Revealed preference is the idea that the best evidence of what people want is what they actually choose, given the prices and budget they face, rather than what they say they want. Economist Paul Samuelson set out the approach in a 1938 paper in Economica and gave it the name “revealed preference” in a 1948 paper, as a way to build demand theory on observed choices. In business it is the counterweight to stated preference: surveys, interviews and wish lists.
Record what people buy, use, renew or abandon when they face real prices and limits.
02
Hold the context in view
A choice only reveals a preference relative to the options, prices and budget the person faced.
03
Check for consistency
If someone picks A over B when both are affordable, they should not later pick B over A in the same conditions; repeated reversals signal noise or changed tastes.
04
Compare with what people say
The gap between survey answers and behaviour is often the most useful finding.
Where you see it today
Training AI on human choices
Modern chatbots are tuned partly on which of two answers people prefer. The method traces to work showing that a reinforcement learning agent could learn complex goals from humans’ choices between pairs of behaviour clips, without a hand-written reward function (Christiano et al., 2017). The model learns from what raters pick, not from a written description of what they want.
Measuring real AI use
Surveys ask people whether they use AI at work. The Anthropic Economic Index instead analysed millions of anonymised Claude conversations, mapped to occupational tasks, and found usage concentrated in software development and writing (Handa et al., 2025). It covers one product’s users, but it is a record of behaviour rather than opinion.
Product analytics over polls
Software firms increasingly trust retention, renewals and feature usage more than satisfaction scores. A customer who says an AI feature is “great” but never opens it has told you something different with their behaviour.
Feeds that learn from clicks
Recommendation systems rank content by what people watch, click and skip. That is revealed preference at scale, with a caveat: an impulse click is a weak signal of what someone would choose on reflection.
Revealed preference vs other ways to read demand
Revealed preference
What it capturesActual choices under real prices and limits
Main weaknessShows what, not why; needs real behaviour to exist
Main weaknessCheap to say; often differs from what people do
Choice experiments (conjoint)
What it capturesTrade-offs in hypothetical choice sets
Main weaknessStill hypothetical; no money at stake
A/B testing
What it capturesBehaviour under a controlled change
Main weaknessAnswers narrow questions; short time frames
Usage telemetry
What it capturesWhat people do inside a product
Main weaknessMisses those who never adopted it
How to use it
Before trusting a survey, find one behavioural metric that should move if the survey is right, and check it.
Track renewals, repeat use and spend for each AI tool you pay for, not staff opinions of it.
When customers ask for a feature, look at how they use the nearest existing feature first.
Price tests beat pricing surveys: offer real options at real prices to a small group.
Read choices in context: a choice made with no good alternative reveals little about real preference.
The Business Engineer lens
AI adoption is full of stated preference: announcements, pilots and survey claims that everyone uses AI. The Business Engineer reads adoption from revealed signals such as usage depth, renewals and budget moved, because that is where the economics of the shift actually show up.
It holds that people's actual choices, given the prices and budget they face, are better evidence of their preferences than what they say.
Who developed revealed preference?
Paul Samuelson set out the approach in a 1938 paper in Economica and named it revealed preference in 1948.
What is the difference between revealed and stated preference?
Stated preference is what people report in surveys or interviews; revealed preference is inferred from what they actually choose when it costs them something.
How is revealed preference used in AI?
Many AI models are tuned on which of two outputs people prefer, and usage data is used to see which AI tools people actually rely on.
The complete explainer, cleaned up and reformatted for mobile.
Revealed preference, is a theory offered by the American economist Paul Samuelson in 1938. The theory asserts that consumers’ behavior – assuming a constant income and an item’s price – is the best indicator of their hidden preferences. In short, that is how people reveal what they really want.
Revealed Preference
ExplanationRevealed Preference is an economic theory that suggests that the preferences of consumers can be inferred from their actual purchasing decisions. It is based on the idea that individuals reveal their true preferences through the choices they make when faced with different options in the marketplace. The theory was developed by American economist Paul Samuelson in 1938.
