Guide6 min readUpdated Aug 4, 2026

Sometimes the fastest way to sell more is to charge more

Most teams shave price to win the yes. Sometimes the opposite is true. Here's why willingness to pay defies surveys, and how to find your real number by watching what buyers do.

Kate O'Keeffe
Kate O'Keeffe
CEO & Founder, Heatseeker

In short: Willingness to pay is the highest price a customer will actually accept, not the number they give in a survey. Because price also signals value, a higher price can increase demand, so stated answers mislead. The reliable way to find it is to observe real purchase behavior in market, under your own brand.

Most people treat price as the thing you shave to win the customer. Lower is safer. Lower is friendlier. Lower gets the yes.

Sometimes. And sometimes the opposite is true, and charging more is the thing that makes people want it.

I know how that sounds. Stay with me, because the evidence is better than the intuition.

In 2008 a group of researchers put people in a brain scanner and gave them wine to taste. They were told the wines were different, sold at different prices. They weren't. Two of the "different" wines were the same wine, just labeled cheap or expensive. People said the pricier one tasted better, and it wasn't only politeness. The pleasure center of their brain actually lit up more when they believed the wine was expensive. The price didn't change the wine. It changed the experience of the wine. (Plassmann et al., PNAS, 2008, if you want to send it to your CFO.)

Or take Williams-Sonoma. They launched a bread maker at $275 and it just sat there. So they added a fancier model at $429. The $429 one barely sold, but sales of the original $275 machine nearly doubled, because next to the expensive one it suddenly looked like the smart buy. The number on the shelf didn't just measure demand. It created it.

That's the thing about price that spreadsheets miss. Price isn't only what you take from a customer. It's information you give them. It tells them what to expect, where you sit, whether to take you seriously. In plenty of categories a low price isn't a bargain, it's a warning. Charge too little and people assume you're cheap for a reason.

You can't survey your way to a price

Which is exactly why you can't ask your way to the right number.

Ask someone what they'd pay and they'll anchor low, because who volunteers to pay more? Nobody has ever told a researcher "honestly, charge me 20% more and I'll want it more." But that is how people behave all the time. It's the say/do gap, and on price it's a chasm.

So don't ask. Watch.

And watch under your own brand, because pricing is the one test you can't run in disguise. What someone will pay is inseparable from who's asking. The goodwill you've spent years building is baked into the number. Strip your name off and you're measuring a stranger's product, not yours.

What a real price test looks like

Here's one that stuck with me. A sneaker brand was launching a new colorway. The latest tariffs had squeezed margins to almost nothing, so there was no room to get it wrong. They had to order the right quantity, of the right shoe, in the right color, to the right store, and land on a new price, all at once. Miss on any one of those and the whole drop loses money.

So they ran the new sneaker in market against an existing one whose sales volume they already knew, broken out by zip code. One experiment, two answers. The known-volume control told them how much to make and where to send it. And the price told them something they hadn't expected: the higher price point won. In every market.

Not "people said they'd pay more." People chose it, at the higher price, everywhere. The inventory problem and the pricing problem, solved in a single read. It's the same behavioral logic behind concept testing: let the choice, not the opinion, tell you what's true. (You don't even have to sell at the test prices to learn this, but that's a trick for another day.)

So, what is your price actually doing?

Before you discount your way into a launch, ask what your price is really doing. Are you shaving it to feel safe? Or have you actually watched what a higher number does to demand?

Stop asking what your customers will pay. Watch what they pay, under your own brand, and let the number surprise you.

That's what Heatseeker is built for. Bring a pricing decision that matters and we'll help you find the real number by watching real buyers, not asking them.

Bring a price you're not sure about.

Book a 15-minute demo and we'll show you what your market actually does.

Book a 15-minute demo →

References

  1. Plassmann, H., O'Doherty, J., Shiv, B., & Rangel, A. (2008). Marketing actions can modulate neural representations of experienced pleasantness. Proceedings of the National Academy of Sciences, 105(3), 1050 to 1054. pnas.org
  2. Williams-Sonoma bread-maker anchoring effect (Farnam Street). fs.blog

Frequently asked questions

What is willingness to pay?

Willingness to pay is the highest price a customer will actually accept for a product. It's a behavior, not the number people give when asked, which tends to be lower and unreliable.

Why can't you measure willingness to pay with a survey?

Because people anchor low when asked, and nobody volunteers to pay more. Stated price and actual behavior are two different things, the say/do gap, and it's especially wide for price.

Can raising the price increase demand?

Yes, in some categories. Price signals quality and status, so a higher price can make a product more appealing, and a price that's too low can read as a warning. A 2008 PNAS study found people experienced identical wine as more pleasant when it was labeled more expensive.

How do you test willingness to pay properly?

Put real prices in front of real buyers, in market, under your own brand, and observe what they actually choose. Pricing has to be tested branded because willingness to pay is tied to trust in the brand.

Why does a price test have to be branded?

Because willingness to pay is inseparable from brand goodwill. Testing an unbranded or lookalike version measures a different product, not yours.

Keep reading