forward you’ve been paying for.
Cohorts by first-order discount. Use cohort analysis to compare customers who first bought with a code against full-price buyers, on repeat rate and margin at 90 and 180 days. If code-acquired customers reorder far less, test a smaller welcome offer. Our article on the first-to-second purchase rate shows how to build that view.
What guardrails should you set before the next promotion?
Many margin leaks come from settings. On Shopify, discounts only combine if you allow it in each discount’s Combinations section, so review those settings for every active code. Restrict eligibility to specific customer segments where it makes sense, use “Limit to one per customer” for single-use offers, and always set an end date, because a Shopify discount without one never expires. Unique codes per recipient, such as Klaviyo’s unique Shopify coupons with an expiration window, are much harder to leak than one shared code.
If your promotions show “was” or compare-at prices, reference-price claims are regulated. Keep records that support them and confirm your approach with counsel.
Before any promotion goes live:
- [ ] Contribution margin per product line is known, and the required sales lift for this discount depth is written down.
- [ ] The offer is the cheapest type that could change behavior (threshold, gift, bundle or credit before % off).
- [ ] Loyal full-price buyers and recent purchasers are excluded or get a non-price version.
- [ ] Combination settings, usage limits, eligibility and end dates are set on every code.
- [ ] A random holdout group is in place for email and SMS sends.
- [ ] Success means incremental contribution margin, including the weeks after the sale.
A margin-first strategy doesn’t mean never discounting. It means every promotion has a cost calculated in advance, a segment that needs it and a result measured against a control. To see what discounted customers are worth over time, compare full-price and discount-acquired cohorts in the LTV & LTV:CAC calculator. Our contribution margin guide shows where promotions fit in a profit-first plan.
FAQ
Are discounts always bad for an eCommerce brand?
No. They make sense for clearing stock you would otherwise write off, for winning back lapsed customers and for well-timed events with a measured result. The problem is discounting by default, to everyone, without knowing the break-even lift or testing against a holdout.
How big should a welcome discount be?
There is no universal right number. Work backward from your contribution margin and from how well code-acquired customers retain compared with full-price buyers. If they reorder far less, test a smaller offer or a non-price incentive.
Is free shipping cheaper than a percentage discount?
Often, but not always. Free shipping costs your actual shipping expense per order, while a percentage discount grows with order size. Tie free shipping to a threshold above your current average order value, and check carrier costs for heavy or remote orders.