The data that market research consistently produces and consistently misleads on

Ask people whether they would pay more for ethically sourced products. Most say yes. Measure whether they actually do. Most don’t. Ask people whether they would choose a healthier option at a restaurant. Most say yes. Observe what they order. Most don’t. The gap between stated preference and revealed behaviour is one of the most replicated and most commercially consequential findings in consumer psychology, and it persists regardless of whether the people being asked are trying to deceive the researcher or genuinely believe what they are saying.

The distinction was formalised by Paul Samuelson in 1938. His revealed preference theory made a foundational claim: the only reliable signal of what someone actually prefers is a real choice made with real consequences. Stated preferences — answers to survey questions, responses in focus groups, reactions during testing — reflect what people believe they want, or what they think they should want, or what they want to be seen wanting. They are not reliable predictors of what they will actually choose when money leaves their account and the decision has a cost.

The implication for entrepreneurs building products and pricing them is specific. Market research that asks people what they want is measuring something real — it is measuring the stated preference, the socially filtered, consciously constructed preference — but it is not reliably measuring the behaviour it appears to predict. The product designed to satisfy stated preferences will serve what customers say they want. The product that observes what they actually do, and why, will serve what they actually buy.

Why the gap exists: the two systems that govern behaviour differently

The stated-revealed preference gap is not primarily about dishonesty. It is about the architecture of decision-making.

Kahneman’s dual-process framework distinguishes between deliberate, effortful, conscious reasoning — System 2 — and the automatic, intuitive, largely non-conscious processing that governs most behaviour — System 1. Stated preferences are System 2 outputs. The survey participant engages their deliberate reasoning to consider what they would want, and they produce a considered, values-consistent answer. But the purchase decision at the point of sale is typically a System 1 event — fast, automatic, driven by habit, salience, and implicit emotional associations rather than by the values that governed the stated preference.

Phil Barden’s framework in Decoded establishes the practical commercial translation. The brain evaluates every potential purchase through a simple formula: perceived reward minus perceived pain. The reward dimension includes functional satisfaction, emotional association, and implicit social meaning — what the product says about you, not just what it does for you. The pain dimension includes price, effort, risk, and the cognitive cost of the decision itself. The purchase happens when the reward calculation exceeds the pain calculation — a process that occurs largely below conscious awareness, in a way that the deliberate values-reasoning that produces stated preferences has limited ability to predict or override.

The specific mechanisms that drive the gap

The social desirability effect is the most straightforward mechanism. When asked about preferences, people answer in the direction of the response they believe is socially acceptable or flattering to their self-image. This is not cynical self-presentation — it is the automatic operation of social cognition that shapes stated responses without the person being aware it is happening. The survey participant who says they would choose the organic option genuinely believes they would. Their actual purchase decision is governed by different cognitive processes operating under different conditions.

Temporal distance compounds this. When people state a preference, they are making an abstract future choice from a psychologically distant position. Research by Trope and Liberman on construal level theory established that temporally distant choices are evaluated primarily on desirability — values, ideals, and abstract principles. Temporally close choices, at the moment of purchase, are evaluated primarily on feasibility — convenience, price, and what is immediately available. The organic food preference stated in a survey is a desirability evaluation. The purchase decision at the supermarket shelf is a feasibility evaluation. These produce systematically different outcomes.

The measurement effect also operates — the act of asking about a preference changes it. Asking someone whether they would pay more for ethical products activates their ethical identity and produces a response that reflects that activated identity. The purchase decision occurs without that identity activation, in conditions governed by habit, salience, and the implicit associations that System 1 processes.

What the research shows actually predicts purchase behaviour

Past behaviour is the strongest available predictor of future behaviour — specifically, the behaviour of the same person in similar contexts. Implicit Association Tests, which measure the strength of automatic associations between concepts rather than explicit attitudes, predict behaviour more reliably than direct preference questions in multiple domains. Observational data — what people actually do in naturalistic settings — outperforms stated preference data across virtually every category where both have been collected.

The practical hierarchy for an entrepreneur designing a product or pricing strategy: observe actual behaviour first, then use stated preferences as a secondary signal about the language and values frame that resonates, not as a primary signal about what will actually be purchased.

What this means for product development and pricing

The most expensive version of this error is building a product that scores well in testing and underperforms in market. Testing produces stated preferences. Market performance reflects revealed ones. The gap between them is the stated-revealed preference gap operating at commercial scale.

Two specific applications. First, weight observational and behavioural research more heavily than survey research when making product decisions — what people are doing in adjacent categories, what they are actually clicking on, what they are actually buying in comparable situations. Second, when testing pricing, real willingness-to-pay experiments — where participants make actual purchase decisions with real money — are categorically more predictive than hypothetical price sensitivity questions, because the real decision activates the System 1 pain calculation that the hypothetical question bypasses.

If the experience of building something that tested well and sold poorly has significantly affected your wellbeing or confidence, that is worth acknowledging. UK: Samaritans (116 123, free, 24/7). Mind (0300 123 3393). International: iasp.info/resources/Crisis_Centres.

A book worth reading alongside this

Decoded by Phil Barden is the most practically applicable treatment of the gap between what consumers say and what actually drives their purchase decisions. Barden spent over 25 years as a senior marketer at Unilever, Diageo, and T-Mobile before immersing himself in decision science, and the book combines that practitioner experience with a rigorous treatment of the neuroscience and behavioural economics research. His reward-minus-pain framework for understanding purchase decisions is the most directly usable available model for translating the stated-revealed preference gap into product and pricing decisions. For any entrepreneur who has discovered that what customers said they wanted in research and what they actually bought were different things, this book provides the most direct and most practically grounded explanation of why — and what to design for instead.

Have questions about this article?

If any part of this article raised questions you want to explore further, courbot.co is built for exactly that. It is courben.co’s AI assistant, designed around the psychology of entrepreneurship. Ask it anything from this article.

If the patterns described here — building for what people say rather than what they do, and experiencing the commercial consequences — are significantly affecting your confidence or wellbeing, speaking with a psychologist is the right next step. UK: Samaritans (116 123, free, 24/7). Mind (0300 123 3393). BACP therapist finder: bacp.co.uk/search/Therapists. International: iasp.info/resources/Crisis_Centres. Crisis Text Line — text HOME to 741741.

This article is for educational and informational purposes only. Sources: Samuelson, P.A. (1938), Economica, 5(17), 61–71. Kahneman, D. (2011), Thinking, Fast and Slow, Farrar, Straus and Giroux. Trope, Y. & Liberman, N. (2010), Psychological Review, 117(2), 440–463. Barden, P. (2013), Decoded: The Science Behind Why We Buy, Wiley.