Van Westendorp sample size calculator (free, no sign-up)
How many participants does a Van Westendorp survey need? Van Westendorp, also called the Price Sensitivity Meter, is the four-question pricing survey that finds the range of prices your market accepts, first published by Peter van Westendorp in 1976 (ESOMAR). Van Westendorp sample size depends on three things: how precise the result needs to be, the smallest customer segment you want to look at on its own, and how many responses you expect to throw away. The calculator below takes those three inputs and tells you how many valid participants you need and how many responses to collect.
Research Toolkit
Van Westendorp Sample Size Calculator
Estimate how many respondents your price-sensitivity study needs. Adjust the inputs and the numbers update instantly.
Distinct groups you'll read separately (e.g. plans, personas). 1 if you're not splitting.
Sets the per-segment floor: 50 / 100 / 200 completed responses.
Answers dropped for contradictions or speeding. Typically 10 to 20%.
Share of people screened who qualify. Lower incidence means more invites to reach your target.
Responses to collect
118
allowing for ~15% unusable answers
Completed responses needed
100
100 per segment × 1 segment
People to invite
—
This is a rule-of-thumb estimate, not a statistical guarantee. Van Westendorp has no exact sample-size formula; the numbers apply a per-group floor (50 / 100 / 200) from practitioner guidance, grossed up for the responses you'll discard.
Why these numbers?
Van Westendorp reads price thresholds off cumulative curves, so each segment needs enough responses to draw a stable line of its own: about 100 as a working floor, or 50 for a rough read and 200 when the number has to hold up. That's why segments multiply the target instead of sharing it. See the sample size guide and the segmentation piece for how that plays out.
Built by OpinionX. It runs Van Westendorp studies, survey through to results.
This guide covers what each input does, why Van Westendorp needs this many participants (using resampled answers from an actual survey), and where the calculator's results stop being reliable.
What drives Van Westendorp sample size?
Van Westendorp sample size comes down to the three inputs in the calculator, and each one multiplies the one before it.
1. How precise the result needs to be. Around 100 valid participants shows you roughly where the optimal price sits, which is enough to rule prices in or out early on. Around 300 valid participants gives you an optimal price you can make a pricing decision on. The pricing consultancy 5 Circles Research suggests at least 400 participants for a consumer product and 200 for a business product, and the calculator's Strict setting uses that 400.
2. The smallest customer segment you'll look at on its own. A customer segment is a group of participants you filter the results down to, such as people on your free plan or people in one country. Each segment needs as many participants as input 1 says, because a Van Westendorp chart for 40 people is as unstable as a survey of 40 people. So if your smallest segment makes up 25% of participants and you want 300 in it, you need 300 ÷ 0.25 = 1,200 valid participants in total. Our guide on how to segment Van Westendorp results explains which segments are worth comparing.
3. Responses you'll throw away. Some people won't finish, some give their four prices out of order, and a few type joke answers. OpinionX ran a practice Van Westendorp survey about a 30oz Stanley Cup, built from plain number questions, and of the 147 people who started, we had to remove 13: 4 didn't answer all four questions and 9 gave prices out of order. That's 9%, which is why the calculator starts at 10%. A question type that stops out-of-order answers, like the Van Westendorp question type on OpinionX, brings that share down.
Why does Van Westendorp need this many participants?
Van Westendorp needs more participants than many surveys because its key prices come from where two lines cross, and each line is only a share of participants. The optimal price point, for example, is where the share of people calling a price "too cheap" equals the share calling it "too expensive". Near that price, both shares are small (14% and 13% of participants in the Stanley Cup survey described below), so a handful of people can move where the lines cross. People also tend to type round prices like $20 or $25, so the lines move in steps, and a few answers can push a crossing from one round price to the next.
To see how much difference sample size makes, we took the 134 valid answers from the Stanley Cup survey, drew 2,000 random samples of each size from them, and found the optimal price point and the edges of the acceptable price range in every sample. The full survey puts the optimal price at $20 and the acceptable price range at $15 to $25. Here's the range that 90% of the samples fell into at each size:
| Sample size | Optimal price point (full survey: $20) | Bottom of acceptable range (full survey: $15) | Top of acceptable range (full survey: $25) |
|---|---|---|---|
| 25 participants | $12 to $30 | $10 to $20 | $20 to $30 |
| 50 participants | $15 to $25 | $12 to $20 | $20 to $30 |
| 100 participants | $15 to $25 | $13 to $19 | $25 to $30 |
| 200 participants | $18 to $25 | $14 to $18 | $25 to $30 |
| 300 participants | $20 to $23 | $15 to $18 | $25 to $28 |
| 400 participants | $20 to $23 | $15 to $17 | $25 to $27 |
With 25 participants, the optimal price point could land anywhere from $12 to $30, which is too wide to choose a price from, while 100 participants narrow it to $15 to $25, enough to see roughly where the price sits. By 300, it stays between $20 and $23, which is precise enough to make a decision on. The edges of the acceptable price range settle more slowly, because they depend on the "getting expensive" and "bargain" lines too.
Don't be fooled though: the ranges keep narrowing past 134 participants because each sample comes from the same 134 people. An actual survey of 400 people would also include people with opinions the first 134 didn't have, so treat the table as a guide to how fast the results settle, not as an exact promise.
What are the limitations of this calculator?
The calculator's precision settings are rules of thumb, not a statistical formula. Van Westendorp has no standard formula for sample size, because its outputs are crossing points rather than a single average, so the settings come from published guidance and from how the Stanley Cup results settled. A product whose buyers disagree a lot about price needs more participants than one whose buyers mostly agree.
The calculator also assumes each segment needs as many participants as the whole survey. If you only want to check whether two segments differ, rather than pick a price for each one, the Rough setting per segment is enough to spot a large difference. One OpinionX survey found a $26 optimal price for people who cycle to work and a $15 optimal price for people who take public transport, a gap that 100 participants per group would show.
Finally, the Newton-Miller-Smith extension, which adds two "how likely are you to buy?" questions to estimate revenue, builds its revenue peak from two answers per person, an approach first published in 1993 (method notes). Use at least the Confident setting if you add it. Our guide to the Newton-Miller-Smith extension shows how it calculates the revenue estimate.
Want to see where these participants' answers end up? Our guide How to read a Van Westendorp chart walks through each line and crossing point with the full Stanley Cup data. No list of customers to survey? Our guide on how to find free participants for your survey covers ways to reach people without a paid recruitment panel. And for the full method, from writing the four questions to choosing your price settings, read our Van Westendorp pricing guide.
Scoping your first Van Westendorp study and want a research expert to check your segments, sample size and recruitment plan before you send it out? Book a free 30-minute setup call, with no obligation.
The Van Westendorp question type is on OpinionX's free plan with up to 25 participants per survey, which is enough to test your survey before you run it, and the Analyze plan, at $900 a year, removes the participant limit.