Consumer Choices
Explanation– Consumer Behavior: Revealed preference theory focuses on studying the choices consumers make when selecting goods and services. These choices reflect individual preferences. – Utility Maximization: Consumers are assumed to make decisions that maximize their utility or satisfaction given their budget constraints and the prices of goods.
Assumptions
Explanation– Transitivity: If a consumer prefers option A to option B and prefers option B to option C, then they should prefer option A to option C. – Completeness: Consumers are assumed to have complete and consistent preferences. They can rank all possible options. – Non-Satiation: More is preferred to less. Consumers prefer having more of a good to less, all else being equal.
Market Equilibrium
ExplanationRevealed preference theory has implications for market equilibrium. When consumers make choices based on their preferences and budget constraints, it contributes to the determination of market prices and the allocation of resources.
Preference Maps
ExplanationPreference maps are graphical representations used to depict consumer preferences. These maps show how consumers make choices when faced with different combinations of goods and their respective prices. They help economists analyze and understand consumer behavior.
Applications
Explanation– Consumer Demand Analysis: Revealed preference theory is used to analyze consumer demand for various goods and services. It helps businesses understand which products consumers prefer and at what price points. – Welfare Economics: It is employed to assess consumer welfare and make policy recommendations related to taxation, subsidies, and market regulations.
Limitations
Explanation– Incomplete Information: In some cases, consumers may not have complete information about all available options, leading to incomplete preferences. – Violations of Assumptions: In practice, consumers may not always exhibit transitive or complete preferences, making it challenging to apply revealed preference theory in all situations.
Hicksian Demand Curve
ExplanationIn the context of revealed preference, economists often use the Hicksian demand curve, also known as the compensated demand curve. It shows the quantity of a good that a consumer would demand at different price levels while maintaining the same level of utility or satisfaction as in their initial choice.
Consumer Surplus
ExplanationRevealed preference theory plays a role in the concept of consumer surplus. Consumer surplus represents the difference between what consumers are willing to pay for a good and what they actually pay. It is a measure of the net benefit consumers receive when purchasing a product.
Behavioral Economics
ExplanationIn recent years, behavioral economics has expanded on traditional revealed preference theory by incorporating insights from psychology and behavioral science. Researchers examine how cognitive biases and heuristics influence consumer choices, sometimes deviating from rational utility maximization.
Policy Implications
ExplanationRevealed preference theory informs various economic policies, particularly in the areas of pricing, taxation, and market regulation. It helps policymakers understand how changes in prices and policies impact consumer behavior and welfare.
Future Developments
ExplanationAs the field of economics evolves, revealed preference theory continues to be a foundational concept. Future developments may involve integrating behavioral insights and advanced econometric techniques to provide a more nuanced understanding of consumer choices.
Taleb’s view on revealed preferences
The axiom of revelation of preferencesstates the following: you will not have an idea about what people really think, what predicts peopleβs actions, merely by asking them βthey themselves donβt know.
There is no such thing as βrationalityβ of a belief, there is rationality of action
The rationality of an action can only be judged by evolutionary considerations
This theory of rationality moves around a few key points which can shift your perspective and make you a better business person.
Science and scientism
As science has been so successful in certain fields (take physics) over the centuries, we matured the belief that scientific principles developed by hard sciences could be applied to soft sciences, or to human nature in general.
This belief brought to the rise of scientism, or the core belief that things could be stripped from their “irrational” part to make them rational.
Here rationality, though, is measured in terms of beliefs rather than the actions that come from those beliefs.
And this is the primary fallacy scientism falls into.
A person imbued with scientism will criticize religion, yet she will believe in economic theories that narrow down reality to a model that is not only far from it, but that from the collective standpoint might have negative consequences for society overall.
Domain dependence
Understanding that the scientific method can be used in certain fields, but it loses value in other domains, is critical. In the realm of human, messy affairs, it is essential to have a different mindset. While experimentation, it’s still the key, it will have a different meaning.
