Will people actually pay? The signals that count before you build

You find out whether people will pay by asking them to pay — not by asking whether they would. Everything short of a payment costs the person giving it less, and tells you less: praise costs nothing, a visit costs a click, a waitlist signup costs an email address. Those can tell you the pain is real and that you can reach the people who have it. Only money, taken before you build, tells you the price works.

Why "would you pay?" gets a yes

The question costs nothing to answer yes. A no costs something — it disappoints a person who is visibly excited, to their face or in their inbox. So the cheap answer wins, and it wins with strangers as reliably as with friends. That is not dishonesty. It is people answering a hypothetical with the answer that ends the exchange pleasantly.

This is why "everyone says it's great but nobody buys" is not a paradox. You asked two different questions. The first — do you like this? — was free to answer. The second — will you hand over money? — was not, and you only asked it later. The gap between the two answers is the expected result. It is not evidence the idea is bad. It is evidence that the first answer was never evidence.

So do not ask whether people would buy. Put a price on a page and see who pays it. Our own validation plans are not allowed to contain a call or a customer interview — the output is scanned and rejected if it does — for exactly this reason: a hypothetical question invites a polite answer, and a polite answer is graded by the person who wanted to hear it.

The ladder, rung by rung

Each rung costs the person more than the one below it, and the three questions a signal can bear on — is the pain real, can you reach them, will they pay this price — fill in from left to right as you climb.

• Praise, likes and followers. They cost nothing, and a validation plan cannot even write them down: of the 5 things our plans can set a bar on, none is a compliment. That is deliberate. A result you cannot pin to a number before the test is a result you will interpret after it. Follower growth is the same thing at scale — attention, not buyers.
• A visit. Measurable, and still silent. Traffic tells you the post got read, not that anyone wanted what it described, so a visitor count on its own evidences nothing about the idea.
• A waitlist signup or a reply. The first rungs that cost something. Someone handed over an address, or took the time to write back. That is fair evidence the pain is real and that your channel reaches people who have it. Neither says a word about price, because nobody was shown one and asked to act on it.
• A payment. The only rung of the 5 that reaches the third column, and the one where the person gave up something they would rather have kept.

Is a waitlist validation?

It is the first rung, not the last. A waitlist tells you that people with the problem found your page and wanted to hear more. It does not tell you they will pay what you plan to charge.

Three things make one worth more:

• Make joining cost a step. An address confirmed through an emailed link is a person. A typed address with no confirmation is the weak version of this signal, and easy to inflate.
• Show the price before they join. A signup under a visible price has at least read it. A signup under "coming soon" has agreed to nothing.
• Then ask the list to pay. The number that matters is not how many joined — it is how many pay when you open the door. A long list where nobody shows up when invited is not a weak yes. It is a no that took longer to arrive.

Our own sample run stops at this rung

Here is the part that is ours to admit. The published sample Discovery run ends with 5 finalist ideas, each with a planned first move. All 5 ask for an email. Not one asks for money.

The plan's own shape leans that way: it has a field for a waitlist target and none for a price, even though a payment or a pre-order is allowed as the first move. For a sweep that ranks many ideas at once, a cheap first rung is a defensible first move — you find out whether anybody is there before you ask anyone for money. But it leaves the price open, and a finalist that clears its waitlist bar has climbed to the middle of the ladder, not the top.

The price question gets answered one step later, on the idea you actually pick. The validation plan for a single idea costs 1 credit, puts exactly one action on the page, and is asked to choose the strongest one you can realistically collect. 3 of its 4 options take money.

Should you charge before you build?

If you want to know whether the price works, yes. No other rung answers that question. Here is what each of the options asks the buyer to give up:

Taking money for something that does not exist yet is only uncomfortable when it is vague. Three things make it honest:

• Say what they are paying for, and when it ships. A pre-order with a scope and a date is a promise you can keep. "Early access" with no date is one nobody can hold you to — including you.
• State the refund terms on the page, before the button. If it does not ship, the money goes back. Written up front, that costs nothing to honor, and it removes the one reason to feel uneasy about asking.
• Charge the price you mean to test. A token deposit tells you who will reach for a card. It does not tell you whether your price works, because nobody paid it. If the question is the price, the price is what goes on the button.

That last one is the trade-off to take with open eyes. A small deposit gets more yeses and answers a smaller question. The real price gets fewer and answers the one you have.

When nobody pays

Write the kill line before the page goes live — the metric, the bar and the deadline — for the reason the guide on pre-registered tests gives: a bar set after the numbers arrive moves toward them. Then read a miss by where the people stopped, because the three places mean three different fixes:

Only the middle row can mean the pain is not there, and even then a sharper promise to the same people is the cheaper test before you conclude it. The bottom row is the one most often misread: people wanted to hear more, so the pain was real enough to act on. It was the offer they turned down.

How many payments are enough is not a feeling either. It comes from how many customers your price requires to reach the income you want — a division you can do before the test, not after it. A handful of pre-orders from one forum post is a start, not a market.

What a payment still does not tell you

A payment proves that someone paid for the promise on your page, at that price, through that channel. It does not prove they will stay, that serving them fits in your week, or that the thing you build will match what they paid for. Support load, retention and platform exposure are separate questions, and no rung on this ladder reaches them.

That is the honest scope of everything above. It is not a claim that a payment predicts success. It is a claim about which question each signal is able to answer — and a method that let a compliment answer the price question would be doing exactly what the encouraging conversation does.

The short version

• Asking whether people would pay collects the cheapest answer available. Ask them to pay instead.
• A signal tells you roughly what it cost the person giving it. Praise costs nothing and answers nothing.
• Signups and replies can show the pain is real and the channel works. Only money shows the price does.
• A waitlist is the first rung. Confirm the address, show the price, then ask the list to pay.
• Our own sample run's plans stop at the email rung. The price question is answered on the idea you pick, where 3 of the 4 options take money.
• Write the kill line before the page goes live, and read a miss by where people stopped.