Van Westendorp Price Sensitivity Meter: How to Run It

By Harry Nguyen · Updated Oct 11, 2026

The Van Westendorp Price Sensitivity Meter (PSM) is a pricing survey that asks each respondent four open-ended questions: at what price is the product too cheap, a bargain, getting expensive, and too expensive. You plot the answers as cumulative curves, and the points where those curves cross mark an acceptable price range and a few reference prices. It tells you how buyers perceive price, not how many will buy.

Key Facts: Van Westendorp at a Glance

• Dutch economist Peter van Westendorp introduced the technique in 1976, in a paper in the Proceedings of the 29th ESOMAR Congress. (pricesensitivitymeter reference manual, CRAN)
• The method asks four price questions and returns a range of acceptable prices bounded by two points, the point of marginal cheapness and the point of marginal expensiveness. (CRAN)
• On its own it does not estimate demand. The Newton-Miller-Smith extension adds two purchase-likelihood questions so you can estimate reach and a revenue curve. (CRAN)
• One survey vendor recommends at least 100 completes for either Van Westendorp or Gabor-Granger, and at least 100 per segment if you plan to compare audiences. (QuestionPro)
• Sawtooth Software cautions that the line-crossing analysis lacks strong economic theory and recommends conjoint or monadic pricing for established products. (Sawtooth Software)

What the Van Westendorp Price Sensitivity Meter Measures

Most pricing debates are guesses with a spreadsheet attached. Someone says $49, someone else says $79, and the loudest voice wins. The PSM replaces that with a quick, cheap piece of evidence: what a sample of your target buyers consider too low to trust and too high to consider.

Notice what it doesn't do. It doesn't show a demand curve, it doesn't account for competitors, and it doesn't know your costs. Think of it as a boundary-finder. It tells you the zone where a price won't trigger either "this must be junk" or "forget it, too much." Where you land inside that zone is a strategy call, which is where value-based pricing and your pricing strategy take over.

It works best when buyers have no strong price anchor, such as a new product or a new tier. For categories with obvious market prices, the answers tend to cluster around those prices and tell you little you didn't already know.

The Four Questions

Each respondent answers all four, in the same order, with a number. Open-ended answers are the point: nobody is shown a price to react to. The wording below follows the common version of the method (Sawtooth Software).

Before the questions, show a short, neutral description of the product, with no price in it. The description matters more than people expect, because every answer depends on what the respondent imagines they're buying.

How to Read the Curves and Intersections

Once responses are in, build four cumulative curves across a price axis. For each price, calculate the share of respondents whose answer falls at or below it (for the "expensive" and "too expensive" questions) or at or above it (for the inverted "too cheap" and "cheap" questions).

The standard method inverts two of the four cumulative frequencies so that four intersection points exist, and conventional practice inverts "too cheap" and "cheap / good value" (CRAN). This article uses that convention. Sources differ on which curves define each point, so check the convention before you compare your output with anyone else's.

The range of acceptable prices runs from the PMC to the PME (CRAN; Sawtooth Software). Don't over-read the name "optimal." The OPP is the price where the same share of people exceed their lower or upper limit. It isn't the revenue-maximizing price, and it isn't the profit-maximizing price.

A worked example (illustrative numbers)

Say you survey 200 buyers about a new analytics add-on. These figures are made up to show the mechanics, not a benchmark for any product.

The acceptable range is $29 to $61. A launch price around $42 to $48 sits in the middle of that zone. If your cost structure needs $65, the survey is telling you the market will push back, and you'd want to redesign the package or find a segment that values it more.

Adding Purchase Likelihood: The Newton-Miller-Smith Extension

Van Westendorp didn't address demand estimation. The Newton, Miller and Smith (NMS) extension, published in 1993, adds two follow-up questions asking how likely each respondent is to buy at the price they called "cheap" and at the price they called "expensive," on a five-point scale (CRAN). Combined with the price answers, those responses produce a revenue curve that points to a revenue-maximizing price.

Two caveats. NMS assumes no purchases at the "too cheap" and "too expensive" prices, and Sawtooth Software calls the too-cheap assumption probably not realistic (Sawtooth Software). And stated intent isn't behavior, so the extension gives you a directional curve, not a forecast.

Van Westendorp vs Gabor-Granger vs Conjoint

These three get mixed up constantly. They answer different questions, so the choice usually comes down to what you already know about the product and market.

