A founder showed us a Black Friday recap last year. Revenue up 41% on the weekend. Orders up 38%. Team thrilled. Then we pulled contribution per order, the money left after product, shipping, fees, and the 25% code they had run. It was down 52%. They had done more work, moved more boxes, and made less money than a normal weekend. Nobody had looked at the number that mattered.
This is not an argument against ever discounting. It is an argument for doing the arithmetic first, because a discount is not a marketing cost. It comes straight out of contribution margin, the pool that has to pay for your ads, your team, and your profit, and it is the single most common leak we find in the accounts we audit.
01 How common, and how deep
Discounting is nearly universal and, at the top end, getting shallower. Klaviyo’s data on Black Friday and Cyber Monday 2025 put the average discount at 26.2%, down from 29.6% the year before, while same-site sales still rose 11%. Salesforce measured the 2024 US holiday season at 23% off on average. An app-data analysis of Shopify stores found the single most common BFCM tier was 20% off, used by about a third of stores, with another quarter going to 30%. Twenty to thirty percent is the default. Almost nobody sets it from their margin.
02 The break-even math
The core formula is short, and it is brutal at low margins. To make the same gross profit after a discount, the extra volume you need is:
Required volume increase = discount ÷ (gross margin − discount)
Run it across two margin profiles:
| Discount | At 70% gross margin, you need | At 35% gross margin, you need |
|---|---|---|
| 10% off | +17% more orders | +40% more orders |
| 15% off | +27% | +75% |
| 20% off | +40% | +133% |
| Buy one get one (about 50% off) | +250% | not achievable |
That is gross margin. Contribution margin, after shipping, payment fees, and returns, is thinner, so the real hurdle is higher. Take an order with $40 of contribution. Offer 25% off a $100 price and $25 comes straight out of that $40, leaving $15. You now need 2.7 times the volume to stand still. A 30% lift in orders, which most promos are thrilled with, leaves you well underwater.
03 Who is buying with the code
The deeper problem is who redeems it. Klaviyo and ProfitPeak analyzed 176 ecommerce brands across a year of promotions. In discounted orders, 68% of buyers were returning customers. In full-price orders, only 11% were. The discount was overwhelmingly used by people who already buy from you, at a price they had already shown they would pay. That is not acquisition. That is a margin refund to your best customers.
The same analysis found revenue in the two to four weeks after a promotion ran about 27% below baseline, because the sale pulled demand forward rather than creating it. And during sitewide events, core bestsellers still produced 90% of revenue while the long-tail stock the sale was supposedly clearing brought in 0.2%. Brands running eleven or more promotional events a year grew GMV 6% with margins down 11%. Brands that discounted rarely grew 12% with margins up 8%.
04 What discounts do to the price in the customer’s head
This part has decades of peer-reviewed evidence behind it. Kalwani and Yim showed in the Journal of Marketing Research back in 1992 that both the frequency and depth of promotions raise the reference price consumers expect, which means every sale lowers the price they consider normal. A 2018 study in the Journal of Marketing by del Rio Olivares and colleagues found the effect of an initial discount on retention is nonlinear, so a deeper first-purchase offer does not simply buy proportionally more loyalty. Train customers to wait for the code and they will, and your full price becomes a suggestion.
05 The paid-ads angle nobody prices in
Discount codes leak in two more places once ads are involved. First, attribution. Public sitewide coupons cannibalize somewhere between 20% and 60% of sales that would have happened at full price, and coupon sites and browser extensions sit at the last click claiming credit for demand your ads created. Honey, the best known extension, is being sued by creators and merchants over exactly that cookie-swapping practice; the claims are still allegations, but a judge let the case proceed in June 2026 and Rakuten dropped Honey from its network in January. Whatever the courts decide, your “code X drove $Y” report is overstated.
Second, your break-even. A 15% average discount can add a full point or more to the ROAS you need just to break even, which is the same margin math we walked through in why your ROAS is lying to you. Run the promo and your affordable cost per customer drops at the exact moment you are spending hardest to win customers. That is why promo months so often show record revenue and a thinner bank balance, a pattern we covered in how much to spend on ads.
