Pricing a new product is difficult when you have no sales history to guide your decision. Even small changes can affect profitability.
A McKinsey analysis of the average S&P 1500 company estimated that a 1% price increase could raise operating profit by about 8%, assuming sales volume remains unchanged.
The Van Westendorp price sensitivity meter helps businesses assess price acceptance before launch.
Developed by Dutch economist Peter van Westendorp in 1976, the method asks potential buyers four questions to identify a range of prices they consider acceptable.
However, the results reflect perceived value rather than actual purchasing behavior. A price that respondents consider acceptable may still fail to generate sufficient sales.
This guide explains how to conduct a Van Westendorp survey, interpret its four curves, assess purchase intent, and combine the findings with competitor pricing and profit margins to make a more informed launch decision.
The Van Westendorp price sensitivity meter is a survey technique that locates the range of prices a target audience considers credible for a product.
It does this with four price questions whose answers are plotted as cumulative curves, and the intersections of those curves mark the boundaries of acceptance.
Van Westendorp designed the meter in the mid-1970s to determine consumer price preferences without asking the one question buyers answer badly, which is how much they would pay.
Instead, it asks where a price starts to feel cheap, fair, expensive and unacceptable, and lets the pattern across respondents reveal the band.
That indirection is the point. Spreading the judgment across several anchored questions is the method's design case against a single willingness-to-pay question, though it does not remove hypothetical bias (see the limitations below).
The meter measures perceived value. It asks where a respondent's sense of value changes; it never asks whether they would buy at a given price.
That is the source of both its speed and its main weakness: you get a band of credible prices, with no demand curve or revenue forecast attached.
Use Van Westendorp when the offer is new and the plausible range is unknown, Gabor-Granger when candidate prices already exist and the question is demand, and conjoint when buyers must trade features against price.
The right method follows from the pricing question, so write that question in one sentence before briefing any panel.
Van Westendorp works without an established price, so its natural use is a new product, tier or line with no live price to test against. It gives a first map of the territory, which is what a launch team usually lacks.
A brand director planning a premium skincare range with nothing comparable in the portfolio needs to know, before pack and media briefs are written, what range buyers will find believable.
Gabor-Granger shows each respondent a specific price and asks whether they would buy, then raises or lowers the price until acceptance flips.
It produces a stated demand curve for a set of candidate prices, so it suits repricing an existing product or choosing between a handful of options already shortlisted.
Conjoint presents configured alternatives at different prices and forces a choice, so price is weighed against features and against rival options. The standard price sensitivity meter never forces that trade-off, which is its most frequently cited weakness.
Conjoint is usually the better fit when configuration is the question; the limitations section below explains why bundles strain the meter.
| Method | Question it answers | Best when | Weakness | Typical output |
|---|---|---|---|---|
| Van Westendorp | Where does the price stop feeling credible | New offer with no price history | No competitive or feature context; perception only | Acceptable price range and anchor points |
| Gabor-Granger | Will people buy at these specific prices | Candidate prices are known | Anchors respondents on the prices shown | Stated demand curve across tested prices |
| Conjoint analysis | Which features justify which price | Configurable bundles or tiers | Costly design and larger samples | Part-worth utilities and simulated share |
A SaaS team launching a single new add-on with no price history runs Van Westendorp first. The same team repricing three existing tiers is better served by Gabor-Granger for demand at known price points or conjoint for the tier structure itself.
The four Van Westendorp questions ask, in a consistent order, at what price the product would be too cheap, a bargain, getting expensive and too expensive.
Researchers still debate which order works best, so pick one and keep it. Each question must reference the same product description and the same billing unit, which the next section specifies, so the respondent reasons along one scale.
A single direct question about willingness to pay invites lowballing, where buyers deliberately state a lower price than they would actually pay.
Four anchored questions make that harder: understating one answer creates an inconsistency with the others, and the analyst can check whether each respondent's prices rise in order.
Standard wording: "At what price would this product be so cheap that you would question its quality?"
This captures the floor of credibility. Below it, buyers assume something is wrong rather than celebrate a deal.
