Gabor-Granger vs Van Westendorp: which pricing method should you use?
How the two survey pricing methods differ, a worked example of each that runs one into the other, and the limits they share.
Van Westendorp asks people to name four price thresholds and gives a range of acceptable prices. Gabor-Granger asks whether they would buy at prices you set and gives a demand curve and a revenue-maximising price. Use Van Westendorp when you don't yet know the plausible range, then Gabor-Granger to choose a price inside it. Neither accounts for competitors.
How do the two methods differ?
Van Westendorp lets respondents name the prices; Gabor-Granger shows them prices you chose and records yes or no.
| Van Westendorp | Gabor-Granger | |
|---|---|---|
| What people are asked | Four open questions: at what price is it too cheap, a bargain, getting expensive, too expensive? | "Would you buy [product] at [price]?" at several prices |
| Who sets the prices | The respondent | The researcher |
| What you get | Four price points and an acceptable price range | A demand curve, revenue per respondent at each price and, with a unit cost, profit |
| Answers "what range is plausible?" | Yes | No, only within the prices you test |
| Answers "which price earns most?" | No | Yes, among the prices tested |
| Main bias to manage | Anchoring by anything shown before the questions | Order and anchoring by the first price shown |
| Origin | Peter van Westendorp, ESOMAR Congress, 1976 (cited by the CRAN package) | André Gabor and Clive Granger, 1966 |
What does a Van Westendorp result look like?
The four answers per person become four cumulative curves, and the prices where they cross mark the range. These five respondents are the illustrative example from our Van Westendorp calculator, far too few for real use but enough to show the arithmetic.
| Respondent | Too cheap | Bargain | Getting expensive | Too expensive |
|---|---|---|---|---|
| A | $15 | $35 | $40 | $60 |
| B | $15 | $20 | $40 | $45 |
| C | $10 | $20 | $55 | $60 |
| D | $20 | $45 | $55 | $60 |
| E | $20 | $35 | $50 | $60 |
- Point of marginal cheapness, $27.50. At $20, two people still call it too cheap and nobody calls it "not cheap"; at $35 the gap runs two the other way. Interpolating between the two prices puts the crossing halfway: 20 + 15 × 2 ÷ (2 + 2) = 27.50.
- Optimal price point, $37.50. Between $35 and $40 nobody calls it too cheap or too expensive, so the crossing is the middle of that stretch.
- Indifference price point, $38.75: where as many call it a bargain as call it getting expensive.
- Point of marginal expensiveness, $52.50: where "too expensive" overtakes "not expensive".
The acceptable range is $27.50 to $52.50. Conventions for the marginal points differ between sources; the calculator uses the "not cheap" and "not expensive" curves, as the pricesensitivitymeter package for R does.
What does a Gabor-Granger result look like?
Multiply each price by the share who would buy at it, and the highest product is the revenue-maximising price. Continuing the example, suppose you test five prices inside the $27.50 to $52.50 range on 150 people. The shares below are invented for illustration, with a unit cost of $12.
| Price | Would buy | Revenue per respondent | Profit per respondent (cost $12) |
|---|---|---|---|
| $30 | 62% | $18.60 | $11.16 |
| $35 | 55% | $19.25 | $12.65 |
| $40 | 47% | $18.80 | $13.16 |
| $45 | 33% | $14.85 | $10.89 |
| $50 | 21% | $10.50 | $7.98 |
Revenue peaks at $35 (35 × 0.55 = 19.25) and profit at $40 ((40 − 12) × 0.47 = 13.16). When demand falls as price rises, a unit cost can only move the profit-maximising price up from the revenue-maximising one, because a higher price earns more on every sale it keeps. Note the noise: on 150 people each share carries a margin of error of about ±8 points, so $35 and $40 are effectively tied on revenue. Read the result as "somewhere from $35 to $40", then test in the market. The Gabor-Granger calculator does this arithmetic and warns when the best price sits at the edge of what you tested.
When should you use one, the other, or both?
Use them in sequence when the price is genuinely open; use one alone when you already know part of the answer.
- New product, no idea of the range: Van Westendorp first.
- A few candidate prices and a revenue or margin goal: Gabor-Granger, with your unit cost.
- Both questions, one survey: Van Westendorp first, then Gabor-Granger. An extension by Newton, Miller and Smith (1993) also lets a Van Westendorp study estimate prices that maximise trial and revenue; the pricesensitivitymeter package implements it.
- Price trades off against features or a rival's offer: neither. Use conjoint analysis, which puts price inside whole products.
What limits do both methods share?
Both record what people say, about one product, in isolation.
- Stated, not revealed. Intentions "often contain systematic biases" and may not predict purchases (Sun and Morwitz, 2010). Use the results to rank prices, not to forecast sales.
- No competitors or reference prices. Neither method shows what rivals charge.
- Anchoring. An example price in the description shifts Van Westendorp answers; the first price shown shifts Gabor-Granger answers, so vary the starting price across respondents.
- Sample. There is no standard minimum for Van Westendorp; recalculate on random halves of your data and check that the price points hold.
To run either method on your own data, our free Van Westendorp calculator and Gabor-Granger calculator work in your browser and send nothing anywhere; the survey question bank has the exact question wording for both. In PulseLake, pricing surveys run as traditional research, with the method and evidence kept in the same study context as the decision they inform.
Frequently asked questions
Can you ask both in the same survey?
Yes. Ask the four Van Westendorp questions first, because they are open-ended and any price shown earlier would anchor the answers, then the Gabor-Granger prices. Keep the product description identical for both.
Does Van Westendorp give the optimal price?
Not in the sense of revenue or profit. Its optimal price point is where as many people call the product too cheap as call it too expensive. It says nothing about how many would buy or what you would earn.
How many prices should a Gabor-Granger study test?
Enough to show where demand falls away, spread across the plausible range, with prices below and above the one you expect to win so the best price is not at an edge. The examples on this page use five prices.
Who developed the two methods?
Peter van Westendorp presented the price sensitivity meter at the ESOMAR Congress in 1976. The Gabor-Granger method is named after the economists André Gabor and Clive Granger, who studied consumers' price limits in the 1960s.
Sources
Sources for the facts on this page, last checked October 9, 2026.
- pricesensitivitymeter: Van Westendorp Price Sensitivity Meter Analysis (R package on CRAN, version 1.3.3, 2025), citing van Westendorp (1976) checked October 9, 2026
- pricesensitivitymeter documentation: the Newton, Miller and Smith (1993) extension checked October 9, 2026
- Sun and Morwitz, Stated intentions and purchase behavior: A unified model, International Journal of Research in Marketing 27 (2010) checked October 9, 2026
Related guides
- Van Westendorp price sensitivity calculator
- Gabor-Granger price calculator
- MaxDiff vs conjoint analysis: when should you use each?MaxDiff ranks a single list of items; conjoint measures how attributes and levels, often including price, trade off inside whole products. When to use each.
- Van Westendorp pricing research explained
Run research end to end. Keep the knowledge working.
One AI-native operating system for market research and insight professionals — from study design and evidence generation to agents, institutional knowledge, delivery and action.