Pricing research
What people say they would pay and what they pay are two different measurements, and the first one is still useful. Pricing research maps the range where your price is read as plausible, the point where it starts to feel expensive, and the segment where the answer flips. That is enough to publish a price list without guessing.
What a price test settles
The acceptable range
Below a floor, the offer reads as cheap or suspicious; above a ceiling, it stops being considered at all. The useful answer is the band between the two, not a single number.
The shape of the curve
Demand rarely falls in a straight line. A price grid shows where a small increase costs almost nothing and where it costs a third of the audience, which is where the decision actually sits.
Which segment moves
A single average hides the only interesting part. The same price is often obvious for one segment and a blocker for another, and the report should show both.
Three ways to ask, and what each one gives
| Method | What you ask | What you read |
|---|---|---|
| Van Westendorp | Four open price questions: too cheap, a bargain, getting expensive, too expensive. | The acceptable band and the indifference point, where the same share calls the price a bargain and calls it expensive. |
| Gabor-Granger | The same offer at several prices, one per respondent, in random order. | Purchase intent by price, and the price point where the curve breaks. |
| Choice pairs | Two offers at two prices; the respondent picks one. | Which attribute carries the decision when price competes with features. |
Van Westendorp is the cheapest place to start, and it is the one that survives a short survey. The grid is better once you have a range and want to see the shape inside it. Choice pairs are for when the price is not the only thing that moves.
The four questions, written out
- At what price would this be so cheap that you would doubt the quality?
- At what price would it be a bargain, a good deal for what it is?
- At what price would it start to feel expensive, but still be worth considering?
- At what price would it be too expensive to consider at all?
They are asked as open numbers, not as a scale. A scale measures agreement with a price you suggested; these four measure where the respondent places the offer on their own map.
How many responses
A range needs around 100 responses to be stable, and every segment you want to read separately needs its own 25 to 50. Below that, the curve still shows a direction, but the crossing points move on the next run.
The sample size calculator gives the number for a given margin. What a run costs is on the cost page.
What it does not do
Stated willingness to pay is not revenue. This kind of test does not forecast how many people will buy, it does not simulate a competitor's response, and it does not price a product whose value depends on usage you cannot describe in a paragraph. On a synthetic population, it gives the reading of a price by an audience calibrated on public statistics, not a margin of error as if it were a probability sample. The reliability page is explicit about that boundary.
The honest use is comparative: two candidate prices, two packagings, two segments, read in the same run. That is where the method earns its keep.
Test the price you are about to publish
Put the offer in front of the audience, ask the four questions, and read the band before the average. It takes minutes, and the free plan covers a first range.
Related: concept testing, message testing, and the plan comparison.