Standard wording: "At what price would you consider this product a bargain, a great buy for the money?"
This captures the lower end of comfortable acceptance, where value clearly exceeds cost in the respondent's mind.
Standard wording: "At what price would you consider this product starting to get expensive, so that it is not out of the question but you would have to think about buying it?"
This marks where hesitation begins while purchase remains plausible.
Standard wording: "At what price would this product be so expensive that you would not consider buying it?"
This captures the ceiling beyond which the offer is rejected outright.
For a monthly consumer subscription, the first question reads "At what monthly price would this plan be so cheap you would question its quality?" and the remaining three follow the same frame, each repeating "monthly" and the plan description.
Copy all four into the survey draft and check that the product description and unit are identical across them before anyone else reviews the wording.
Write the one-paragraph concept description with its billing unit, the screener, the response format and the exclusion rule before the survey goes live. These four decisions determine whether the curves are interpretable, and none of them can be repaired after fielding.
State the billing unit in the description and repeat it in every question: per month, per seat, per unit of usage, per box, one-off.
If one respondent prices a year while another prices a month, their answers land on the same axis and the curves become noise.
A brand testing a new coffee pod range writes "price per box of 20 pods" into all four questions.
Show the same concept stimulus the launch creative will use, because a price judged against a vague description will not transfer to the offer buyers actually see; our creative team typically shares the launch concept with the research panel for exactly this reason.
Recruit respondents who match the launch target and know the category well enough to judge value, using the same screening discipline as any voice of customer research.
The coffee pod study screens for people who bought pod coffee in the past three months; a general consumer panel would price the box against instant coffee or nothing at all.
Typed prices feel realistic, but recent evidence favors structured formats.
A 2024 study by Schmidt and colleagues in the International Journal of Research in Marketing found that open-ended questions estimate willingness to pay less accurately than closed-ended questions for new products.
However, respondents' knowledge and cognitive ability can influence which format performs better.
That study tested general willingness-to-pay questions; it did not test the meter itself. A practical compromise is a fixed list of price points spanning well beyond the plausible range, which also removes typing errors and currency confusion.
Set the rule before analysis starts. Drop any respondent whose four prices are not monotonic, for example a "too expensive" price below their "bargain" price, and define a threshold for extreme outliers such as prices several multiples above the next highest answer.
The lowballing risk described earlier is why these consistency checks matter; an inconsistent sequence usually signals inattention or gaming. Document the sample size, screening criteria and exclusion rules in the report so the acceptable range can be defended when finance or sales challenges it.
A Van Westendorp graph plots four cumulative curves, and their four crossing points give you the acceptable price range (from marginal cheapness to marginal expensiveness), the optimal price point and the indifference price point.
Record which convention produced them and take the whole range into the pricing discussion. The chart is where reporting errors tend to creep in, so the steps below are deliberately mechanical.
For each question, calculate the cumulative percentage of respondents at every price on the axis.
The "too cheap" and "bargain" curves accumulate downward in price (the share who think a price is too cheap rises as price falls), while the "getting expensive" and "too expensive" curves accumulate upward.
Use the same price axis for all four curves and the same cleaned sample. Apply any weights first, then compute the cumulative percentages.
The optimal price point sits where the "too cheap" and "too expensive" curves cross, which is the price at which the fewest respondents reject the product for being either too cheap or too costly.
It reflects alignment of perceived value, not the profit-maximizing price, and it is routinely over-read as a recommendation.
The indifference price point sits where the "bargain" and "getting expensive" curves cross, so equal shares call the product a bargain and a stretch.
It often lands close to what the market already expects to pay, which makes it a useful check against competitor list prices.
The point of marginal cheapness is where "too cheap" crosses "getting expensive"; the point of marginal expensiveness is where "too expensive" crosses "bargain".
The acceptable price range runs between them, and because the meter measures where a sense of value changes, prices outside it carry rising rejection on one side or suspicion on the other.
