
How to price a new product when you have no idea
Price a new product from what it changes for the buyer: four questions, a worked example, how to say the number and what to do with the answer.
Many founders price a new product by guessing. A buyer asks what it costs, the founder says a number that feels safe, and that number goes into every deal that follows. It's one of the most expensive guesses you'll make early on, because every customer you sign pays it.
How to price a new product comes down to one idea: a price is a claim about value. It's a slice of what changes for the buyer once they have your product: their hours, their mistakes, the customers they lose, the risk they carry. So work out the value first, and only then the price. Below: how to find that value, a made-up example with the sums, and what to do after you say the number.
Two starting points to avoid: your costs and a competitor's price
Your costs. Total your costs, put a margin on top, and call that your price. The buyer doesn't care how hard it was to build, though. They pay for what it does for them. With software, one more customer costs you almost nothing, so cost-based pricing pushes you toward a tiny fee for something that may save the buyer whole days.
A competitor's price. You don't know how they picked it, or whether it's working for them. Copy it and you've handed your most important number to someone who may have guessed too. Treat their price like a boundary, not an anchor. That number shows the range buyers in that market are used to, but it can't tell you where you belong in it. Setting out to undercut them is worse. Competing on price alone is one of the startup mistakes to avoid, because the customers it wins leave for the next cheaper offer.
How to price a new product in three moves
- Name what changes for them. Make it concrete, something you could point at. "We help teams work better" isn't a change. "The office manager no longer spends Friday afternoon matching invoices to payments" is one. If you can't say what changes, the product needs work before the price does.
- Put money on that change. A rough figure is fine: hours saved times what an hour of that person costs the business, mistakes avoided times what each one costs, customers kept times what each one pays. Expect the same change to be worth far more to one buyer than to another, so price for the buyer you're talking to.
- Take a slice. Leave most of the value with the buyer, or there's nothing in it for them. When the price sits well below the value, the sum sells it for you.
The buyer won't do this sum alone, so work it out and show it to them in the conversation. The result is what goes in the revenue streams box of a Lean Canvas, filled in box by box: a real price, how you charge it, and why it's worth it in the customer's money.
Four questions that find the value
The figures for move 2 come from the buyer. Get them by asking about the problem, without mentioning your product.
- How do you deal with this today? Whatever they describe is what you're really up against. Often it's a spreadsheet, someone doing it by hand, or nothing at all.
- What does it cost you each month? Let them work it out aloud. A number they say themselves is one they believe. Write it down in their words; you'll need it later.
- What happens if nothing changes? If the honest answer is "not much", the problem is a nice-to-have, and any price will feel high.
- Who else needs to sign off on something like this? It saves you pricing carefully for someone who can't say yes.
They follow the rule in the customer interview questions for startups: ask about their world, never about your idea. That page also warns against asking "Would you pay €50 a month for this?" in discovery, because a hypothetical price gets a polite yes. Your price comes later, in a real offer.
Pick the shape of the charge before the number
How you charge changes which numbers make sense. Four common shapes:
| Shape | How it works | What to know |
|---|---|---|
| Per project | One result, one fee | Quick to test, but no repeat income |
| Per seat | A fee for each person using it | Grows with their team, but they may keep colleagues off it |
| Per month | A flat fee for ongoing use | Predictable, and what software buyers expect |
| Per outcome | You're paid when the result happens | Buyers like it, but it's the hardest to run. Needs a result you can both see |
Charge the way the value arrives: monthly value gets a monthly fee, and one-off value a one-off fee.
A made-up example: video quotes for small moving companies
Everything in this example is invented: the product, the firms and every number.
Imagine you've built a tool for small moving companies. Their customer films a walk-through of the home on a phone, the tool lists the furniture and boxes and suggests the van and crew, and the moving company quotes from the video instead of visiting first.
What changes. You ask the owner of a firm with two vans the four questions. Today the owner visits every home before quoting: about 20 visits a month, around two hours each with the driving. Big or awkward jobs would still get a visit, about 5 a month, so 15 visits go: 30 hours a month back.
What it's worth. The owner reckons an hour of their time costs the business about €35. 30 hours × €35 = €1,050 a month. Same-day quotes might win more jobs too, but nobody can put a number on that yet, so it stays out of the sum.
The slice. You set €220 a month, about a fifth of the value. The firm keeps around €830 a month of what changes.
The shape. Monthly fits, because the visits happen every week. Per seat makes no sense when the owner is the only user.
The number you started with. Your gut said €49 a month, copied from an app for tradespeople. The €220 came from this firm's 30 hours.
Selling to consumers
The same thinking holds for consumers: the price is about what it's worth to them. The test is what changes. A consumer won't do the sum with you in a meeting, and asking a stranger what they'd pay gets a guess. So test with real money:
- A pre-order. They pay now and get it later. Money moved, so it counts.
- A waiting list with the price shown. Weaker, since joining costs nothing, but it filters out people who would never pay that price.
- Two prices on a landing page, each shown to a different batch of visitors, with a pre-order or a deposit behind the button. Then compare how many paid at each. With only a few hundred visitors, a gap between the two can be chance, so treat it as a signal and confirm it with real sales or pre-orders.
The startup landing page covers how to build a page as a test, and why sign-ups alone aren't validation.
A second invented example: a service that scans a shoebox of family photos into an online album shows €39 a box to one batch of visitors and €69 to the next, each with a €10 deposit. If about as many pay at €69, keep selling at €69 to check it holds, and call the people who paid to ask why.