Experimentation won’t be tied to a narrow domain, but close to creativity and the ability to draw ideas from fields and domains that are far off from the one where you’re trying to apply a solution. Thus, it will require a broad range of creativity, a lot of experimentation, and the ability to have a broad spectrum of experiments.
Rationality based on survival of fit actions
Your eyes are not sensors aimed at getting the electromagnetic spectrum of reality. Their job description is not to produce the most accurate scientific representation of reality; rather the most useful one for survival.
Still, in his article “How to be Rational about Rationality,” Taleb highlights how eyes are wired to survival rather than giving an accurate or scientific representation of reality.
This view highlights how rationality is not about logical explanations. As many things in the real messy world can’t be explained, as there is a hidden meaning, which might not be revealed.
We need to trust more the subconscious side, which calls up to look us for actions that survive time, rather than explanations and how those might fit the current context or prevailing intellectual view.
Skin in the game
What matters, in the end, is what they pay for goods, not what they say they βthinkβ about them, or what are the reasons they give you or themselves for that. (Think about it: revelation of preferences is skin in the game).
Another core concept connected to this revealed preference view is that of skin in the game or the fact that in order to reveal what people really want.
One way is to have them pay for something, as this implies an action, which isn’t frictionless, thus if performed might reveal strong subconscious preferences, hard to explain rationally, but work as a strong hidden signal.
Beliefs are cheap; actions do cost
For beliefs are β¦ cheap talk. A foundational principle of decision theory (and one that is at the basis of neoclassical economics, rational choice, and similar disciplines) is that what goes on in the head of people isnβt the business of science. First, what they think may not be measurable enough to lend itself to some scientific investigation. Second, it is not testable. Finally, there may be some type of a translation mechanism too hard for us to understand, with distortions at the level of the process that are actually necessary for think to work.
For the sake of understanding this theory, it’s essential to look at actions!
Key takeaways
Revealed preferences can help us uncover what people really want.
Revealed preferences are about what people do when they are skin in the game.
It’s also about rationality meant as the survival of actions that gain us fit in the real world.
It’s not about elegant theories or beliefs, but it’s about those beliefs that lead to actions that make the collective survive.
In short, that is about uncovering hidden cues from the subconscious mind, so that we avoid attaching rationality to things in hindsight.
Case Studies
Purchasing Behavior:
Decision: Consumers claim they prioritize eco-friendly products.
Action: However, when shopping, they consistently choose cheaper, non-eco-friendly alternatives.
Revealed Preference: Price is more important to these consumers than eco-friendliness.
Transportation:
Decision: A city’s residents often voice support for public transportation.
Action: Yet, when public transit is available, many continue to use their personal vehicles.
Revealed Preference: Convenience or personal space might be more valued than supporting public transport.
Dietary Choices:
Decision: Individuals frequently state a desire to eat healthily.
Action: However, fast-food chains have consistent long lines and booming sales.
Revealed Preference: Taste and convenience may be more important than health considerations.
Work-Life Balance:
Decision: Employees express the importance of work-life balance.
Action: Yet, when offered overtime or promotions that require longer hours, many accept.
Revealed Preference: Financial incentives or career progression might be valued more than personal time.
Online Privacy:
Decision: Internet users claim to value online privacy and data security.
Action: However, they frequently use and sign up for platforms and apps without reading terms of service.
Revealed Preference: Using the platform or app immediately might be more valuable than ensuring data privacy.
Education:
Decision: Students say they are committed to their studies.
Action: Many skip classes or don’t study regularly.
Revealed Preference: Leisure or other activities might be more appealing than academic diligence.
Environmental Concerns:
Decision: People voice concerns about plastic waste.
Action: They continue to buy bottled water and use single-use plastics.
Revealed Preference: Convenience may outweigh environmental concerns.
Social Activism:
Decision: Many express support for various social causes online.