The table below uses the convention applied throughout this guide.
| Price point | Intersecting curves | What it signals | How to use it in a launch decision |
|---|---|---|---|
| Point of marginal cheapness | Too cheap and getting expensive | Below this, quality doubts outweigh appeal | Lower bound of the band |
| Optimal price point | Too cheap and too expensive | Fewest respondents reject the price | Perception anchor, not the answer |
| Indifference price point | Bargain and getting expensive | Price the market treats as normal | Check against competitor list prices |
| Point of marginal expensiveness | Too expensive and bargain | Above this, rejection accelerates | Upper bound of the band |
Analysts disagree on whether the "bargain" and "getting expensive" curves are inverted before intersecting; the Penn State course notes on the meter warn that published plots vary in which questions are flipped.
Under one convention, the "bargain" curve is plotted as "not a bargain" and the "expensive" curve as "not expensive"; under the other, the raw curves are used, and the labeled bounds move as a result.
Neither is wrong, but mixing them across waves or between the analyst and the slide deck produces mismatched bounds nobody can reconcile. State the convention in the report.
A hypothetical illustration: a cleaned sample yields a marginal cheapness point of 18, an optimal point near 24 and a marginal expensiveness point of 32 per month, units illustrative.
The team treats 24 as a perception anchor and 18 to 32 as the band for revenue modeling.
The width of the band is information too. A narrow band means buyers broadly agree on what the offer is worth, so there is little room to move. A wide band usually signals a mixed audience or a vague product description, so split the sample by segment or tighten the stimulus before acting on it.
The Newton-Miller-Smith extension adds two purchase-likelihood questions to the standard four, so the acceptable range can be turned into rough demand and revenue curves.
Add it when the stakeholder question is where in the range to land, and agree the probability weights with finance before fielding. The extension gives the band a revenue shape, but every step after the survey rests on assumptions you should make visible.
Newton, Miller and Smith proposed in 1993 a two-question purchase-intent extension to the meter so results could be tied to likelihood of choice.
After the standard four questions, each respondent rates how likely they would be to buy at their own "bargain" price and at their own "getting expensive" price, usually on a five-point scale.
The intent ratings are converted to purchase probabilities with researcher-set weights, interpolated across each respondent's range, then aggregated and multiplied by price to sketch demand and revenue curves.
The same survey answers under two calibrations: the revenue peak moves from 29 to 26 when the weights change (illustrative data).
A common starting calibration, used in the Penn State course notes, counts 70% of "very likely" and 50% of "likely" answers as buyers. With weights like these, revenue often peaks nearer the expensive end of the band than the optimal perception point.
The weights are assumptions, so present the revenue curve as a scenario and show how the peak shifts under an alternative calibration. If the peak moves materially between reasonable weightings, the extension is telling you the data cannot settle the position on its own.
The estimate remains perception-based, and the revenue curve inherits every limitation of the band it was built on, including the missing competitive context. It belongs beside the margin and competitor inputs covered later.
The meter has five main limits: stated prices overstate real spending, it has no competitive context, it fits complex bundles poorly, it misreads luxury goods, and its range goes stale as markets move.
Attach a limitations note to every pricing report that names which decisions the range supports and which it cannot. The meter's flaws are well known among researchers and rarely communicated to the executives who read the chart.
What consumers say they would pay can exceed what they actually spend.
A meta-analysis of 77 studies by Schmidt and Bijmolt, published in the Journal of the Academy of Marketing Science in 2020, found that hypothetical willingness to pay for consumer goods averaged about 21% higher than real willingness to pay.
This finding comes from broader willingness-to-pay research rather than Van Westendorp studies specifically. It highlights the need to validate survey findings through purchase experiments, preorders, or other behavioral evidence.
The standard meter never asks respondents to choose between your product and a rival at given prices. It therefore misses how competitor pricing shapes acceptance, which for most launches is the dominant force on the day.
The crossing lines rest on a psychological model of how price perception shifts, not on an economic theory of demand. Treat the outputs as directional, useful for bounding a discussion and sharpening the next test rather than proving a price.
