Sales are flat this week, so you take twenty percent off. It works. Orders come in, the number on the dashboard goes back up, and the problem goes away for about nine days. Then it comes back, so you do it again, a little deeper this time, because the last one has worn off.
Most store owners know this pattern from the inside. What almost nobody is told is why it happens, and it is not because you lack discipline. It happens because the discount is the only lever you ever built. Every other way of making a customer want to buy right now takes planning, and the price field takes four seconds.
A Discount Is Not an Offer. It Is One Ingredient in One.
Here is the distinction the whole thing turns on, and it is worth slowing down for.
An offer is the complete answer to "why should I buy this, from you, today." It has five moving parts. What the product is actually for, and who it is for. How the price is structured, which is a different question from what the price is. What else the customer gets in the box. Who carries the risk if it does not work out. And why any of this should happen now rather than in March.
A discount touches exactly one of those five. It adjusts the price. It says nothing about what the product is for, adds nothing to what they receive, moves no risk off the buyer, and creates no real reason to act beyond an expiry date you invented.
That is the whole diagnosis. Stores that seem to never discount are not more disciplined than you. They built the other four.
The Discount Costs More Than the Discount
Run the arithmetic once and it changes how the price field feels.
Take a hundred dollar order. Your cost of goods is forty dollars, which leaves sixty dollars of gross profit before you have paid for anything else. Now take twenty-five percent off. The customer pays seventy-five, your cost of goods has not moved, and your gross profit is thirty-five dollars.
The discount was twenty-five percent. The profit cut was over forty percent. That gap is the part nobody sees, because the number you typed was the percentage off the price, and the number that matters is the percentage off what you keep.
There is a second bill behind that one. If you are paying for traffic, every dollar of advertising now has to work considerably harder to get you back to where you were, because it is buying orders that carry less profit. And that requirement does not climb in a straight line. Small discounts cost you a little more efficiency. Deep ones can demand a level of performance from your advertising that you were never going to hit, which is how a sale can post record revenue and lose money at the same time.
Then there is the cost that does not show up anywhere at all. A sale does not usually create demand. It reschedules it. The customer who would have paid full price a week from Tuesday just bought at twenty-five percent off today. You did not win an order. You moved one forward and paid for the privilege.
Do that often enough and next month arrives structurally short, which is the reason the discount that used to be occasional becomes the thing holding the whole month up.
The Same Discount Stops Working Because It Was Never Speaking to Strangers
This is the part that explains why your winning promotion died.
"Twenty percent off" carries information only if the person reading it already knows what full price is and already wants the thing. For a customer who bought from you twice last year, that is a genuine reason to act today. For a stranger who has never heard of you, it is close to meaningless. They have no idea what your normal price is, so they cannot tell whether twenty percent is generous or routine.
People walk around with a rough number in their head for what a thing should cost. Eight dollars for a good coffee, say. Twelve feels expensive and needs justifying. What surprises most owners is the other direction. A two dollar coffee does not read as a bargain. It reads as suspect. Something must be wrong with it.
Your discount is being measured against a number the stranger does not have. So it does not say "good deal." It says the only thing it can say, which is that you needed the sale.
That is why the promotion works beautifully on your email list and dies on cold traffic. It was built for people who already know you, and you pointed it at people who do not.
The Four Levers You Are Not Pulling
None of these require a bigger budget. They require deciding something in advance, which is most of what the work actually is when we build an offer for an online store: not finding a cleverer percentage, but choosing what goes in the box and who carries the risk before the traffic arrives.
Bundles that sell an outcome. Most bundles are two products and a percentage, which asks the customer to do the math and gives them no reason to want the combination. A bundle works when it sells a job getting done. The starter set, which is everything you need to begin. The routine, which is the three things that go together and the order to use them in. The restock, which is the month's supply so they stop thinking about it. Same products, but now you are selling a solved problem rather than a slightly cheaper pile.
A gift with purchase, which is the counterintuitive winner. Say a towel is fifty dollars and costs you fifteen. Two towels is a hundred, and you throw in a bag that costs you four dollars. You have added four dollars of cost and roughly doubled what you keep on the order, because the second towel brought its own margin and shipping two together costs barely more than shipping one. Compare that with taking twenty-five percent off two towels, which raises no order value and hands away real money.
The part that makes it work is that the bag has to be a real product with a real price on your site. A free gift of unknown value is worth nothing in the customer's head, because they have nothing to measure it against. A free gift they can see listed at twenty-five dollars is a hundred and twenty-five dollars of value for a hundred, and your headline price never moved.
Risk reversal. A guarantee is not a trust badge in the footer. It is the specific answer to the specific thing that stopped someone buying. The only way to know what that is, is to ask the people who did buy what nearly stopped them. If the answer is that they assumed it would not last, then a replacement promise is worth more than any percentage you could have offered, and it costs you nothing on the orders that go fine.
Access instead of money off. Give the people who buy from you regularly first look at the new thing, twenty-four hours before everyone else. It costs nothing. It works because of what it says about the relationship.
The Threshold Mistake Almost Everyone Makes
This one is worth checking this afternoon, because it is common and it is quietly expensive.
You want to lift order values, so you set free shipping at a bit above your average order. Your average is a hundred dollars, so free shipping starts at a hundred and ten. Sensible.
Except averages hide their own shape. Plenty of stores do not have one cluster of orders around the average. They have two. A large group buying a single item at around seventy dollars, and a smaller group buying several at around a hundred and eighty. Those two groups average out to a hundred, but almost nobody actually orders a hundred dollars' worth. The average is sitting in the empty space between your real customers, describing a shopper who does not exist.
So your hundred and ten dollar threshold does two unhelpful things at once. Everyone in the upper group was already spending more than that, and now gets free shipping they would have happily paid for. Everyone in the lower group is being asked to nearly double their order to qualify, so they do not. You gave away margin to one group and motivated neither.
Pull your last few hundred orders and look at how they are actually spread out, not what they average. If you find two clusters, you need two nudges, each sitting just above where a real group of people already is.
When a Discount Is Genuinely the Right Tool
None of this means never run a sale. It means knowing which job you are hiring it for. There are three where it is the right tool and little else is.
Turning stock into cash. If something is not moving and it is occupying money and shelf space, a discount converts a dead asset into working capital. That is a real job with a real answer.
A moment customers can learn. One predictable event a year that people come to expect is a different animal from a flash sale every third weekend. The first builds anticipation. The second teaches everyone that your prices are a suggestion.
Giving a committed buyer a reason to move today. Someone who has been circling for a month and needs a nudge over the line is exactly the customer a discount was designed for.
What those three share is a reason. An anniversary, a clearance, a genuine occasion. Customers are much better at telling the difference than we give them credit for.
The Offer Starts With Them, Not With You
The most common way a well-built offer fails is that it was designed backwards. The owner works out what protects the margin, structures the bundle around that, and then cannot understand why take-up is poor.
Nobody buys an offer because it is good for your contribution margin. They take it because it removes something that was in their way. The fear that it would not fit. The suspicion that it would not last. The reluctance to commit before trying it. The offer has to start at that objection and work backwards to the mechanics, not the other way around.
Which means the research is not complicated, it is just rarely done. Ask twenty recent customers what nearly stopped them buying. Ask the ones who returned something what they expected instead. That is your offer, sitting in plain language in your inbox, and it costs nothing to collect.
You do not need a bigger discount. You almost certainly need a better offer, and the difference between those two sentences is most of the margin you have been giving away.
We do this all day, so if you would rather hand it over than work through it yourself, we can. And if you would rather build it yourself, everything above is the actual method. Either way, stop reaching for the price field first.