Free plans and free trials are not free pilots
I'm against free pilots. I'm not against free plans or free trials. Those are two different things: a free plan and a free trial are part of a business model, and a free pilot isn't.
- A free plan is limited on purpose, so there's enough inside it to make people want what sits behind the limit. You won't convert everyone, but it will convert.
- A free trial always has a clear end. That can be a date, a number of tokens, or whatever suits the product. When it ends, they pay or they stop.
- A free pilot, the kind I'm against, is free work for a business with no price and no end. In my experience it gets you a logo for your website and little more. The person who asked for it often moves on, and nobody with a budget has decided anything.
A pilot with an agreed price and a date when payment starts is not a free pilot. If you'd like to be generous with a business, agree the money first:
- a paid pilot: a set length, a real price and one result you promise to deliver;
- a discount with a stated reason and a set end, for example half price, locked in for a year;
- a free start, with the price agreed and a date when payment begins.
How to get your first customers covers how to agree the last two before any work starts.
Why most first prices are too low
In my experience, most first prices are too low, and the same reasons keep coming up. You price the rough version on your screen, bug list and all, while the buyer prices what it does for them. You price for your own wallet: as a founder with little cash, every euro feels big, while the buyer is spending a company budget. And a low price spares you a no. It feels like the sensible choice. Often it's fear.
It costs you more than the money:
- It sends a signal. Buyers see a cheap price and wonder what's missing.
- It brings the wrong customers. In my experience, the ones paying least tend to ask for most.
- It leaves you no room. Start at the bottom and you have nothing left to offer when a buyer pushes for a better deal.
- It tells you less. A first small payment still proves someone will pay, which is why you charge your first customer. But it says little about what the problem is worth, so don't set the price for the next buyers from it.
Say the number, then stop talking
Say it to an empty room first, as one short sentence that ends there: "It's €220 a month." Does your voice go up at the end, or do you rush it? Repeat until it sounds ordinary. A number you flinch at in private will wobble in the meeting, and the buyer will hear that.
In the meeting, say the number and stop talking. Let the silence sit.
Name a price early: it's the sharpest test you have
A friendly meeting where the buyer nods and asks when it launches proves very little, because interest costs them nothing. A real offer with a real number makes them decide, which makes it the strongest test of whether the problem is worth paying to fix. How to validate a startup idea sets out what counts as proof, with money at the top.
"It isn't ready yet" is the usual reason to wait. Charge for what you can deliver today, and leave the roadmap out of it. A painful enough problem lets a rough product sell, as long as it works: that's the hair-on-fire point in how to build an MVP. If it helps the buyer say yes, offer a paid pilot.
Reading the answer
| What they say | What it usually means | What to do |
|---|---|---|
| Yes, straight away | You were below what they'd have paid | Take it, and raise the price for the next buyer |
| Questions, like "Would it work with what we use now?" | They're picturing using it | Answer them. Don't offer a discount nobody asked for |
| "That's too expensive" | They're asking about value. It isn't a refusal | Ask "Compared to what?" (next section) |
| Nothing at all | The problem isn't big enough for them to pay to fix | Go back to the problem, and leave the price alone |
Silence is the real no, and a lower price won't make the problem any bigger.
"That's too expensive": ask "Compared to what?"
Keep your tone warm and ask: "Compared to what?" Often the honest answer is doing nothing. It costs nothing up front, and it's the competitor you'll meet most often. Then give them back their own number, the one from the four questions. In the moving example: "You said the visits take about 30 hours of your month." Say it calmly and stop. The gap between their number and your price makes the case.
Never discount for nothing, and decide your floor first
A discount given because the buyer pulled a face says your first number was made up, and every later deal will start from the discount. If the price moves, something else moves with it:
- they take a smaller package,
- they become a reference other buyers can call, or
- they pay a year in advance.
They still get a win, and your price still means something. A year up front also helps your cash: bootstrapping vs fundraising shows how billing in advance lets a company grow on its customers' money.
Before the meeting, decide your floor: the lowest number you'll accept, and what you'll ask for in return at each step down to it. Far more deals are lost to improvising under pressure than to a high price. In the moving example, the floor might be €180 a month, and only for a year paid up front.
Treat your first price as a hypothesis, with a review point
Your first price is a guess you're testing. Decide before you start when you'll look at it again: after the first five customers, for example. If every buyer said yes without a pause, the price was too low.
If you're going to be wrong, be wrong on the high side. Too high costs you a few deals, and a quick price change fixes it. Too low is baked into every contract you've already signed.
Price is the first assumption in startup financial projections, where each assumption carries a one-line reason. A price worked out from the buyer's value is one you can defend when an investor asks where it came from.
How to raise your price
- Start with new customers. The next person to sign pays the new price, so you see whether it holds before touching anyone who already pays you.
- Give existing customers at least 30 days' notice, or more if their contract says so. Tell them yourself, in your own words, never as a line at the bottom of a newsletter.
- Give the real reason. What do they get now that they didn't get a year ago?
- Let them come out ahead. For example, they stay on today's price for 12 more months if they switch to paying yearly now.
Some customers will leave. In practice fewer go than founders expect, and the ones who go are often the cheapest and the most demanding.
Ready to Move Faster?
Get actionable feedback and insights on your challenges and goals. 30 minutes focused on helping you move forward.
No sales pitch. No obligations. Just helpful advice.