Pricing surveys: Van Westendorp and Gabor-Granger explained

Use Van Westendorp's four price questions to find the range of prices your market will accept, then a Gabor-Granger purchase ladder to choose a price inside that range. Van Westendorp tells you where the edges are; Gabor-Granger estimates how demand and revenue change between them. Both rely on what people say rather than what they pay, so confirm the answer with real buyers before you change a price list.

This guide gives the exact questions for both methods, shows how to read the curves with a worked example, covers sample sizes, and lists the limits a finance lead will raise.

The four Van Westendorp questions

The Price Sensitivity Meter, introduced by the Dutch economist Peter van Westendorp in 1976, asks each respondent for four prices instead of one. Asking "what would you pay?" invites a single anchored guess; asking for four thresholds shows where the product starts to feel cheap, good value, expensive and out of reach.

Three rules make the answers usable:

1. Describe exactly what is being priced before the questions: what is included, the billing unit (per month, per year, per seat) and the currency. Two people pricing different things produce noise.
2. Show no price anywhere in the survey before these questions. A number in the introduction becomes the anchor.
3. Check the order. Each respondent's prices must rise from too cheap to a bargain to getting expensive to too expensive. Remove answers that do not, and remove obvious jokes (a price of 1,000,000) before you plot anything.

How to read the Van Westendorp chart

Plot four cumulative curves against price, then read the range of acceptable prices between two crossings. For every price on a grid (say every dollar or every five dollars), count:

Too cheap at p          = share whose "too cheap" price is at or above p
A bargain at p          = share whose "bargain" price is at or above p
Getting expensive at p  = share whose "getting expensive" price is at or below p
Too expensive at p      = share whose "too expensive" price is at or below p

The two "cheap" curves fall as the price rises and the two "expensive" curves climb. Where they cross gives four reference points.

The range of acceptable prices runs from the PMC to the PME. Practitioners draw the two marginal points in two ways: the version above, and a version that first inverts the bargain and getting-expensive curves into "not a bargain" and "not expensive" before crossing them with the too cheap and too expensive curves, which moves the edges. Both are in use. Pick one and name it in your report.

A worked example (illustrative numbers)

The table below is made up to show the reading, not taken from a study. It shows the share of respondents at each price for a team software subscription priced per seat per month.

Reading the crossings, with straight-line interpolation between rows:

• Too cheap falls below getting expensive between $15 and $20, at about $18. That is the PMC.
• Too cheap meets too expensive between $20 and $25, at about $22. That is the OPP.
• A bargain meets getting expensive between $20 and $25, at about $24. That is the IPP.
• Too expensive rises above a bargain between $25 and $30, at about $28. That is the PME.

So this sample accepts roughly $18 to $28, with $22 and $24 as reference points. A current price of $35 would sit above the range: more of this sample rejects it as too expensive than sees it as a bargain. That does not prove $35 is wrong, because the method cannot see how buyers weigh your product against alternatives, but it is a strong reason to test it.

The OPP is not the revenue-maximizing price, despite its name. It is the point of least resistance in the sample. To estimate revenue, add a Gabor-Granger ladder or the purchase-likelihood questions of the Newton-Miller-Smith extension, which asks how likely each respondent is to buy at their own bargain and getting-expensive prices.

Gabor-Granger: the purchase-likelihood ladder

Gabor-Granger shows one price at a time, asks how likely the person is to buy at that price, and moves up or down a ladder of prices depending on the answer. The method is named after the economists André Gabor and Clive Granger, who described it in the 1960s. Its output is a demand curve (the share who would buy at each price) and a revenue curve.

How to run it:

1. Pick five to seven prices that span the plausible range. The Van Westendorp range is a good source.
2. Start each respondent at a middle price (or a randomly chosen one, to spread the anchoring). If they would buy, show the next price up; if not, the next price down. Stop when the answer changes.
3. Decide what counts as "would buy". Stated intent overstates real purchasing, so count only "Definitely would buy", or apply a discount you can defend, and use the same rule every time.
4. For each price, multiply the price by the share who would buy. The peak of that revenue index is the revenue-maximizing price in this sample. If your costs rise with each customer, use price minus unit cost instead.

In this illustrative example the index peaks at $25, inside the Van Westendorp range from the earlier table. The gaps between neighboring prices are smaller than the margin of error of a typical sample, though, so the survey cannot reliably separate $20, $25 and $30. That final choice belongs to a live test.

Which pricing method to use

Match the method to the decision. The table covers the common pricing questions and what each method can and cannot tell you.