06 What to run instead
- Free-shipping thresholds set 20% to 30% above current AOV. Shoppers add items to qualify, AOV rises 12% to 24% in the published data, and your margin per order goes up instead of down.
- Bundles with a single-digit discount. Bundling lifts AOV 20% to 30%, but only protects contribution if the discount covers real fulfillment savings, usually 5% to 10%. The default 20% bundle discount is margin-dilutive on most accounts we review.
- A gift with purchase over a percentage off. A $5 gift on a $60 order costs less than 8% and feels bigger than 8% off.
- Member pricing inside a loyalty structure, not a public code. Loyalty-based pricing has shown roughly 19% better retention than generic coupons, and it keeps the discount away from the public reference price.
- Retention instead of promotion. Moving average orders per customer from 1.2 to 1.6, with no discount at all, lifted contribution per order 18% in one case we found. Repeat purchase is the margin lever that does not train anyone.
07 Where to start
Before the next promo, do one calculation: contribution per order at full price, then at the planned discount, then the volume lift needed to match. If the lift you need is bigger than any promo has ever delivered for you, the discount is a loss you have chosen in advance. Our unit-economics calculator runs it in a couple of minutes. And if promo months keep setting revenue records while cash gets tighter, show us the calendar; that is the leak, and it is fixable. What did your best sale day of last year leave behind per order, after the code?
Sources: Klaviyo BFCM 2025 data (26.2% average discount, down from 29.6%); Salesforce Shopping Index 2024 holiday (23% US); app-data analysis of Shopify BFCM 2025 discount tiers (Fabrikator); Klaviyo and ProfitPeak analysis of 176 brands, Q1 2025 to Q1 2026 (68% vs 11% returning share, 27% post-promo dip, GMV and margin comparison); Digital Applied 2026 discount playbook (break-even volume table, cannibalization ranges); Kalwani and Yim, Journal of Marketing Research, 1992; del Rio Olivares et al., Journal of Marketing, 2018; free-shipping and bundling AOV data (SellersCommerce, Swell, Eightx); Access Development loyalty data; 2026 coverage of the Honey litigation (allegations, not findings).
Want the full playbook?
The 12 Growth Leaks burning your ad budget — the same internal doc we hand to new clients on day one. One short form, no spam.
Get the playbook →FAQ
Do discounts hurt profit margins?
Usually, yes, unless the volume lift is large. The extra orders needed to make the same profit equal the discount divided by your margin minus the discount. A 20% discount at a 35% gross margin needs 133% more orders to break even; at a 70% margin it still needs 40% more. Most promotions deliver far less than that.
How much can I discount before I lose money?
Compute contribution per order at full price, subtract the discount in dollars, and compare the volume you would need against what promos have delivered before. On a $100 order with $40 of contribution, 25% off leaves $15 and requires 2.7 times the volume. For most DTC brands the safe ceiling is single digits to low teens.
Do customers acquired with a discount have lower lifetime value?
The best evidence says discounts mostly reach existing customers rather than new ones: in a 176-brand Klaviyo and ProfitPeak analysis, 68% of discounted orders came from returning buyers versus 11% of full-price orders. Peer-reviewed research also shows promotions lower the price customers consider normal, which erodes willingness to pay over time.
Why do sales drop after a promotion ends?
Because the promotion pulled demand forward instead of creating it. Klaviyo and ProfitPeak measured revenue in the two to four weeks after a promo running about 27% below baseline. Customers who would have bought anyway bought early at a discount, leaving a hole afterward.
What can I offer instead of a percentage discount?
Free-shipping thresholds set 20% to 30% above your current average order value, bundles with a single-digit discount that covers real fulfillment savings, a gift with purchase, member-only pricing inside a loyalty program, and retention programs that raise repeat orders. Each protects contribution margin instead of refunding it.