When a business is preparing to price a new Product, the most obvious research question is:
“How much are customers willing to pay?”
But the most direct question does not necessarily produce the answer closest to real behavior.
In a Survey, customers are not spending real money. Competitors may not appear beside the Product. Their actual Budget Constraint may not be active, and their answer has no financial consequence.
This creates a possible gap between Stated Willingness to Pay and Actual Purchase Behavior. A Meta-analysis covering 77 Studies on consumer goods found measurable Hypothetical Bias in WTP estimates and showed that accuracy varies across methods and contexts.
A better Pricing Research design therefore does not search for one magical number. It creates situations where customers evaluate Price together with Product Value, alternatives and trade-offs, then uses the result as evidence for the Pricing Decision.

Instead of asking “How much would you pay?”, create a choice customers must evaluate
A direct question such as “What is the maximum you would pay?” is simple, but the answer is still hypothetical. The respondent is not necessarily facing a real Budget Constraint, competitive alternatives or financial consequences.
Pricing Research therefore often uses methods that create more structured price decisions:
1. Gabor-Granger: show defined Price Points and ask whether the respondent would buy
2. Van Westendorp Price Sensitivity Meter: identify prices perceived as too cheap, good value, expensive but acceptable and too expensive
3. Conjoint / Discrete Choice: require customers to choose between Product Packages with different Features and Prices
4. Monadic Price Testing: show different respondent groups the same Product at different Prices
5. Behavioral Price Experiments: observe real purchasing responses to Price changes where feasible
No method eliminates Hypothetical Bias completely. Meta-analytic research on Consumer Willingness to Pay shows that hypothetical WTP can differ from real WTP, and the size of that difference depends on the Method and Context.

Before measuring Willingness to Pay, clarify what the business actually needs to know

Willingness to Pay is often used as a broad label for several different Pricing Questions.
A business may really need to know:

• What Price Range feels acceptable?
• How much would Purchase Intent fall if Price moves from 299 to 349 baht?
• Does a new Feature create enough Value to justify a Price Premium?
• Can a 499-baht Package compete with a 399-baht alternative?
• Which Price Point could maximize expected Revenue?
• Which Features will customers trade away for a lower Price?

These are different questions.
They should not automatically use the same research method.
Qualtrics distinguishes common Pricing Research approaches such as Van Westendorp, Gabor-Granger and Conjoint Analysis according to the type of Pricing Decision being investigated.

Why does the direct “maximum price” question have limitations?

Imagine testing a new premium Ready-to-drink Coffee.
You ask: “What is the maximum you would pay per bottle?”
Respondent A answers: 65 baht
We still do not know:
Whether they would actually purchase at 65 baht
Whether a 49-baht competitor would change their decision
Whether different Packaging would change the answer
What “maximum” means to that respondent
Whether they intentionally understate WTP because they do not want the Brand to set a high price
Whether they overstate it because no money is actually being spent
This is why Stated WTP should be treated as evidence about preference under a specific hypothetical setting, not as an automatic market price.
Research on Hypothetical Bias shows that stated valuations can differ from WTP observed under real financial consequences, with the size and direction of the gap depending on Research Design and Context.

Method 1: Gabor-Granger asks customers to react to Price Points rather than invent a number

Gabor-Granger presents respondents with predetermined prices and asks whether they would consider purchasing at each level.
Example:
Weekly Meal Subscription 599 baht → Would you buy?
If Yes, the next price may be 699 baht.
If No, it may move down to 549 baht.
The logic gradually identifies where the respondent switches from acceptance to rejection.
Qualtrics describes Gabor-Granger as a method for measuring Price Sensitivity by testing several Price Options and using the responses to build Demand and Revenue Curves.

When is Gabor-Granger useful?

It is useful when:

• The Product Concept is relatively well defined
• The business already has a plausible Price Range
• The objective is to understand Price Sensitivity
• Several specific Price Points need to be compared

Example Business Question: “What happens to Purchase Intent when Price increases from 399 to 449 baht?”

What does Gabor-Granger not solve?