Action: Fewer participate in physical rallies, donations, or volunteer work.
Revealed Preference: Online expression might be easier and more appealing than on-ground activism.
Fitness Goals:
Decision: Individuals purchase gym memberships at the start of the year as a New Year’s resolution.
Action: A significant portion stops attending after a few weeks.
Revealed Preference: The idea of fitness might be more appealing than the actual effort required.
Real Estate:
Decision: Homebuyers claim they want a house in a quiet, suburban area.
Action: They end up buying homes in bustling city centers.
Revealed Preference: Proximity to work, entertainment, or urban lifestyle might be more important than tranquility.
Key Highlights
Revealed Preferences: Revealed preferences theory suggests that consumers’ behavior, assuming a constant income and item price, is the best indicator of their hidden preferences. It means people reveal what they really want through their actions.
Nassim Nicholas Taleb’s View: Taleb’s perspective on revealed preferences revolves around the idea that judging people solely on their beliefs is not scientific. Rationality of action can only be judged by evolutionary considerations, and it’s not about elegant theories or beliefs, but about the actions that lead to survival in the real world.
Scientism and Domain Dependence: Scientism, the belief that scientific principles can be applied to all areas, including human nature, can lead to fallacies. In domains like messy human affairs, a different mindset is needed, relying on creativity and broad experimentation.
Rationality Based on Survival: Rationality is not always about logical explanations; it’s about actions that lead to survival and adaptation over time. Trusting the subconscious and focusing on actions that have stood the test of time can be more reliable than trying to explain everything rationally.
Skin in the Game: Revealed preferences are best observed when people have skin in the game, i.e., when they have a personal stake or financial commitment to their choices. Actions that require commitment and resources can reveal strong subconscious preferences.
Beliefs vs. Actions: Beliefs are cheap and may not be measurable or testable enough for scientific investigation. In understanding revealed preferences, the focus should be on actions rather than beliefs.
Uncovering Hidden Cues: Revealed preferences help us uncover hidden cues from the subconscious mind, allowing us to understand what people truly want, avoiding attaching rationality to things in hindsight.
Stated Preference Methods
Description– Research techniques used to directly ask individuals about their preferences, attitudes, or choices regarding specific products, services, or policies. Stated Preference Methods include surveys, questionnaires, and choice experiments.
When to Apply– When seeking to understand individuals’ preferences or attitudes. – Using Stated Preference Methods to gather data on consumer preferences, perceptions, and willingness to pay for products or services, informing marketing strategies, product development decisions, and policy design.
Conjoint Analysis
Description– A market research technique used to measure consumers’ preferences and trade-offs for different product attributes or features. Conjoint Analysis presents respondents with hypothetical product profiles and asks them to make choices, revealing their preferences based on trade-offs.
When to Apply– When evaluating product features or attributes. – Conducting Conjoint Analysis to identify the relative importance of different product attributes, assess market demand, and optimize product design and pricing strategies, informing Revealed Preference models and decision-making processes.
Market Basket Analysis
Description– A data mining technique used to analyze patterns of co-occurrence or association between products purchased by customers. Market Basket Analysis identifies frequently co-purchased items or product combinations, revealing consumers’ preferences and purchase behavior.
When to Apply– When analyzing consumer purchase behavior and preferences. – Applying Market Basket Analysis to identify cross-selling opportunities, understand consumer preferences, and personalize marketing strategies and product recommendations based on revealed purchase patterns, supporting Revealed Preference modeling and decision-making processes.
Choice Modeling
Description– A quantitative research method used to model and predict consumer choices or decisions based on observable attributes, preferences, and contextual factors. Choice Modeling encompasses various techniques such as discrete choice models, multinomial logit models, and random utility models.
When to Apply– When predicting consumer behavior or preferences. – Using Choice Modeling techniques to estimate choice probabilities, forecast market demand, and simulate consumer decisions under different scenarios, guiding Revealed Preference analysis and informing strategic decision-making in marketing, product development, and pricing.
