The first price you set will be a guess. That is fine, and unavoidable. What is avoidable is arriving at it the usual way, by counting the hours you spent and multiplying by what you would like your time to be worth. That number feels rigorous because it involves arithmetic. It is unrelated to what the thing is worth to anyone else, and it usually fails to cover what selling it actually costs you.
Two questions are hiding inside "what should I charge", and they are much easier apart than together. The first has an answer you can calculate. The second does not.
- Hours worked multiplied by a wished-for rate is the one number certain to be wrong.
- The floor is not what it cost to build. It is what it costs you to sell one more.
- Payment fees, tax set-aside, refunds and support are all real and all forgotten.
- Buyers compare you to their alternative, never to your effort.
- Your first number is a hypothesis. Publish it and watch what gets challenged.
The number everyone reaches for first
It goes like this. Sixty hours went into it. Your time is worth fifty an hour. So the thing is worth three thousand, and if you expect to sell a hundred copies, that is thirty each.
There are two problems and the second one is fatal. The first is that the hundred is invented, and every conclusion downstream of an invented number inherits its confidence without inheriting any of its accuracy. The second is that the buyer does not know how long it took you and would not pay differently if they did. Nobody has ever chosen a product because it was laborious to make.
What the hours actually tell you is something else, and it is worth knowing: whether the project was worth doing at all. That is a question about your last sixty hours. It is not a question about the price tag, and the two only look similar because both are measured in money.
The floor is the cost of one more sale
Before you decide what to charge, work out the number below which each sale takes money out of your pocket. This is not the cost of building the thing, which has already happened. It is the cost of the next transaction, and it is almost always higher than people expect because most of it arrives after the sale.
| What comes out | Why it gets missed |
|---|---|
| Payment processing | It is a percentage plus a flat fee. The flat fee is what hurts on anything cheap. |
| Platform or marketplace cut | It shows up in the payout, not in the sale notification you actually read. |
| Tax set aside | It lands in your account, so it feels like income. It is not yours. |
| Refunds and chargebacks | Everyone budgets zero. Nobody gets zero. |
| Support | Ten minutes per buyer is not free, and unlike the build it repeats forever. |
| Admin | Receipts, invoices and the occasional tax question from a buyer abroad. |
Put invented numbers into that table for a moment, purely so the shape is visible. On a sale of fifty, processing might take two, the platform another five, tax set aside twelve, and a refund rate of one in twenty removes another two and a half on average. That is twenty one and a half gone before you have answered a single email. If answering email takes you ten minutes per buyer, the remaining twenty eight and a half is your wage for those ten minutes plus everything else you did.
Those figures are made up and yours will differ, sometimes a lot. Run the same six rows with your own and the useful part is not the total. It is discovering which row is largest, because that is the one worth attacking.
Why cost-plus falls apart on things that copy for free
Cost-plus pricing works when each unit costs something to produce. Add a margin to the unit cost and you have a defensible number. For anything digital the unit cost rounds to nothing, so cost-plus returns nothing plus a margin, which is a margin on nothing.
The cost that does scale with sales is your attention. Every buyer is a small ongoing liability: a question, an edge case, a refund request, a note that something broke on their machine. That is the real per-unit cost, and it means the honest version of the question is not "what does this cost to make" but "at this price, how many buyers can I carry before support eats the money".
Price low enough and you can win on volume while losing on time, which is the specific failure that makes people quit and conclude the market was not there.
The three things a buyer is comparing you to
Nobody evaluates a price in isolation. They put it next to something, and there are only three somethings.
What it would cost to have a person do it. This is the strongest anchor you have and the one most people never mention. If your thing replaces four hours of someone's afternoon, four hours of that someone is the comparison, not the other software.
What the problem costs them now. Late invoices, a missed deadline, a report rebuilt by hand every month. This anchor is powerful and hard to use, because you have to know their situation well enough to name it without guessing.
What similar products charge. The weakest anchor, and the one everybody reaches for, because it is the only one you can research without talking to anyone. It tells you what the market has trained buyers to expect. It does not tell you what you are worth to them.