Single, clearly described offers work best. Feature-laden tiers, seat-based enterprise agreements and negotiated contracts ask respondents to price something they cannot hold steady in mind, so conjoint or sales-side evidence from recent deals will serve you better.
The band is a snapshot. Competitor launches, category inflation and shifts in what buyers compare you against move the acceptable range, so plan periodic re-tests using the same billing unit, the same stimulus and the same curve convention so the waves are comparable.
Watch: Weaknesses of the Van Westendorp Price Sensitivity Meter
Build a one-page overlay that shows the band, the margin floor, competitor prices and the intent or in-market signal, then record the launch price and its follow-up test.
Exclude any part of the band below the margin floor before discussion starts. Pricing low to win volume rarely pays back: the same McKinsey analysis found that volume would have to rise 18.7% just to offset the profit impact of a 5% price cut.
The pricing model itself, whether subscription, usage or seat, is a strategic choice settled before research, and it has consequences beyond the launch price.
Bessemer Venture Partners' State of the Cloud 2021 found usage-based companies averaged roughly ten percentage points higher net dollar retention than subscription peers, a historical figure that moves with SaaS market conditions.
Place competitor prices on the same axis as the band. Because respondents never chose among rivals, this overlay supplies the context the meter omits; the same competitive intelligence work that tracks rival positioning should feed it.
Use Newton-Miller-Smith intent evidence or an in-market offer test to pick a position within the surviving band, and avoid defaulting to the optimal perception point. Our measurement team favors a live offer test where traffic permits, since it is the only input that observes payment.
Set the launch price inside the surviving band, write down the evidence it rests on and schedule the test that will confirm or move it.
Continuing the hypothetical band of 18 to 32 per month, the lower third falls below the margin floor, competitors cluster near 25 and intent weighting favors the upper half, so the team launches at 29 with an offer test planned at 27 and 32.
| Evidence source | Question it answers | What it misses | How it narrows the band |
|---|---|---|---|
| Acceptable price range | Where price stops feeling credible | Competition, actual purchase | Sets the outer bounds |
| Unit economics | What price covers cost and target margin | Buyer perception | Removes the unprofitable lower end |
| Competitor benchmarks | What buyers compare against | Your differentiation | Flags positions that invite comparison |
| Purchase intent extension | Where stated likelihood peaks | Real payment, rivals | Points to a region within the band |
| In-market offer test | What converts at a given price | Long-term retention | Confirms or moves the launch price |
Pricing research is most useful when its findings inform decisions beyond the price itself. An acceptable price range can guide brand positioning, shape how creative communicates value, and influence where media investment goes.
At Fieldtrip, we connect market research and competitive insights with brand strategy, creative development, and media planning. Our specialist teams work together to help brands translate customer insights into consistent messaging, campaigns, and measurable performance.
A Van Westendorp study can provide a starting point, but launch decisions also need competitive context, clear positioning, and real-world validation.
Planning a product launch or refining your market positioning? Reach out to Fieldtrip to explore how our research, strategy, creative, and media expertise can support your go-to-market decisions.
A Van Westendorp survey commonly uses a few hundred qualified respondents, but the appropriate sample size depends on the target population and required precision. If you plan to compare customer segments, recruit enough respondents within each group to produce meaningful price sensitivity curves.
Yes. Van Westendorp results can be calculated separately for customer segments, such as new and existing customers or different income groups. Comparing their acceptable price ranges can reveal pricing differences that disappear when all responses are combined into a single average.
Yes, if they are potential buyers of the new tier. However, their existing subscription prices may influence their responses. Analyze existing customers separately from qualified prospects to identify differences in price expectations and potential upgrade resistance.
Include the brand name if customers will see it when making a purchase. Brand familiarity and reputation can influence price perceptions. For an unfamiliar brand, researchers may also test a blinded concept separately to understand how brand recognition affects the acceptable price range.
Yes, but each market needs a sufficiently representative sample and a locally appropriate survey. Currency conversion alone cannot account for differences in purchasing power, taxes, and consumer expectations. Analyze each country's responses separately before comparing the acceptable price ranges.
