Willingness to pay describes the maximum amount a customer would be prepared to pay for a particular offer under particular conditions. In practice, research usually estimates this threshold rather than observing it exactly. To estimate it, define the buyers you want to understand, show them a clear offer, collect responses using a suitable pricing method, and check the findings against actual purchases.
The useful result is usually a picture of how responses differ across prices and customer groups. A single average can hide shoppers who would never buy the product and others who would accept a higher price.
Key Takeaways
- Define the audience, product and purchase conditions before asking about price.
- Price acceptability, stated purchase intent and completed purchases measure different things.
- Keep uncertain answers and sample sizes visible when interpreting results.
- Use survey findings to design a purchase test before treating them as evidence for a selling price.
What willingness to pay tells you
Willingness to pay belongs to a buyer and an offer, rather than to a SKU alone. The same kettle with slower delivery, an unfamiliar seller or a different warranty is a different buying proposition.
It also differs from the price a customer has already paid. A purchase at $69 shows that the buyer accepted that amount on that occasion. It does not reveal whether they would have accepted $79. Equally, an abandoned checkout does not prove that the buyer’s limit was below $69: they may have encountered a payment problem or changed their plans.
Keep the research task distinct from the final decision. Measuring willingness to pay establishes evidence about buyers; choosing a price point also requires a business objective and workable economics.
Define whose willingness to pay you are measuring
Recruit people who could realistically make the purchase you are studying. “Everyone interested in coffee” is a weak audience definition for a temperature-controlled kettle. Someone who enjoys coffee but never prepares it at home may like the concept without having a reason to buy it.
For a hypothetical kettle study, you might recruit US adults who brew pour-over coffee at home and expect to buy or replace a kettle within the next 90 days. That definition makes the findings relevant to near-term category shoppers. It also limits them: the results would not describe all US adults or everyone visiting your store.
Screen for category use, buying responsibility and purchase timing before showing prices. Then record distinctions that might explain responses, such as first-time purchase versus replacement, current equipment and previous category spending.
If you recruit only existing customers, label the study accordingly. Their experience with your store may differ from that of new shoppers. Where both groups matter, recruit both and retain the group labels in the results.
There is no sample size that makes every study reliable. Plan around the groups and comparisons you need. A sample that supports an overall view may become too small after splitting it by customer type and purchase occasion. Report the number of usable responses in each group, and avoid treating a few enthusiastic respondents as a market segment.
Describe the complete offer before showing a price
Write one offer description and use it consistently. Respondents should understand what they receive, who sells it and what they must pay.
For our hypothetical study, the offer could specify:
- A 0.8-liter stainless steel electric gooseneck kettle.
- Adjustable temperature and a 30-minute keep-warm function.
- A one-year warranty and a 30-day return window.
- Delivery within three to five business days.
- A one-time total in US dollars, including delivery and applicable tax.
Use the same product image and seller information throughout the study. Avoid adding “premium,” a sale badge or a crossed-out comparison price unless those elements are deliberately part of the offer being tested.
Be explicit about the price basis. A shopper answering about a $69 item plus an unspecified delivery charge is evaluating something different from a $69 delivered purchase. The distinction between list price and net price matters when discounts or fees change the amount the buyer actually pays.
Pilot the description with a few eligible people. Ask them to explain the offer in their own words. Correct misunderstandings before collecting more responses.
Match the research method to the question
Choose the method according to what remains unknown. A study of acceptable price ranges needs different questions from a study of whether shoppers would trade temperature control for a lower price.
Van Westendorp: perceived price boundaries
Van Westendorp asks for four amounts: a price so low that quality becomes doubtful, a bargain price, an expensive but still acceptable price, and a price too high to consider. The combined responses describe perceived price boundaries. They do not establish how many units a store would sell. Any “optimal” intersection is an output of that analysis, rather than proof of the best commercial price.
Gabor-Granger: purchase intent at specified prices
Gabor-Granger asks whether someone would buy a defined offer at a stated price, then tests higher or lower prices according to the response. It helps locate stated buying thresholds within the tested range. Varying the starting price can reduce reliance on one initial anchor, but repeated questions can still encourage respondents to recognize and react to the pricing exercise.
Monadic testing: one price per respondent
Monadic testing assigns comparable respondents to groups, each seeing the same offer at one price. It avoids showing each person a sequence of prices to compare. Random assignment helps make the groups comparable, but each price needs enough responses. The outcome remains stated intent unless respondents are making real purchases.
Choice-based conjoint: price and feature trade-offs
Choice-based conjoint presents sets of offers whose prices and attributes vary. Respondents choose among them; a “none” option allows them to decline all offers. Use it when features, brands or competing alternatives are central to the question. Its estimates depend on the design and model, and modeled preference shares are not guaranteed sales shares.
For the fixed kettle offer, a Gabor-Granger study can explore stated thresholds. A monadic study can check reactions without exposing respondents to a price sequence. If you are still deciding which features to include, conjoint addresses that different question.
Ask about a concrete buying decision
A useful research question specifies an action, an offer, an amount and a time frame. Asking “What is a fair price?” leaves room for respondents to answer about fairness without considering a purchase.