How many respondents a pricing survey needs

Each curve and each price point is a percentage with its own margin of error, so plan for at least 100 to 200 completed answers from the target segment, and that many again for every segment you want to compare. That is a common working rule rather than a formula: below it, the crossings move noticeably when a handful of answers change. For Gabor-Granger, remember that a 40% share from 100 answers carries a 95% margin of error of about plus or minus 10 points at each price.

Ask the people who would actually make the purchase: target buyers in the category for a new product, and current customers separately from prospects for a price change, since current customers anchor on what they pay today. Before a price change, check whether price is really why customers leave; the reason mix from a cancellation survey shows it. Use the sample size calculator to set the target, and the survey sample size guide for the reasoning.

The limits of stated-preference pricing

Every method on this page measures what people say about prices, which differs from what they do when money leaves their account. Use the results to narrow the options and spot problems, not to set a price on their own.

• No money changes hands, so answers are hypothetical. People can be generous or stingy on paper in ways they would not be at checkout.
• Anchors move the answers: the price they pay today, competitors' prices, a number in the introduction.
• Customers asked about a price change have a reason to answer low.
• Van Westendorp and Gabor-Granger show the product alone. Real buyers compare it with alternatives, including doing nothing.
• Respondents can only price what they understand. For a new category, show a demo or a clear description first.
• An average curve hides segments with very different willingness to pay. Segment by company size, use case or plan before reading the crossings.
• The results describe the sample. Panel respondents who will never buy may price very differently from qualified prospects.

Confirm a survey result with behavior: a test on the pricing page, a price stated in sales conversations for a quarter, or a pilot with new customers only. If the survey and the test disagree, trust the test.

Common mistakes

• Pricing a vague concept. Describe what is included, the billing unit and the currency before the questions.
• Keeping answers that are out of order. Validate that the four prices rise, and remove those that do not.
• Calling the OPP "the optimal price" in a board deck. Report the range, and use Gabor-Granger or a live test to choose within it.
• Mixing current customers with prospects. Analyze them separately.
• Showing a price before the pricing questions. Remove every number from the introduction and earlier questions.
• Treating "probably would buy" as a sale. Count only the top answer or apply a discount you can defend.
• Reading the curves from 40 answers. Collect enough per segment that the crossings stop moving.

Doing this in Claform

Claform has a Price Sensitivity question type for Van Westendorp, available on Professional and higher. Respondents see four price fields labeled Too cheap, A bargain, Getting expensive and Too expensive, each with a one-line prompt and the currency symbol you set, and you can set a minimum and maximum price. When the four prices do not rise in order, the respondent sees a warning and the answer is not accepted until it is corrected. In the results, the question shows the four curves with the optimal price point, the indifference price point and the acceptable range, calculated with the crossings in the table above. For Gabor-Granger, build the ladder from rating questions, one per price, with skip logic that moves up after a positive answer and down after a negative one, and use the MaxDiff question type to rank features for packaging. See Claform for market research and pricing for what each plan includes.

Frequently asked questions

What are the four Van Westendorp questions?

At what price would the product be so expensive you would not consider buying it, so cheap you would doubt its quality, starting to get expensive so that you would think before buying, and a bargain, a great buy for the money. Describe the product, billing unit and currency first, and check that each respondent's prices rise from too cheap to bargain to getting expensive to too expensive.

How many respondents do you need for a Van Westendorp survey?

Plan for at least 100 to 200 completed answers from the segment you are pricing for, and the same again for each additional segment you want to compare. This is a working rule, not a formula. With fewer answers the curves are jagged and the crossings move noticeably when a few people change their answers.

Is the optimal price point the price I should charge?

Not on its own. The optimal price point is where equal shares find the product too cheap and too expensive, which marks low resistance, not maximum revenue or profit. Treat the range from marginal cheapness to marginal expensiveness as the shortlist, estimate demand with Gabor-Granger, and confirm with a live test.

What is the difference between Van Westendorp and Gabor-Granger?

Van Westendorp asks each respondent for four price thresholds and gives you a range of acceptable prices. Gabor-Granger asks how likely they are to buy at specific prices and gives you demand and revenue curves. Use Van Westendorp early to find the range and Gabor-Granger to compare a few prices inside it.

Can you use Van Westendorp for B2B or subscription pricing?

Yes, if the unit is precise: per seat per month, per year with annual billing, or per order. Survey people who influence or approve the purchase, and segment by company size, since a small team and an enterprise buyer read the same price very differently. Confirm the result in sales conversations before you publish a new price.

Start building for free