Somebody in your category announces 40% off. You match it, because that is what everyone does, and now the biggest sales week of your year is running on a number you picked in an afternoon. Nobody checked whether your margin could carry it.
That is how a store ends Black Friday with a record revenue screenshot and less money in the bank than last November. Revenue went up. Gross profit went down. The discount, the ads you bought to feed it, the returns that came back in January, and the payment fees all came out of the same pocket.
Black Friday 2026 is November 27. Cyber Monday is November 30. Between now and then you have time to pick a discount you can actually afford, and the math that decides it takes one afternoon.
The number that matters is your gross margin per product, not your average
Most founders can tell me their store’s average margin. Very few can tell me the margin on their five best sellers, one at a time.
The average is dragged up by products nobody buys, and your bestseller is often your worst-margin item, because volume products get priced against competitors and everything else gets priced against cost. Discount by an average and you cut hardest on the products that can least afford it.
So build a real landed cost per SKU before you build the offer. Unit cost from the supplier. Inbound freight and duty. Pick, pack, and the box. Your payment processor’s cut, which is on your statement, not in your head.
Subtract all of that from the selling price. What is left is what a discount comes out of. Do this for your top ten sellers and you already know more than the store next door running 40% off on vibes.
What a discount actually costs, in arithmetic
Take one product. It sells for $100. Landed cost is $50. Gross profit is $50, a 50% margin.
Now take 20% off. The price becomes $80. The cost is still $50. Gross profit falls to $30.
Read that again. You cut the price by 20% and cut the profit by 40%. To make the same gross profit dollars as before, you now need to sell roughly 1.7 units for every one you used to sell.
Go to 40% off and it gets sharper. Price $60, cost $50, gross profit $10. That is an 80% cut in profit per unit, and you need about five times the volume to stand still.
Then subtract what nobody puts in the model: the traffic you bought to make the sale happen, the percentage that comes back as a return, the support time on a busy week. On a 40%-off order carrying $10 of gross profit, one paid click can be enough to make that order cost you money.
None of this means do not discount. It means know which number turns profitable volume into expensive volume, and refuse to cross it.
Stacking is where a good plan turns into a bad one
You plan 30% sitewide. Then the welcome popup keeps handing out its 10% code. Then free shipping kicks in over $75. A buyer combines all three, and the order that was supposed to carry a thin margin now carries none.
This is a settings problem. On Shopify, discounts fall into three classes: product, order, and shipping. A customer can apply up to five product or order discount codes plus one shipping discount code on the same order, and combinations only happen when the setting on each discount allows it. The rules are in Shopify’s own documentation on combining discounts, worth reading before November, not during it.
Decide now, in writing, which of your discounts may combine. Then place a real test order and try to stack every active code you have. Whatever the checkout lets you get away with in a test is what a deal-site comment thread will do at scale on November 27.
The bill that arrives in January
A discounted order still costs full price to fulfil, and it still comes back at the same rate. Sometimes higher, because a price-driven buyer is less attached to the product.
When it comes back you pay return freight, inspection, and restocking, and you refund the discounted revenue. The gross profit on that order was $10. The cost of handling its return is often more. That is a straight loss on a sale you celebrated eight weeks earlier.
So look at your return rate per product before you decide which products go on sale. Your highest-return SKU is the worst thing to put at 40% off. If you have never quantified it, start with why returns are eating your margin.
Offers that hold margin better than a sitewide percentage
A flat sitewide discount gives the deepest cut to the buyer who was going to purchase anyway. Alternatives that do more work per dollar given away:
- Spend thresholds. “$25 off orders over $150” raises order value instead of cutting price, and you control exactly how much margin leaves per order.
- Bundles built around a high-margin item. The perceived discount is on the bundle, and the mix protects the blended margin.
- Free shipping above a number set higher than your current average order value. Cheap to give, strong at moving people up.
- Gift with purchase, using slow inventory. You pay landed cost instead of a percentage of revenue, and you clear stock you were carrying anyway.
- Early access with no discount at all. Give your email list first pick on limited stock. It costs nothing and works on the people who already trust you.
Pick one or two. A promotion nobody can explain in a sentence just breaks in more places.
Write the floor down before October
Decide the minimum gross profit per order you will accept, in dollars, and write it on the same page as the offer. Anything that lands below the floor after discount, ads, and expected returns does not run. Not for a competitor, not for a Slack message on the Wednesday of that week.
Do this in August, while there is still room to change your mind calmly. October goes to creative and campaign production, and November should be a freeze. That sequencing is the point of the Q4 checklist I run starting in August, and the discount decision belongs in the same window as the technical work.
What I would do this week
Pull landed cost on your top ten sellers. Run the arithmetic above on each one. Pull the return rate for the same ten. You will finish the afternoon knowing which products can carry a real discount and which ones cannot.
If you would rather have somebody find the numbers with you, that is a Growth Audit. In 72 hours, for $497, every fixable problem in your store gets found, named, and ranked by what it is costing you, so you go into Q4 knowing where the money actually goes. The fee credits toward any Sprint within 30 days, and Sprints start at $3,500 and run 14 days. Booked in July, the fixes are done and settled long before November 27.
Frequently asked questions
How much should I discount for Black Friday?
There is no universal percentage, because the answer comes out of your landed cost. Work out gross profit per product first, then test the discount against it. On a product that sells for $100 with a $50 landed cost, 20% off cuts your profit per unit by 40%, and 40% off cuts it by 80%. Pick the deepest discount that still leaves an order profitable after ads and returns, and hold that as your floor.
Why did my revenue go up on Black Friday but my profit go down?
Because a discount comes out of gross profit, not out of revenue, so profit falls much faster than price. Add the cost of the paid traffic that produced the order, the payment fees, and the returns that arrive in January, and a heavily discounted order can cost more to fulfil than it earns. Revenue records are easy to set with a big enough discount. Profit records are not.
When is Black Friday 2026?
Black Friday 2026 is November 27 and Cyber Monday is November 30. Set your discount floor and your promotion rules in August, because October is consumed by campaign production and November is when you should be changing as little as possible.
Should I let customers stack discount codes during a sale?
Only where you have deliberately allowed it and checked the result. On Shopify a customer can apply up to five product or order discount codes plus one shipping discount code on a single order, and discounts combine only when each one’s settings permit it. Decide which of your codes may combine, then place a real test order and try to stack every active code you have before the sale goes live.
Ready to fix what's broken?
Stop guessing and get a real diagnosis. Book a Growth Audit or jump straight into a Sprint.
Find this useful? Make Ecomm Decoded a preferred Google source → so it shows up first when Google answers your ecommerce questions.