The respondent is still answering a Survey.
Therefore: Purchase Intent ≠ Actual Purchase
The sequence of prices can also create Anchoring or Context Effects if the design is poor.
Treat the result as a Demand Signal and Pricing Input rather than a literal Sales Forecast.

Method 2: Van Westendorp estimates an Acceptable Price Range rather than asking for one maximum price

The Van Westendorp Price Sensitivity Meter uses four types of Price Perception questions, typically covering:

• A Price so low that quality becomes questionable
• A Price that feels like good value
• A Price that feels expensive but still acceptable
• A Price that becomes too expensive to consider

The distributions are then used to identify psychological price points and an Acceptable Price Range.
The respondent is therefore not answering one direct: “What is your maximum price?”
question. They are evaluating Price from several perspectives.

When is Van Westendorp useful?

It is useful when:

• The Product or Service is relatively easy to understand
• The business is uncertain about the market’s broad Price Range
• Directional evidence about Price Perception is useful
• The team wants to understand both “too cheap” and “too expensive” boundaries

What does Van Westendorp not tell you?

An Acceptable Price is not automatically the Optimal Business Price.
The result still needs to be combined with:

• Cost
• Margin
• Competitor Pricing
• Positioning
• Demand
• Channel Economics
• Strategic Objectives

Qualtrics also notes that Pricing Survey responses require validation because stated preference may not hold in an actual purchase situation.

Method 3: Conjoint Analysis asks customers to choose Product and Price together

In real markets, customers rarely evaluate Price separately from the Product.
They might choose between:
Package A
500 GB
1-day Delivery
Premium Support
599 baht

and: Package B
1 TB
3-day Delivery
Standard Support
649 baht
Choice-based Conjoint creates tasks where respondents choose between Product Profiles made up of different Features and Prices.
Those choices are used to estimate Relative Utility for Attributes and the trade-offs customers make.
When Price is included as an Attribute, Conjoint can also estimate Willingness to Pay for particular Feature Levels relative to a Base Case.

When is Conjoint useful?

It is particularly useful when the Pricing Decision also involves Product Configuration.
For example:

• How much more can we charge for a Premium Feature?
• Which Features belong in each Package?
• Does the customer value Delivery Speed more than Brand?
• What happens to Preference if Price rises but an additional Feature is included?
• How should Good / Better / Best Packages be structured?

The key advantage of Conjoint is Trade-off

A direct question asks: “Is Feature A important?” Many customers may say yes.
Conjoint requires a choice between: More Features + Higher Price and: Fewer Features + Lower Price
That moves closer to the decision logic customers face in a market.
Qualtrics describes Conjoint as a method for evaluating Price-Feature trade-offs and recommends using simulation to compare preference across alternative packages.

Conjoint is not “the truth of the market” either

Although Conjoint creates more realistic choice situations, many Conjoint exercises remain hypothetical.
Respondents may not spend real money.
The real market may contain more competitors.
Brand, Promotion, Availability and context may differ.
A review of Hypothetical Bias in Choice Experiments found that bias remains a real concern, although its direction and magnitude vary substantially by application and study design.
Conjoint is therefore most useful for estimating:
Relative Preference
Trade-offs
Scenario Comparison
rather than claiming: “Customers will definitely pay this exact amount.”

Method 4: Monadic Price Testing compares reactions without showing every respondent all prices

In a Monadic Price Test, respondents are split into groups.
Each group sees the same Product Concept. But each group sees a different Price.
Example:
Group A → 299 baht
Group B → 349 baht
Group C → 399 baht
Then measure:

• Purchase Intent
• Value for Money
• Appeal
• Uniqueness
• Expected Quality

Because one respondent does not see the entire Price Ladder, each Price can be evaluated more independently.
The trade-off is Sample Size.
Each Price Cell needs enough respondents to support a useful comparison.

