You cannot simply ask a customer what they will pay and trust the answer. This article covers research methods for measuring willingness to pay, from conjoint and MaxDiff to Van Westendorp and Gabor-Granger, and when to use each.
Willingness to pay is the foundation on which the whole framework rests, and it is the hardest thing to measure honestly. Ask a buyer directly, and they will understate it, because they have every reason to. The methods that work get at willingness to pay indirectly, through structured choices and observed behavior.
This article is the final one in our Pricing and Packaging series, and it provides the measurement behind the Customer and Value decisions and the elasticity model in this article that determines the profit-maximizing price. Throughout the series, we follow Meridian Software, an eighty-million-dollar workflow platform; here we choose its methods.
The simplest approach, asking how much someone would pay, is also the least reliable. People guess at what a product should cost rather than reflecting on its worth, and they shade their answer downward when they sense a price is being set. The number that comes back does not mean much.
Good pricing gets it right because the payoff is large. Directionally, a one percent improvement in price has been associated with a profit lift of around eleven percent, and a large share of new products miss their revenue targets due to weak pricing rather than weak demand. The methods below are how you replace a guess with evidence.
Each method answers a different question. Strong studies combine them, and run separately by segment.
The Acceptable Range
The Van Westendorp Price Sensitivity Meter asks four questions: at what price the product is too cheap to trust, a bargain, getting expensive, and too expensive to consider. Plotting the answers reveals an acceptable price range and an optimal point where the fewest buyers hit either extreme.
Its strength is that it needs no predefined prices, so it works early, when you do not yet know what the market will bear. It measures perception rather than purchase intent, which makes it a strong first step in a broader study rather than a final answer on its own.
The Demand Curve
Gabor-Granger, also called price laddering, shows a buyer a price and asks whether they would buy, then raises or lowers the price to find their ceiling. Repeated across respondents, it builds a demand curve and identifies the revenue-maximizing price.
Take the demand curve rather than the revenue-maximizing price it reports. As the elasticity modeling shows, the profit peak sits above the revenue peak by roughly half the cost to serve, so the price Gabor-Granger names is a floor on what you should charge rather than a target. The curve is the valuable output because it is what lets you subtract cost and find the peak that matters.
Its limitation is that it tests one product at a time and, because respondents know they are being asked about price, it can understate the true ceiling.
The Value of Each Feature
Conjoint analysis, or discrete choice, presents buyers with competing product configurations that trade features against price and asks them to choose. Because the choices mirror real buying decisions, statistical modeling can isolate how much each feature is worth and how buyers trade it against price.
This is the most powerful method for designing tiers because it shows which features drive the decision to upgrade and what each feature is worth to each segment. It is more complex to run than the direct-pricing methods, which is the price of its realism.
What Matters Most
MaxDiff, or best-worst scaling, presents buyers with small sets of features and asks them to identify which are the most and least important. Across many such sets, it produces a clean ranking of what buyers actually value, free of the flattening that happens when people rate everything as important.
It pairs naturally with conjoint: MaxDiff identifies which features matter, and conjoint quantifies their value. Together, they answer what belongs in each tier and how to price the step between them.
No single method is the answer. Van Westendorp finds the range, Gabor-Granger draws the curve, conjoint values the features, and MaxDiff ranks them. The strongest studies triangulate across several sources and then confirm the results against real-world data. The order matters as much as the mix. Start with the method that matches what you don't know: the range if you have no price yet, or the feature values if you have a price and a packaging problem. Running all four on a question you could have answered with one is how a research budget gets spent before it produces a decision.
The four methods above are stated-preference techniques: they ask buyers to make structured choices. The other family is revealed preference, which reads willingness to pay from what buyers actually did.
For a company with sales history, revealed-preference data is often the richest source of all. Win-loss records, discount patterns, and how quotes converted at different prices show real willingness to pay, with no survey bias, which is why we always ground survey findings against the deal data before trusting them.
The two families also fail in different ways, which is why they check each other. Stated preference overstates what buyers will do because saying yes in a survey costs nothing. Revealed preference only shows you what happened at the prices you have already charged, so it cannot tell you what a segment would have paid above your current price. Survey work explores beyond the range you have tested. Deal data is validated inside it.
Whatever method you choose, run it separately for each segment. Enterprise and mid-market buyers have different acceptable ranges, and the segment that depends on your product will pay far more than the one with easy alternatives.
A blended, market-wide willingness-to-pay number averages those differences into a figure that fits no one. Segment-level measurement makes the research actionable because it produces a different price for each group rather than a single compromise price for all.
Meridian Software, the platform we follow in this series, needs to measure willingness to pay for its two mid-market segments, which it has only ever priced based on assumptions. The method mix follows from what it is trying to learn.
Because Meridian has years of sales history, it starts with revealed preference: its own win-loss and renewal data, segmented into the mission-critical and occasional-use groups. It adds conjoint and MaxDiff to understand which features drive upgrades and belong in each tier, and a Gabor-Granger study on the mission-critical segment to build the demand curve that feeds the elasticity model.
Note what Meridian skips. Van Westendorp is the method for a company that does not know its range, and Meridian has been selling at $2,000 a month for years, so the range is already known from deal data. Skipping a method is part of designing the study, and the discipline is choosing what to learn rather than running everything available.
Within a few weeks, Meridian replaces the assumption behind its $2,000 price with measured evidence, confirming the higher price its mission-critical segment supports.
AI is compressing research that used to take quarters into weeks. It can continuously estimate willingness to pay from behavioral and transaction data, run and analyze choice-based studies faster, and keep the picture current as the market moves rather than freezing it in a one-time study.
The methods still matter, because they frame the questions correctly, but the speed and cost of answering them have dropped sharply. We cover this two-sided shift and how it reaches every pricing decision in our article on pricing power in the age of AI.
Willingness-to-pay research is the measurement layer beneath the whole framework. It grounds the six decisions in evidence, feeds the elasticity modeling that finds the profit-maximizing price, and is what carries a company across the instrumentation line on the maturity ladder.
This work sits inside our broader commercial transformation practice, where willingness-to-pay research connects pricing to product, marketing, and the evidence base the whole revenue engine relies on.
You cannot just ask what a customer will pay, but you can measure it. Van Westendorp finds the range, Gabor-Granger draws the curve, conjoint values the features, and MaxDiff ranks them, and revealed preference from your own deals grounds them all. Run every method by segment.
For Meridian, the right mix of methods turns a two-thousand-dollar assumption into a measured, defensible price. For any company, willingness-to-pay research is what turns the entire framework from a set of good decisions into evidence-based ones.
Download the Pricing and Packaging Framework
The complete framework, with the maturity ladder, the diagnostic, and a sequenced ninety-day plan.
The Pricing and Packaging Assessment
We design and run willingness-to-pay research with you, by segment, and turn the results into a price.
This completes our Pricing and Packaging series. To put it into practice, start with the framework overview, assess where you stand on the maturity ladder and in the diagnostic, and download the full guide for the complete toolkit. When you are ready to turn it into a price, we are here to help.
Sources
Method descriptions draw on standard pricing-research literature. The one-percent-price-to-profit figure is a widely cited McKinsey finding; the new-product pricing-failure figure draws on Simon-Kucher research. Figures are stated directionally rather than as precise claims.