A price is not a description of your effort. It is a claim about what changes for the buyer if they say yes.
Cheap is a claim, and you have to defend it
A very low price does not remove a question from the buyer's mind. It adds one. If this does what it says, why is it nine.
There are good answers. It is narrow on purpose. It is one person with no office to pay for. It does one job and stops. Any of those, stated plainly, turns a suspicious price into a credible one. What does not work is leaving the number to speak for itself and hoping it reads as generosity.
There is also a practical cost that catches people out. Cheap products attract the buyers who need the most help, because a low price lowers the bar for buying without lowering expectations at all. The support load per pound taken is worst at the bottom of the market. That is not a reason to price high, but it is a reason not to price low by accident.
What to do when you have no market data
You will not have data before your first sale, so the goal is to make the guess less arbitrary rather than to eliminate the guessing.
Do not ask people whether they would pay a number. It is the least reliable question in business, because saying yes to a hypothetical costs nothing and being encouraging is free. People will tell you a price is fair and then not buy it, and both things are sincere.
Ask about the past instead, because the past already happened and does not need to be imagined. What are you using for this now. What did that cost. What did you try before it. How long did it take you last time. Someone who has already paid for a worse solution has told you more than any hypothetical answer could.
The people worth chasing hardest are the ones who nearly bought and did not. They are unpleasant to contact and they know exactly what stopped them, which is more than your buyers can tell you.
Moving the number afterwards
Prices move up more easily than people fear and down more permanently than people expect.
Raising is mostly a communication job. Tell existing buyers before it happens, keep them on the old price if you can afford to, and give the new number a reason. Almost nobody objects, and the ones who do were rarely going to be happy anyway.
Cutting is the move you cannot take back. A public price cut tells everyone who paid the old price what you now think their money was worth, and it tells everyone else to wait for the next one. If you need to test a lower number, do it with a discount that has an end date rather than by changing the price, because one of those is an experiment and the other is a decision.
Where to start this week
Fill in the six rows above with your own numbers and find your floor. Then write down the three comparisons a buyer might make, and pick which one your page is actually arguing against. Most pages argue against the third by accident, because competitor pricing is the easiest thing to look up.
Set the number, publish it, and leave it alone long enough to learn something. If the thing you are pricing came out of a problem in your own work, our piece on turning a repetitive work problem into a product covers the stage before this one, and the capacity maths behind doing it alongside a job covers the stage after.
Frequently asked questions
Should I charge anything at all for my first product?
Charge something. A price of zero removes the only reliable signal you have, which is whether somebody will part with money. Free users are generous with encouragement and tell you almost nothing about demand. If you want reach rather than evidence, give away a smaller thing and charge for the larger one.
What if nobody buys at my price?
Silence is ambiguous. It can mean the price is wrong, but it far more often means nobody saw it, or the page did not explain what the thing does. Check that the page is being read at all before you touch the number, because a price cut is hard to undo and will teach you nothing if the problem was traffic.
Is a launch discount a good idea?
A discount with an end date is a reversible experiment. A permanently low price is a decision. Launch pricing is fine as long as the higher number is the one on the page afterwards, and as long as you actually put it back.
How do I price against a competitor who is free?
You do not win that comparison on price, so do not enter it. Free products carry costs elsewhere, usually in setup time, missing support or the buyer becoming the product. Name the cost your version removes and let the price sit next to that.
Should the price be public?
For anything sold to individuals, yes. A hidden price filters out people who would have bought and invites everyone else to assume the worst. Contact-for-pricing makes sense when every job is genuinely different, and almost nowhere else.
What about pay what you want?
It reads as generous and behaves like a discount with extra steps. Most people pay the minimum shown, so the suggested figure becomes the real price. If you try it, set a floor and treat the suggestion as the number you actually want.
How this was put together: the framework here comes from pricing our own tools and from watching which decisions people regret, not from a survey. Every figure in the worked example is invented and labelled as such, because real percentages vary enormously by product, country and payment provider, and a made up number presented as research would be worse than no number at all. Nothing here is tax or financial advice.