Method 5: Behavioral Price Experiments observe Actual Behavior where feasible

If a business has sufficient Traffic, transactional systems and appropriate operating conditions, real-world Price Tests can provide evidence closer to Actual Purchase.
For example:
Market A → 390 baht
Market B → 420 baht
Then compare:

• Conversion
• Units Sold
• Revenue
• Gross Margin
• Repeat Purchase
• Refund / Cancellation

In some cases, A/B Price Tests may be possible, subject to appropriate legal, ethical and operational constraints.
The advantage is that customers make decisions with real consequences.
However, businesses must consider:

• Fairness
• Customer Trust
• Channel Conflict
• Price Transparency
• Legal or Regulatory Requirements
• Contamination between groups
• Concurrent Promotions

Behavioral Experiments can provide stronger evidence for some causal pricing questions, but they are not feasible or appropriate in every setting.

Stated Preference and Revealed Preference work better together than as competing philosophies

Pricing Research operates in two evidence worlds. Stated Preference. What customers say or choose in Research.
Examples:

• Purchase Intent
• Van Westendorp
• Gabor-Granger
• Conjoint Choice

Revealed Preference
What customers actually do under real constraints.
Examples:

• Transactions
• Conversion
• Churn after a Price increase
• Purchase Mix
• Actual response to a Promotion

Stated Preference is valuable when testing Products or Prices that do not yet exist.
Revealed Preference is closer to actual behavior.
But Revealed Preference is not perfect either because historical market data often contains Confounding Factors such as Promotion, Distribution Changes and Competitor Actions.
A stronger workflow is often: Research → Price Hypothesis → Market Test → Learn → Adjust
rather than trying to discover one permanent “true WTP” estimate.

Do not rely only on Average Willingness to Pay

Suppose Average WTP is: 520 baht
But the distribution is:
Segment A = 700–800 baht
Segment B = 450–550 baht
Segment C = below 350 baht
The 520-baht average may not describe any real Customer Segment particularly well.
Pricing decisions should therefore consider:

• Segment
• Use Case
• Need Intensity
• Purchase Frequency
• Income / Budget Context when relevant
• Existing Alternatives
• Brand Relationship
• Channel
• Geography

Conjoint can estimate WTP at the respondent level before aggregating, allowing researchers to examine variation rather than relying only on the overall average.

Price Sensitivity is not only a customer trait—it changes with the situation

The same customer may be willing to pay differently when:
Buying for themselves
Buying a gift
Facing urgency
Having many Alternatives
Receiving a Promotion
Experiencing Stock-outs
Trusting one Brand more than another
Entering a different Budget Period
WTP should therefore not be treated as a permanent number attached to a customer.
A more accurate interpretation is: Willingness to Pay under a specific Product, Context, Alternatives and Timing.
That is why a Pricing Study needs a clearly defined buying situation.

Make sure the Product Concept is understood before testing Price

If respondents do not understand the Product, their Price answers become difficult to interpret.
Consider: “What is the right Price for an AI-powered Business Analytics Platform?”
One respondent imagines a Dashboard.
Another imagines a CRM.
Another imagines an AI Consultant.
Another imagines Enterprise Software.
Their Price responses are not based on the same Product.
Before Pricing Research, the Stimulus should clearly communicate:

• Product
• Target
• Main Benefit
• Important Features
• Usage Model
• Package / Quantity
• Payment Period
• Relevant Alternatives

A sophisticated Pricing Method cannot compensate for an unclear Product Definition.

Do not ask about Price before the respondent has enough Value Context

Question Order matters.
If the Survey begins with: “How much would you pay?”
before explaining the Product, the respondent may anchor on the first Category Price or Budget that comes to mind.
But if the Product description becomes a Sales Pitch, Value Perception may be artificially inflated.
The Stimulus should therefore be:
Neutral
Clear
Comparable
Realistic
and contain approximately the information a customer would normally have at the point of choice.

Revenue-maximizing Price is not always Profit-maximizing Price

Gabor-Granger can be used to estimate Demand and Revenue Curves from Price Acceptance data.
But the Price that maximizes Revenue does not automatically maximize Profit.
Example:
Price 400 baht
Expected Demand = 1,000 Units
Revenue = 400,000 baht

Price 500 baht
Expected Demand = 850 Units
Revenue = 425,000 baht
The 500-baht Price looks stronger on Revenue.
The business still needs to consider:

• Variable Cost
• Gross Margin
• Fulfillment
• Promotion Cost
• Customer Acquisition Cost
• Repeat Purchase
• Strategic Positioning

Pricing Research provides Demand-side Evidence.
The Pricing Decision still requires Business Economics.

