The willingness-to-pay evidence ladder: which signals actually predict revenue

Ask ten people whether they would pay for your idea and most of them will say yes. A few will even name a price. That feels like demand, and usually it isn't.

People aren't lying to you. A hypothetical question gets a hypothetical answer. Saying "I'd pay $20 a month for that" costs the person nothing and lets an awkward conversation end on a friendly note, and neither of those has much to do with whether they'll enter a card number three months from now.

So I stopped asking whether people will pay. I sort what they have already done, and what they're willing to do right now, into a ladder. The higher the rung, the better it predicts revenue. The rule I take from it is to count the rungs, not the yeses.

The six rungs

Rung	Signal	What it proves	What it does not prove
0	Compliments: "love it", "great idea"	That they are polite	Anything about demand
1	Stated intent: "I'd use it", "I'd pay $X"	The idea is understandable and not offensive	That they will pay, or how much
2	Time already spent on the problem	The problem is real and recurring for them	That they will spend money on it
3	Money already spent on the problem	A budget exists and the pain is bad enough to pay for	That they will switch to you
4	Something scarce committed now	They value your specific approach enough to give something up	The final price
5	Money now	Real willingness to pay, at a real price	That it scales beyond early adopters

Most early research collects rungs 0 and 1. Rungs 2 and 3 describe the past, which is why I trust them: you're asking about behaviour that already happened. Rungs 4 and 5 happen in the present, and they're the only ones where the person gives something up because of you.

Rung 0: compliments

"This is a great idea" is what you'll hear most, and it tells you the least. It reflects how the conversation went. If an interview ends with nothing but compliments, it went badly, even if it felt good at the time.

When you get one, use it as a cue for a better question: "Thanks. When was the last time this problem actually came up for you?"

Rung 1: stated intent

This is where "would you pay?" lives. Its one honest use is checking that the idea makes sense to people. If nobody can even picture using it, you want to know.

It can't predict purchase. Stated intent runs well ahead of behaviour because the question strips out every real cost: the effort of switching, getting a budget approved, the fact that the problem only hurts on the last Friday of the month.

If you collect it anyway, ask for a range and compare it with rungs 2 and 3 from the same person. Someone who says "$30 a month" while spending no money and no time on the problem today is giving you a polite number.

Rung 2: time already spent

Here the evidence starts to mean something. People spend time on problems that hurt. They build a spreadsheet, chase invoices by hand, copy and paste between three tools, or ask a colleague the same question every week.

Ask:

• "Tell me about the last time this happened. What did you do?"
• "How long did that take?"
• "How often does it come up?"

A messy workaround is one of the best signals you'll find in customer discovery, because the person cared enough to invent something. When there's no workaround at all, the pain is often too mild for anyone to act on.

Rung 2 can't tell you whether that time will turn into money. Plenty of people spend hours on something they'd never pay to fix, especially when their own time feels free to them.

Rung 3: money already spent

Next, ask what they already pay to deal with the problem today.

• "What do you use for this right now? Do you pay for it?"
• "Have you ever paid someone to handle this?"
• "Is there a budget line this would come out of?"

Existing spend is the strongest evidence you can get before you have a product. It shows the problem is bad enough that someone already pays to deal with it, and it gives you a real price anchor. A person paying $40 a month for a clunky tool has told you more about pricing than any survey slider will.

It still doesn't show they'll switch to you. Switching has its own costs, and that belongs to a separate hypothesis (solution fit) that you test on its own.

Rung 4: something scarce, committed now

Before there's anything to sell, you can still ask for something that costs the person something real:

• Thirty more minutes of their time to try a prototype
• An introduction to the person who owns the budget
• Access to their actual data or workflow
• A signed letter of intent that names a price
• A place on a waitlist that asks for more than an email address, such as five questions about how they handle this today

Watch whether they agree to something that isn't free for them. If someone hesitates at a second call, that tells you how much the problem matters to them, and it's worth knowing.

Rung 5: money now

The top rung is money before the product is finished: a pre-order, a deposit, a paid pilot, or a founding-customer price paid upfront. It's the one signal that answers "will they pay?" outright, because they did.

The amount can be small. A refundable deposit filters out a surprising share of the rung-1 yeses. And if asking for money before the product exists makes you nervous, pay attention to that, because it says something about how confident you are in the offer.

Write your hypothesis against a rung

Most willingness-to-pay hypotheses are written so that a rung-1 answer confirms them, which is why they so rarely fail. Tie the hypothesis to a rung and you can test it properly.

Weak:

Freelance designers would pay for a tool that chases late invoices.

Testable:

At least 4 of 10 freelance designers we talk to already spend more than 2 hours a month chasing late payments (rung 2), and at least 2 of them pay for a tool or service that partly does it today (rung 3).

The second version names the customer, the rung and a threshold you set before collecting anything, so it can come back false. There's more on structure in how to write a startup hypothesis.

What ten responses can and cannot tell you

Solo founders rarely get a hundred interviews. Ten is more typical, and ten is fine if you're clear about what it can show. Treat what follows as rules of thumb; none of it is statistics.

Ten responses won't give you a price point. Price-sensitivity methods need more data and a more representative sample. When you need a number and have the reach, use willingness-to-pay survey questions.

What ten responses can show you:

• Whether the higher rungs exist at all. If none of ten people in your target customer has spent time or money on the problem, that's a real result. Either the pain is too mild to pay for, or you're talking to the wrong customer.
• Who feels it most. Look at the two or three people highest on the ladder and what they have in common. Their role, company size or situation is often a sharper customer definition than the one you started with.
• What they pay today. Even two rung-3 answers give you a price anchor and a competitor you can name.

Reading the result, and what to do next

Read the ladder across all your responses rather than one person at a time.

• If several people are at rungs 3 to 5, the signal is strong. Test the offer next by asking for a deposit or a paid pilot at a real price.
• If most are at rung 2 and few reach rung 3, the problem is real but the budget isn't proven. Look for the segment that already pays. A narrower customer often has the same problem with money attached.
• If most answers sit at rungs 0 and 1, demand for this version is weak. Go back to the problem and ask about the last time it happened and what it cost them. If that keeps coming up empty, it may be the wrong problem for this customer, and an adjacent one could be the better bet.

Whichever result you get, it points to the next test. (For how many conversations to have before you decide, see how many customer interviews you need.)

If you want a structured way to do this, Hypothis turns your idea into testable hypotheses, including one for willingness to pay. It writes questions about past behaviour instead of hypotheticals and weighs each answer as evidence. The ladder works just as well on a notepad.