How to run a discount without losing money

Everybody else is running a sale, so you knock 20% off, and you get busier, and at the end of the month there's less money than there was before. You sold more. You made less. It doesn't feel like it should be possible.

It is, and the arithmetic is short.

A discount doesn't come out of your price. It comes out of your margin, and your margin is a much smaller number than your price. If you make 30% margin and you knock 20% off, you have to sell three times as many units to end up where you started. Not 20% more. Three times.

That single multiplier is the whole article. Everything below is how to calculate yours, when a discount is still worth running, and what to do instead most of the time.

Work out the real margin first, because yours is probably fiction

Before any of this means anything, you need an honest cost per unit. Most small sellers don't have a margin, they have a difference: selling price minus what they paid for materials. That's not a margin. That's a subtraction.

A real cost per unit includes:

• Materials, including the bits you don't count. Thread, tape, labels, the poly bag, the tissue paper.
• Anything you outsource. Printing, tailoring, roasting, cutting.
• Wastage. If one in eight prints comes out wrong, your cost per good print is your cost divided by 0.875.
• Packaging that gets it there intact.
• Payment costs, if any.
• Your own time, at a number. Any number. ₹200 an hour, £12 an hour, whatever you'd accept from someone else.

That last one is where the fiction usually lives. If your cost sheet doesn't have your hours in it, every discount you run comes out of your wages, and you won't notice until you're exhausted and can't work out why the busy month felt worse than the quiet one.

Delivery is separate and you should keep it separate. It doesn't get discounted with the goods and it distorts the maths if you mix it in.

The break-even table

Here's the number that matters. Read down your current margin, across to the discount you're considering, and that's how many times your normal sales volume you need to hit just to make the same profit as before.

Margin here means gross margin as a percentage of your selling price, after real costs including your time.

Your margin	10% off	15% off	20% off	25% off	30% off
20%	2.0×	4.0×	never	never	never
25%	1.7×	2.5×	5.0×	never	never
30%	1.5×	2.0×	3.0×	6.0×	never
40%	1.3×	1.6×	2.0×	2.7×	4.0×
50%	1.25×	1.4×	1.7×	2.0×	2.5×
60%	1.2×	1.3×	1.5×	1.7×	2.0×
70%	1.2×	1.3×	1.4×	1.6×	1.75×

"Never" means the discount is larger than your margin, so every additional sale loses you money and selling more makes it worse.

The formula, if you want to run your own: multiplier = margin ÷ (margin - discount), both as percentages of selling price.

Two things fall out of that table immediately.

If you sell physical goods you buy in and resell, your margin is probably between 20% and 40%, and at that end almost every discount you can advertise is a bad deal. A 25%-off sale on 30% margin needs six times your normal volume. Nobody does six times.

If you make the thing yourself and your margin is genuinely 60% or 70%, discounts are survivable. Digital products, where the marginal cost is close to zero, are a different game entirely, which is why every course and template gets discounted constantly and physical sellers who copy that behaviour go broke.

What Meera's Diwali sale actually cost her

Meera sells hand block-printed cotton kurtas from Jaipur. She sells them at ₹1,450, mostly through Instagram, mostly paid by UPI to her own UPI ID, with about a third going cash on delivery.

Her real cost per kurta:

Line	Cost
Cotton fabric, 2.5m	₹380
Block printing (outsourced)	₹150
Tailoring	₹260
Label, tag, poly bag	₹45
Her own time, 45 min at ₹250/hr	₹190
Total cost	₹1,025
Selling price	₹1,450
Margin	₹425 (29.3%)

In a normal month she sells around 22 kurtas. That's 22 × ₹425 = ₹9,350 of gross profit.

For Diwali she ran 20% off. New price ₹1,160. Cost unchanged at ₹1,025. New margin: ₹135, or 11.6%.

From the table: 30% margin, 20% off, roughly 3×. She'd need about 69 kurtas to stand still.

She sold 41. Nearly double her normal month, the busiest four weeks she'd ever had.

41 × ₹135 = ₹5,535.

She sold 86% more and made 41% less. And that's before the two things that made it worse: she offered free delivery over ₹2,000 to push the bundle, which 14 orders qualified for at about ₹80 a shipment, and she worked most of two weekends. Take the ₹1,120 of delivery off and she cleared ₹4,415 against a normal ₹9,350, for double the work.

The part she found hardest to accept afterwards: at least half of those 41 buyers were her existing customers, who bought in October at ₹1,160 the kurta they'd have bought in November at ₹1,450. The sale didn't create those sales. It moved them, and marked them down on the way.

That's the failure nobody warns you about.

A discount to your own list is mostly a rebate on revenue you already had.

When a discount is actually the right call

There are real reasons. They're all specific, and none of them is "it's Black Friday".

Clearing stock that's costing you. Fabric you'll never use, a colourway nobody wants, the last eleven of a design you've discontinued. Here the alternative isn't full price, it's zero, so any price above your marginal cost of shipping it is a win. Take 40% off and be glad.

Buying a first order from a stranger. A one-time first-purchase discount is a customer acquisition cost, and you should compare it to what acquiring a customer costs any other way. If 10% off closes people who otherwise wouldn't risk you, and a third of them come back at full price, that's cheap. Judge it on the second order, not the first.