Example: Should a new Product be priced at 399 or 449 baht?

Suppose the Product Concept is relatively clear and the business is deciding between:
399 baht and: 449 baht
A direct: “How much would you pay?”
question might produce answers ranging from 250 to 600 baht, with little guidance.
A stronger workflow may be:
Step 1 Run a Concept Test to confirm Need, Benefit and Value Perception.
Step 2 Use Gabor-Granger to test Purchase Intent at 349 / 399 / 449 / 499 baht.
Step 3 Compare Demand Curves across Customer Segments.
Step 4 Calculate Revenue and Margin Scenarios.
Step 5 Where feasible, Pilot 399 and 449 baht in comparable markets or periods.
Step 6 Evaluate Conversion, Margin and Repeat Behavior.
The result might show that 399 baht creates higher Purchase Intent, while 449 baht generates stronger Contribution Margin with only a modest decline in Demand.
In that case: “Which Price do customers like more?”
is not the same question as: “Which Price should the business choose?”

Another example: Is a Premium Feature worth an extra 100 baht?

Suppose the question is: “How much extra will customers pay for Same-day Delivery?”
A direct WTP question can generate a number.
But it does not reveal how customers make the trade-off when other Features also change.
A Conjoint exercise could present:
Option A
Standard Delivery
Basic Support
499 baht

Option B
Same-day Delivery
Premium Support
599 baht

Option C
Next-day Delivery
Basic Support
549 baht
Choice Patterns help estimate the Relative Value of Same-day Delivery in the presence of other Features and Prices.
This is a situation where Conjoint is more appropriate than Van Westendorp because the Business Question concerns Attribute Trade-offs rather than only an Acceptable Price Range.

Eight questions before choosing a Willingness to Pay Method

1. Do we need a Price Range or a specific Price Point?
2. Is the Product Concept clear enough to price?
3. Do customers need to trade Price against Features?
4. Do we already have a plausible Price Range to test?
5. Do we need Directional Evidence or evidence for a High-stakes Decision?
6. Do we need results by Customer Segment?
7. Can we test Actual Behavior in the market?
8. How will Cost, Margin, Competition and Positioning enter the final Pricing Decision?

These questions are more useful than asking only: “Which Method is the most accurate?”

What Pricing Research does not allow you to claim

A Survey shows: 62% Purchase Intent at 499 baht
Do not automatically claim: “62% of the market will buy.”
Van Westendorp produces: Acceptable Range = 400–550 baht
Do not automatically claim: “The best Price is the midpoint.”
Conjoint estimates: WTP for Feature X = +80 baht
Do not automatically claim: “An 80-baht Price increase will definitely increase Revenue.”
An Experiment shows: Conversion -5% after a Price increase
Do not automatically claim: “Price was the only Cause.”
Every Pricing Method has Assumptions, Context and Limitations.
A useful Pricing Research report should state:
Method
Target Population
Product / Price Stimulus
Sample
Context
Outcome
Key Assumptions
Limitations
That makes the evidence easier to interpret and easier for others to cite accurately.

The takeaway: Do not try to discover Willingness to Pay with one question

Willingness to Pay is not a fixed number hidden inside the customer waiting for a researcher to ask for it.
It changes with:
Product
Benefit
Alternatives
Context
Price
Timing
and real purchasing constraints
If the goal is an Acceptable Price Range, Van Westendorp may be a useful starting point.
If the goal is Demand Response across defined Price Points, use Gabor-Granger.
If the question is how customers trade Price against Features, use Conjoint / Discrete Choice.
If the business needs evidence closer to Actual Behavior and the context allows it, add a Behavioral Price Test.
Research on WTP shows that differences between hypothetical and real WTP can occur, so no Survey Method should be treated as perfect proof of future purchasing behavior.
The more useful question is therefore not simply: “How much are customers willing to pay?”
It is: “Given the Product, Alternatives and Price Choices that best represent the real decision, what trade-offs do customers make—and when we combine that evidence with Cost, Margin and Strategy, which Price makes the most sense for the business?”