To figure out how to price a product, estimate what your intended buyers will pay, then check that estimate against real purchase behavior and your business economics. Willingness to pay (WTP) is not necessarily a fixed number customers already know: answers depend on how you describe the offer, what alternatives people consider, how familiar they are with it, and whether their choice has real consequences.
What a willingness-to-pay study can—and cannot—tell you
A WTP study gives you evidence for a pricing decision, not a guaranteed price. Someone answering “What’s the most you would pay for this?” may respond differently from someone choosing between your product and competitors, or deciding whether to make a consequential purchase.
As an Amazon Associate I earn from qualifying purchases.
That distinction matters because hypothetical answers can diverge from buying behavior. A 2020 meta-analysis by Jonas Schmidt and Tammo Bijmolt covered 77 studies in 47 papers and 115 effect sizes; it found average hypothetical bias of 21% across the included studies. That average is not a correction factor for your survey, product, or market. The authors also caution against assuming that indirect methods automatically outperform direct questions.
Use survey results to narrow or challenge a pricing decision. When feasible, compare them with actual purchases, a live market test, or other revealed-preference data. A study can also help determine whether to launch an offer, which features to include, or how to position it—not just what price to print.
#1 Best Overall
Choose a method that matches the buying decision
The four methods below differ in what people evaluate, how directly they consider price, and how much design and analysis they require. The Product Hunt guide to WTP, written by Kristen Berman and published December 6, 2024 (originally published in February 2024), discusses these approaches and favors comparative designs for some unfamiliar or high-ticket offers. Those are recommendations, not universal rules.
| Method | What the respondent does | When it may fit | Main cautions |
|---|---|---|---|
| Van Westendorp Price Sensitivity Meter | Gives four prices: suspiciously cheap, a bargain, expensive, and too expensive. | A relatively quick, open-ended way to explore perceived price boundaries, particularly when respondents know the product category. | People imagine a purchase rather than necessarily make one; the output is a survey signal, not a validated selling price. |
| Becker-DeGroot-Marschak (BDM) | States a maximum price; a price is drawn at random, and the respondent buys at that drawn price if their stated maximum meets or exceeds it. | When you can implement the mechanism with real purchase consequences and participants understand its rules. | The mechanism can confuse respondents. Incentive alignment depends on a correctly implemented, understood process. |
| Multiple price list (MPL), also called Gabor-Granger in the guide | Answers yes or no to a sequence of prices; in some designs, one response is randomly selected for implementation. | When you want a structured set of price responses for a clearly described offer. | Price intervals, order, framing, and inconsistent answers can affect results. |
| Discrete choice or choice-based conjoint | Chooses among competing offers that vary in price and features, potentially including a no-purchase option. | When buyers compare alternatives or must trade features against price, including for unfamiliar or more complex offers. | Requires realistic offer bundles and more demanding design and analysis than a short direct question. |
Use open-ended questions when the offer is easy to imagine
Van Westendorp asks respondents to identify four thresholds: the point where the product seems so cheap that quality is suspect, where it seems like a bargain, where it starts to seem expensive, and where it is too expensive. The resulting range can help you see whether the price you are considering falls outside buyers’ stated comfort zone. It does not show what they will actually buy.
Use a consequential mechanism only if people can follow it
In BDM, a respondent reports the most they would pay and a random price is drawn. If their reported maximum is at least that drawn price, they buy at the drawn price; otherwise, no purchase occurs. The design aims to make truthful reporting advantageous, but the logic is less intuitive than a simple price question. Check comprehension before relying on the answers, and do not describe a hypothetical survey as incentive-aligned if no purchase consequence is actually attached.
Read price-list responses as a pattern, not a string of perfect answers
In an MPL, respondents say yes or no at multiple prices. Responses can reveal how acceptance changes as price rises, but the answer sequence may be affected by the order and spacing of prices. Jack, McDermott, and Sautmann’s 2022 NBER working paper notes that multiple switches and never switching can signal high error rates. Their proposed approach uses randomization and a random-effects latent-utility model to detect bias and account for error; they also report order effects in South Africa data.
Use choice designs to expose trade-offs
Discrete choice presents competing product profiles with varied features and prices and asks which the respondent is most inclined to purchase. This makes the comparison explicit and can show which features buyers trade against price. It is only useful if the profiles resemble plausible offers and the analysis fits the design; an assortment of arbitrary feature combinations can produce misleading results.
Evidence does not establish that indirect methods are always more accurate. Miller and co-authors’ 2011 comparison of open-ended questions, choice-based conjoint, BDM, and incentive-aligned conjoint against real purchase data reported greater price sensitivity in incentive-aligned settings. That finding does not make one method the right choice for every product.
Design the study around the actual offer
- Specify the decision. Decide whether you need to choose a launch price, compare price points, test a feature bundle, or assess whether the product should launch at all. Keep the study focused on that decision.
- Write the offer as customers would see it. Show the relevant features, limitations, positioning, and price context. Copy and framing can affect perceived value, so consider testing distinct, realistic descriptions when positioning is still in question.
- Recruit likely buyers. Use screening questions when needed so respondents resemble the people who might actually buy. A general audience is not a substitute for the intended market.
- Choose a response format that fits the offer. A familiar, frequently bought product may be easier to value with a direct price question. A less familiar product, a high-ticket decision, or an offer whose features need to be weighed against alternatives may call for a comparative design. Treat this as a design choice to validate, not a rule.
- Allow people to decline. Include a “would not purchase” option in choice questions so respondents are not forced to select a product. For mechanisms involving a real purchase, explain the rules and check that participants understand them.
- Inspect response quality. Look for comprehension problems, implausible patterns, and—when using price lists—multiple switches, no switches, or order effects. Do not treat every recorded answer as equally informative.
- Compare the estimate with market evidence. If feasible, test actual purchase behavior or run a live market test. A survey can guide what to test; it cannot substitute for observed demand.
There is no universally established sample size in the evidence summarized here. Set one based on your research design, target population, analysis, and the uncertainty you can tolerate rather than treating a single number as a guarantee.
What’s actually slowing this PC down?
Pick the symptom - the matching free tool is one click away.
Turn responses into a pricing decision
Start by examining the shape of responses, not just a single average or cutoff. For a direct price question, check whether answers cluster or vary widely. For Van Westendorp, compare the four thresholds and see how the candidate price sits within the stated range. For MPL, inspect acceptance across price points and assess whether switching patterns are coherent. For discrete choice, interpret prices alongside the features and alternatives respondents saw.
Then make the commercial decision separately from the measurement. A price buyers say they would accept is not automatically a viable price for your business. Compare candidate prices with costs, margins, channel economics, and the product’s role in your wider offer. If plausible price points imply very different outcomes, run a targeted follow-up or market test rather than pretending the survey has resolved the uncertainty.
Keep the conditions attached to any reported result: who was asked, what offer they saw, how the question was framed, whether decisions had consequences, and when the study was run. If a study changes the description, audience, or method, treat its estimates as answers to a different question unless the designs support a valid comparison.
Quick Recap
What the evidence does not establish
- No method returns a universal “correct” price, and the evidence here does not establish one best method for every product.
- The 21% average hypothetical bias reported by Schmidt and Bijmolt is an average across their meta-analysis, not a discount to apply to an individual WTP estimate.
- Indirect questioning is not automatically more accurate than a direct question; method performance depends on the study and its implementation.
- A result from one audience, offer description, or market should not be generalized to a different one without evidence.
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