For the hypothetical kettle, the question could be:
Imagine you are choosing a kettle for your own use within the next 90 days. Based on the product and seller information shown, would you buy this kettle for a total of $69, including delivery and tax, using your own money? You may choose another kettle or buy nothing.
Offer “Yes,” “No” and “Not sure” responses. Keep uncertainty separate in the data; do not silently count it as willingness to buy. If you use a purchase-likelihood scale instead, decide before analysis which response categories count as positive and retain the complete distribution.
For a Gabor-Granger sequence, predefine the tested prices, starting-price assignment and branching rules. Keep the offer unchanged as the price changes. Include a rule for uncertain responses rather than forcing them into a buying threshold.
Ask a short follow-up after the pricing exercise: “What most influenced your answer?” Someone rejecting the kettle because of its capacity has given different information from someone rejecting only the price.
When alternatives matter, establish a consistent competitive context before the price questions. An isolated offer does not test whether shoppers would choose it over available substitutes. Pricing intelligence for ecommerce can help document those alternatives, while the buyer research supplies evidence about their appeal.
Interpret the answers without inventing sales
Calculate a stated purchase share using a clear denominator:
Stated purchase share = respondents classified as willing to buy at that price ÷ all usable respondents evaluated at that price.
With a branched questionnaire, distinguish direct answers from classifications inferred from the price sequence. Document the threshold assumption and retain unresolved responses. Do not silently remove uncertain respondents to improve the percentage.
Suppose a hypothetical study has 200 usable respondents, complete price sequences and consistent answers. Their stated thresholds produce the following illustrative results:
| Total purchase price | Respondents classified as willing to buy | Share of the 200 respondents |
|---|---|---|
| $59 | 148 | 74% |
| $69 | 118 | 59% |
| $79 | 82 | 41% |
These are invented teaching figures, not findings from a customer study.
At $69, the calculation is 118 ÷ 200 = 59%. Moving from $69 to $79 reduces the stated share by 18 percentage points in this example. That identifies a range worth investigating; it does not mean that 59% of store visitors will purchase at $69.
Nor does a “yes” at $69 followed by a “no” at $79 establish an exact maximum of $69. It places the respondent’s stated threshold between tested prices, subject to the study’s assumptions. Someone accepting the highest tested price may have a higher threshold that the questionnaire never measured.
Check whether the pattern differs between the customer groups defined before fieldwork. Show counts alongside percentages. Review contradictory sequences and doubtful-quality responses using consistent rules, and disclose exclusions. Do not delete inconvenient answers because they weaken the result.
Check stated intent against actual purchases
Use the survey to identify prices and assumptions worth testing in a real buying situation. A completed purchase involves money, competing needs and checkout friction that a questionnaire cannot fully reproduce.
For an available product, a controlled test can compare completed purchases at different prices. Check applicable consumer rules and operational constraints before showing different prices to comparable shoppers. Randomly assign eligible shoppers, keep each shopper’s assigned price consistent, and hold product presentation, delivery terms and stock availability constant. Define the eligible audience, measurement window and primary outcome before starting.
For example, imagine two hypothetical test groups:
| Total purchase price | Eligible visitors exposed to the offer | Completed purchases | Observed purchase rate |
|---|---|---|---|
| $69 | 2,000 | 56 | 2.8% |
| $79 | 2,000 | 44 | 2.2% |
These figures are also invented. They illustrate measurement, not an expected conversion rate.
The purchase rates are 56 ÷ 2,000 and 44 ÷ 2,000. Their observed difference is 0.6 percentage points. Assess statistical uncertainty before attributing that difference to price. Equal visitor counts alone do not make the finding conclusive.
The survey’s 59% and the test’s 2.8% also have different denominators and conditions. One describes stated intent among screened respondents; the other describes purchases among exposed visitors. Neither number is a direct correction factor for the other.
Record cancellations and returns separately, and define whether your commercial assessment uses completed or retained orders. Check contribution margin alongside the purchase outcome before making a pricing decision.
If the product is unavailable, a genuine paid preorder can provide evidence under its stated delivery and refund terms. An email signup or a click on “Buy” measures an earlier action. Label it accordingly; it does not establish that someone paid the tested amount.
Hand over a research result with clear limits
Finish with a short record that another person can inspect and use:
- Audience: eligibility, recruitment channel and usable sample sizes.
- Offer: product, seller, currency, delivery, returns and total-price basis.
- Questions: wording, tested prices, assignment and classification rules.
- Results: response counts, uncertain answers, group differences and exclusions.
- Purchase evidence: actual orders, exposure counts, test conditions and remaining uncertainty.
Keep the conclusion tied to that record. “Among the screened category shoppers, stated acceptance fell between $69 and $79; a purchase test is needed” is a usable finding. “Customers will pay $69” discards the audience, conditions and uncertainty that make the result meaningful.
Once that evidence is assembled, it can inform the selling-price decision. If the seller, product specification or purchase terms change materially, revisit the assumptions before reusing the estimate.
See how your own catalog compares with the market
SkuPeek finds matching competitor offers and compares their prices and availability with your products.