Van Westendorp price sensitivity meter: definition, formula and examples

Van Westendorp's price sensitivity meter turns four survey questions into an acceptable price range. The questions, the math, a worked example and its limits.

Definition

The Van Westendorp price sensitivity meter is a survey method that asks people at what prices a product would feel too cheap, cheap, expensive and too expensive. Plotting the answers as cumulative curves gives an acceptable price range, bounded by where the curves cross, before you have any sales data.

  • The lower bound is where 'too cheap' crosses 'expensive'; the upper bound is where 'cheap' crosses 'too expensive'.
  • It records what people say, so treat the range as the bounds of a live price test.

Van Westendorp asks potential customers for four prices instead of one. Someone evaluating a scheduling tool might say $10 is so cheap they'd doubt it works, $20 is a bargain, $35 is getting expensive and $50 is too much. Collect those four numbers from a few hundred people and you can see where most of them start to balk.

Dutch economist Peter van Westendorp introduced the method in 1976 (overview). The standard questions:

  1. At what price would this be so cheap that you'd question its quality? (too cheap)
  2. At what price would it be a bargain? (cheap)
  3. At what price would it start to feel expensive, though you'd still consider it? (expensive)
  4. At what price would it be too expensive to consider? (too expensive)

How to calculate it

For every price, compute the share of respondents whose answers put that price in each group, then find where the curves cross:

Text
too cheap(p)     = share whose "too cheap" answer ≥ p
cheap(p)         = share whose "cheap" answer ≥ p
expensive(p)     = share whose "expensive" answer ≤ p
too expensive(p) = share whose "too expensive" answer ≤ p
lower bound = price where too cheap(p) = expensive(p)
upper bound = price where cheap(p) = too expensive(p)

Example survey results:

PriceToo cheapCheapExpensiveToo expensive
$1544%82%6%1%
$2018%66%12%2%
$256%46%22%4%
$302%27%36%7%
$351%14%54%14%
$400%6%70%24%

"Too cheap" drops below "expensive" between $20 and $25, at about $21, and "cheap" meets "too expensive" at about $35. The acceptable range is roughly $21 to $35. Two more crossings are often reported: too cheap against too expensive, called the optimal price point (about $26 here), and cheap against expensive, the indifference price point (about $29).

Example data: the shaded band between the two crossings is the acceptable range, $21 to $35.

Why it matters for pricing

Before launch you have no conversion data, and a survey is a cheap way to avoid a price that is badly off. It is quick to run, and our Van Westendorp calculator draws the curves from your answers.

Its limit is that people describe prices without paying them, and nothing in the method says which price inside the range earns the most. Use the range as the floor and ceiling of a live test instead of picking a point from the chart. With Leverprice you set a minimum and a maximum price, and real purchases decide the number in between.

Common mistakes

  • Surveying people who would never buy. Friends, followers and other founders give cheerful numbers, so ask people who have the problem and a budget.
  • Asking before explaining the product. Respondents need to know what they'd get, or they price the category instead of your product.
  • Treating the optimal price point as the answer. It marks where the two extreme groups balance, which says nothing about revenue.
  • Reading curves from a few dozen answers, where a single respondent can shift a crossing by a few dollars.

Van Westendorp is one way to estimate willingness to pay. It can't measure price elasticity, which needs real purchases, and its answers can be skewed by price anchoring. A price test turns the range it produces into a measured result.

Frequently asked questions

How many respondents does a Van Westendorp survey need?

Enough that no single answer moves the curves much. With 40 respondents each person is worth 2.5 percentage points, with 200 each is half a point, so a few hundred answers from likely buyers give stable crossings.

Can Van Westendorp predict revenue?

Not by itself, because it never asks whether someone would buy. Extensions such as the Newton/Miller/Smith variant add purchase-likelihood questions to estimate demand (overview).