Preference Mapping
Description– A statistical technique used to visualize and analyze consumers’ preferences and perceptions regarding products, brands, or attributes. Preference Mapping identifies underlying dimensions of preference and segments consumers based on their preferences.
When to Apply– When analyzing consumer preferences or perceptions. – Conducting Preference Mapping to identify key dimensions of preference, visualize consumer segments, and understand the relationships between products, attributes, and consumer preferences, supporting Revealed Preference analysis and strategic decision-making in market segmentation, product positioning, and brand management.
Demand Estimation
Description– Involves quantifying the relationship between product demand and various factors such as price, income, demographics, and marketing activities. Demand Estimation helps businesses forecast sales, optimize pricing strategies, and allocate resources effectively.
When to Apply– When analyzing market demand and consumer behavior. – Estimating demand functions using econometric techniques to quantify the responsiveness of demand to changes in prices, incomes, or other variables, informing Revealed Preference models and strategic decision-making in pricing, product planning, and resource allocation.
Discrete Choice Experiments
Description– A research method used to elicit individuals’ preferences and choices among a set of alternatives with varying attributes or levels. Discrete Choice Experiments present respondents with choice scenarios and ask them to select their preferred option, allowing researchers to estimate choice probabilities and attribute preferences.
When to Apply– When assessing consumer preferences or trade-offs. – Conducting Discrete Choice Experiments to design choice scenarios, measure attribute importance, and estimate choice models, revealing consumers’ preferences and trade-offs, informing Revealed Preference analysis and strategic decision-making in product development, marketing, and policy design.
Hedonic Pricing Model
Description– A statistical model used to estimate the value of goods or services based on their attributes, characteristics, or quality. Hedonic Pricing Modelsanalyze how variations in product attributes affect prices and consumer preferences.
When to Apply– When pricing products or valuing attributes. – Applying Hedonic Pricing Models to estimate the implicit prices of product attributes, amenities, or features, quantify consumer preferences, and assess the value of quality improvements or changes in product specifications, supporting Revealed Preference analysis and pricing strategy formulation.
Brand Equity Analysis
Description– Involves assessing the value and strength of a brand based on consumers’ perceptions, attitudes, and associations with the brand. Brand Equity Analysis measures brand awareness, brand loyalty, perceived quality, and brand associations.
When to Apply– When evaluating the strength and value of a brand. – Conducting Brand Equity Analysis to measure brand performance, assess consumer perceptions and preferences, and identify opportunities to enhance brand equity through marketing initiatives, informing Revealed Preference analysis and strategic decision-making in brand management and marketingstrategy.
Experimental Economics
Description– A branch of economics that uses controlled experiments to study individual behavior, decision-making, and market outcomes in controlled laboratory or field settings. Experimental Economics provides insights into consumer preferences, decision-making processes, and market dynamics.
When to Apply– When studying consumer behavior and preferences in controlled environments. – Conducting controlled experiments to elicit consumer preferences, test economic theories, and analyze decision-making processes, providing empirical data for Revealed Preference analysis and informing strategic decision-making in marketing, pricing, and policy design.
Gennaro Cuofano is a CRO and tech executive who has worked in AI since late 2015, starting with bringing NLP, natural-language generation, voice and chatbot products to the marketing industry. His background is in law and finance: he holds a Master's degree in Law and an International MBA with an emphasis on corporate finance (LUISS Business School and the University of San Diego, 2012), and worked as a financial analyst at a real-estate investment firm in San Diego and as an assistant controller. His work focuses on business model strategy, business engineering and, more broadly, structural analysis: how companies make money, read from their own filings. He created FourWeekMBA and leads research there and at The Business Engineer, his newsletter on AI and business strategy, with over 95,000 subscribers and more than 1,000 published analyses. His writing has also appeared on Entrepreneur (for example, How to Digitize a Family Business in Europe) and HackerNoon. Find him on LinkedIn and Substack. See how we source.
