Paul Samuelson’s (1938) revealed preference argument is nearly ninety years old and remains the most important idea in customer research. The argument is precise: preferences are not directly observable psychological states. They are inferences drawn from choice behaviour. When researchers ask customers to state their preferences, they are asking customers to introspect on and report internal states they have limited access to, that are context-dependent, and that are routinely shaped by the social dynamics of the research situation. The purchase reveals preference; the survey reveals what the customer believes about themselves, what they think the researcher wants to hear, and what they can articulate about decision processes that are mostly not deliberate.

Why customers cannot accurately report what drives their decisions

Two distinct mechanisms produce the gap between stated and revealed preference. The first is introspective limitation. Nisbett and Wilson’s (1977) research established that people have significantly less access to their own mental processes than they believe. System 1 drives the majority of purchasing decisions — the heuristic evaluation, the emotional response, the automatic pattern matching — and System 2 has limited ability to accurately report on System 1’s operations after the fact. The customer who says they bought the product because of the features is typically constructing a post-hoc explanation for a decision that was made on different grounds than the explanation implies.

Johansson et al.’s (2005) choice blindness research provided the most striking laboratory demonstration: participants chose between products, then evaluated a product that experimenters had secretly switched. The participants not only failed to notice the switch — they justified their preferences for the switched product with explanations that bore no relationship to why they had made the original choice. The justification was constructed in real time; it was not a report of the actual decision process.

The second mechanism is social desirability. The survey context activates the impression management function that Leary and Kowalski’s (1990) research documented: people present the self-image they want the researcher to perceive. The customer who actually buys a product primarily for status reports that they bought it for environmental reasons, because the environmental motivation is more consistent with the identity they want to project. The gap is not dishonesty; it is the automatic social performance that the research context activates, operating below the threshold of awareness.

The 1.5x finding and its commercial implications

Behavioural Science Insights’ study comparing stated and revealed preference methodologies for predicting weekly supermarket sales found that revealed preference techniques outperformed the best-performing stated preference technique by a factor of 1.5 across commodity categories. The gap between stated and revealed preference is not merely philosophical. It produces measurably worse commercial predictions when stated preference data is treated as a reliable predictor of actual purchase behaviour.

The electric vehicle case makes the motivational divergence concrete. Over 70% of EV buyers in the UK in 2021 were motivated primarily by status rather than the rational cost savings of electricity over petrol. The stated preference surveys had predicted an environmental and economic rationale; the revealed behaviour showed a social status motivation — a fundamentally different psychological driver that requires entirely different messaging, positioning, and product feature prioritisation. The survey captured what buyers were comfortable saying; the revealed behaviour captured what was actually driving the purchase.

When stated preference is necessary and when to reject it

The research boundary conditions are specific. Stated preference is necessary in two situations: when the product does not yet exist (you cannot observe revealed preference toward something that is not available), and when you need to understand the motivation behind behaviour that revealed preference data has already captured. What people actually buy tells you what they choose; understanding why requires a different research method.

Revealed preference is superior in every other situation: predicting how existing products will perform, understanding actual willingness to pay, evaluating feature priorities in an existing product, and testing messaging. A/B testing, click-through analysis, real purchase pilot programmes, and waitlist commitment are more reliable substitutes for stated preference surveys in these contexts because they observe actual choice under real constraints — including the financial constraint that hypothetical willingness-to-pay questions systematically ignore.

The Segway launch demonstrates the distinction. Pre-launch stated preference surveys showed strong consumer interest in the product concept. The revealed behaviour that followed launch showed that the practical constraints — pavement legality, storage, range, social context of use — that survey respondents had not registered as relevant made the product commercially non-viable for the urban commuter use case it was positioned toward. The surveys captured concept appeal under frictionless hypothetical conditions; the revealed behaviour captured the actual constraint environment in which purchase decisions are made.

The Mom Test as a practical methodology for the reveal-preference problem

Fitzpatrick’s (2013) The Mom Test reoriented early-stage customer research around the same insight that revealed preference theory establishes. Customers are too socially motivated to tell you your product is bad, which means direct product questions produce stated preference data contaminated by the same approval and politeness norms that survey research has always been vulnerable to.

The methodological shift is from stated preference questions — “would you use this?”, “how much would you pay?” — to behavioural revelation questions: “tell me about the last time you tried to solve this problem”, “what did you actually do when you encountered this situation?”, “what have you already tried?” These questions access the revealed preference data that the customer’s actual behaviour has already generated, rather than asking the customer to simulate a hypothetical future choice that will be confounded by all the social and introspective limitations that stated preference research is subject to.

What the combination looks like in practice

The research supports a specific sequencing: use revealed preference wherever it is available — existing behavioural data, purchase records, A/B test results, pilot programme outcomes — as the primary source of preference data. Use stated preference methods for concept testing before products exist and for motivation research after behaviour has been observed. Weight revealed preference evidence substantially more heavily than stated preference evidence when the two diverge.

The customer who says they would pay £50 for a product and then buys the £30 competitor is giving you two pieces of data. The stated preference is the £50; the revealed preference is the competitor purchase. The revealed preference is the one that predicts next quarter’s sales.

Books worth reading on this

The Mom Test by Rob Fitzpatrick is the most directly applicable entrepreneurial methodology for getting past stated preferences to revealed behaviour in early-stage customer research — the specific conversational techniques for reorienting customer conversations from hypothetical preference to behavioural revelation. For the entrepreneur who is conducting customer discovery and wants the most practically structured available method for accessing the information that stated preference questions systematically fail to produce, this book provides the most concrete and immediately applicable framework.

If the dynamics described here are significantly affecting your wellbeing, speaking with a psychologist is the right next step. UK: Samaritans (116 123, free, 24/7). Mind (0300 123 3393). BACP: bacp.co.uk/search/Therapists. Crisis Text Line — text HOME to 741741 (US, UK, Canada, Ireland). International: internationaltherapistdirectory.com.

This article is for educational and informational purposes only. Sources: Samuelson, P.A. (1938), A Note on the Pure Theory of Consumer’s Behaviour, Economica, 5(17), 61–71. Nisbett, R.E. & Wilson, T.D. (1977), Telling More Than We Can Know: Verbal Reports on Mental Processes, Psychological Review, 84(3), 231–259. Johansson, P., Hall, L., Sikström, S. & Olsson, A. (2005), Failure to Detect Mismatches Between Intention and Outcome in a Simple Decision Task, Science, 310(5745), 116–119. Sheeran, P. (2002), Intention-Behaviour Relations: A Conceptual and Empirical Review, European Review of Social Psychology, 12(1), 1–36. Fitzpatrick, R. (2013), The Mom Test, CreateSpace. Ariely, D. (2008), Predictably Irrational, HarperCollins. Kahneman, D. (2011), Thinking, Fast and Slow, Farrar, Straus and Giroux. Eyal, N. (2014), Hooked, Portfolio. Sutherland, R. (2019), Alchemy, WH Allen.