Moving volume to hit a real threshold. A supplier price break at 200 units, a courier rate that changes at 50 parcels a week. If the discount unlocks a genuine cost reduction, the maths changes because your cost changes. Rerun the table with the new cost.

Making room. Physical space, or a batch with a shelf life. Food, plants, anything dated.

Notice what's not on that list: because sales are slow. A discount will not fix low demand, and if you're at zero sales the reason is almost never price. What to do when nobody is buying works through the six things it's more likely to be, in order of how cheap they are to check.

Discounts that cost less than a percentage off

Most of the time you want the effect of a discount, which is a reason to buy now, without giving away the margin. There are cheaper ways to buy that.

Tactic	What it costs you	What it does well
Free delivery over a threshold	The delivery cost only, on qualifying orders	Raises average order value instead of cutting margin
Bundle of two at a small saving	Margin on one unit, spread over two	Sells the slow item alongside the fast one
A small free extra	Your cost of the extra, not your price	Feels generous, costs a fraction of a percentage cut
Early access for past buyers	Nothing	Rewards loyalty without a markdown
A genuinely limited batch	Nothing	Urgency without a price change
Payment split into two	Some admin, some risk	Removes a cash-flow objection, not a price objection

The free-delivery threshold is the one most small sellers underuse. Set it about 30% above your average order value. If your average order is ₹1,450, set free delivery at ₹1,900, and a chunk of buyers add a second item to get there. You give up ₹80 of delivery and gain ₹1,450 of goods, which is a far better trade than giving up ₹290 of margin on the order they were already placing.

The gift-with-purchase trick works for the same reason. A ₹60 pouch reads as a gift worth ₹200 and costs you a quarter of what a 20% cut would.

Don't raise the price so you can discount it

The oldest trick in retail is to put the price up to ₹1,900 for a fortnight, then advertise "40% off, now ₹1,140", which is your normal price with a bigger number crossed out.

Two reasons not to.

The first is that it's regulated. In the EU, a price-reduction announcement has to state the lowest price the item was sold at in the previous 30 days, which exists specifically to stop this. The UK's pricing guidance takes a similar line on what counts as a genuine previous price. If you sell to European buyers, the crossed-out number has to be real.

The second reason is worse and applies everywhere. Your regulars know what you charge. They've been in your DMs for a year. The single fastest way to make someone stop trusting your prices is for them to notice that the "was" price was invented, and once they've noticed, every future price you post gets mentally discounted before they even ask.

If you want a comparison price, use a real one: the price you actually charged last season, or the price of the two-item bundle broken down per item.

Time-box it, or it becomes your price

A discount with no end date isn't a discount. It's a price cut you haven't admitted to, and once buyers learn that your things go on sale, they stop buying between sales. That's the trap that eats specialist retailers alive and it works exactly the same at 22 kurtas a month.

Rules that keep it contained:

• Put an end date on it and honour it. If it ends Sunday, it ends Sunday.
• Don't run one more than three or four times a year.
• Never discount your best seller. It sells anyway. Discount the thing that doesn't.
• Don't stack. One offer at a time, or you'll wake up having sold something at 45% off you meant to sell at 20%.
• Say what the discount is for. "Clearing the last of the indigo run before the new batch" is credible. "Mega sale" trains people to wait.

And if you find yourself running a sale every month to hit your number, the problem isn't the discount, it's the price. When to raise your prices is a more useful conversation than another markdown.

The small costs that come out after the discount

Two things reduce your margin further, and both are calculated after the discount is applied, so they hurt slightly more than you'd expect.

Payment costs. Whatever your card processor charges is on the discounted total, which is fine, but it's a fixed-plus-percentage structure on most rails, so the fixed part becomes a bigger share of a smaller order. A flat fee that was 2% of a ₹1,450 order is 2.5% of a ₹1,160 one. On small baskets this gets ugly fast, and how payment fees eat a small order has the detail.

Platform commission, if any. Worth knowing exactly what yours is charged on.

For Sailo specifically, since this is the article where the number actually matters: card orders carry a 1–3% platform fee, taken on the goods after discount, excluding delivery and tax. On Meera's discounted ₹1,160 kurta that's about ₹5.80. That's not the number that hurts. The number that hurts is the rate itself: 3% on the free plan, and it doesn't get cheaper because you had a bad sale. Manual rails, including the UPI transfer she actually uses, carry no Sailo commission at all, because Sailo never touches that money. Coupon codes are also a Business plan feature, so on the free and Pro plans a discount is you changing the price and changing it back, by hand, on the day.

That's a real constraint and it argues for the manual version at her volume, not for the upgrade.

Before your next sale, do three sums

Not a strategy. Three numbers on paper, in about fifteen minutes.

1. Your real cost per unit, with your own hours in it at a number you'd accept from someone else.
2. Your margin as a percentage of your selling price. Cost divided by price, subtracted from one.
3. Your break-even multiplier from the table, or from margin ÷ (margin - discount).

If that multiplier is above about 2×, don't run the discount. Run a free-delivery threshold or a bundle instead, and put the end date on it before you post.

And when the sale is over, check where the buyers came from. If most of them were people already on your list, you didn't run a promotion, you ran a rebate, and the same effort spent on getting your first ten orders from people who've never bought from you would have been worth more at full price.