Gabor-Granger was developed in the 1960s by economists Andre Gabor and Clive Granger. It shows each respondent a series of prices, one at a time, and asks how likely they are to buy at each, which lets you plot demand and revenue curves (QuestionPro). A QuestionPro webinar describes it as generally better for product extensions, with Van Westendorp fitting new products better (QuestionPro). Sawtooth Software points out that conjoint can cover thousands of potential product formulations in a single study, while Van Westendorp is typically limited to one or a few (Sawtooth Software).

A practical sequence: use Van Westendorp to find the range, then test specific price points inside it with a pricing experiment on real traffic.

How to Run a Van Westendorp Study, Step by Step

1. Define the product and the buyer. Write a two or three sentence neutral description and name the audience. A study of "everyone" produces a price nobody pays.
2. Choose a survey tool and screen respondents. Include a screener so only people who'd plausibly buy answer. Segment-level questions (company size, role, current spend) help later.
3. Ask the four questions in order, with numeric inputs. Add a validation check: the answers should rise as you go from too cheap to too expensive. Drop respondents whose answers contradict that order.
4. Add the NMS questions if you want a revenue view. Two five-point purchase-likelihood items, one at each of the respondent's own "cheap" and "expensive" prices.
5. Collect enough completes. The one sample-size recommendation worth citing is a vendor rule of thumb: at least 100 completes, and at least 100 per segment you plan to compare (QuestionPro). Treat it as a floor, not a guarantee.
6. Build the cumulative curves and find the four intersections. A spreadsheet or the survey tool's built-in analysis does the job.
7. Cut the data by segment. A range that looks wide overall often splits into two narrower ranges for two different buyer groups.
8. Turn the range into a decision. Pick candidate prices inside the range, then test them against cost, competitor prices, and positioning.

Ready-to-Copy Survey Template

Paste this into your survey tool. Replace the bracketed text.

Intro screen: "Please read the description below. [Two or three sentences describing the product and who it's for. No price.]"

Q1 (too cheap): "At what price would you consider [product] so inexpensive that you'd question its quality? ($ per [month/seat/unit])"

Q2 (bargain): "At what price would you consider [product] a bargain, a great buy for the money?"

Q3 (getting expensive): "At what price would [product] start to feel expensive, though you'd still consider buying it?"

Q4 (too expensive): "At what price would [product] be so expensive that you wouldn't consider buying it?"

Optional NMS items: "How likely would you be to buy at [respondent's Q2 answer]?" and "How likely would you be to buy at [respondent's Q3 answer]?" (1 = definitely not, 5 = definitely would)

Then lay out your analysis sheet like this:

Read the sheet row by row. The first price where the "too cheap" column drops below the "expensive" column is near your PMC. Where "too cheap" and "too expensive" cross is your OPP. Repeat the same logic for IPP and PME.

Examples by Business Function

These scenarios are illustrative, showing how different teams use the same four questions.

For the structure behind those add-ons and tiers, see SaaS pricing models, and for how tiers line up from a free entry offer to a premium one, see the value ladder. And once you have a number, pricing page optimization covers how to present it.

Limitations to Respect

The PSM is easy to run, which is partly why it's easy to misuse.

• Stated, not revealed. Respondents say what they feel, not what they'd pay at a checkout.
• No competitive context. People judge price in isolation unless you prompt them to think about alternatives (Sawtooth Software).
• Weak theory. Sawtooth Software notes that the line-crossing approach has been criticized as lacking good theory and doesn't follow from economic theory (Sawtooth Software).
• No demand estimate without extension. The original method doesn't predict volume; reach and revenue estimates come from the NMS purchase-intent questions (CRAN).
• Unrealistic answers. Open-ended price questions can invite illogical responses, so clean the data first.
• Ignores cost and margin. A range tells you nothing about whether you can profitably sell inside it.

Van Westendorp Dos and Don'ts

Do

• Describe the product in neutral terms. Leave out price anchors and marketing hype.
• Screen for real buyers. Ten engaged prospects beat a hundred random clicks.
• Cut results by segment. Averages hide the split between small and large accounts.
• Treat the range as input, not an answer. Combine it with cost, competitor prices, and a live test.
• Run it again after a major product change. The range moves when the product does.

Don't

• Don't show a price before the questions. It anchors every answer.
• Don't call the OPP the best price. It's a balance point between two kinds of rejection.
• Don't mix conventions. Pick one curve convention and state it in your report.
• Don't use it alone for an established product. Add a method that sees competing choices.
• Don't skip data cleaning. Contradictory rows distort the curves.

Where This Fits in a Growth Program

Pricing research is one input to a broader growth experimentation framework: the survey generates a hypothesis about where the acceptable range sits, and a live test confirms it with real buyers. Run them in that order and each does the job it's good at.