A consumer survey finds that 73% of respondents say sustainability is a top consideration in their purchase decisions. Sales data for the more sustainable, moderately more expensive option in that same category tells a very different story. This gap is one of the most consistently observed patterns in consumer research, and it isn't primarily explained by people lying on surveys — it reflects a genuine, well-understood difference between what stated preference and revealed preference research actually measure.
Two different methods, two different things being measured
Stated preference research asks people directly — through surveys, interviews, or conjoint exercises — what they want, value, or would choose under described circumstances. Revealed preference research instead infers preference from what people actually do when facing a real decision with real consequences and real trade-offs, typically using actual purchase or behavioral data. These aren't two ways of measuring the same thing with different levels of noise — they measure genuinely different psychological processes, which is exactly why they frequently produce different answers.
Why stated preference specifically drifts from actual behavior
Answering a survey question about what matters to you carries no real cost or trade-off — a respondent can say sustainability matters most without having to actually give up anything to say so, which makes stated answers vulnerable to social desirability bias (answering in the way that reflects well on the respondent) and hypothetical bias (people are systematically less accurate predicting their own behavior in a hypothetical scenario than in a real one with real stakes). A real purchase decision forces an actual trade-off — sustainability against a real price difference, at a real moment of parting with real money — and that trade-off often resolves differently than the same trade-off imagined abstractly in a survey.
This isn't a reason to distrust stated preference research generally
Stated preference methods are often the only option available when the relevant choice doesn't exist yet to observe — a genuinely new product concept, a pricing tier that hasn't launched, a feature that hasn't been built. In these cases, there's no revealed preference data to draw from, because the actual behavior in question hasn't happened yet and can't be observed. The error isn't relying on stated preference in these situations — it's treating stated preference results with the same confidence you'd give revealed preference data when a real behavioral alternative exists and could have been measured instead.
Where each method actually fits
Revealed preference is more reliable whenever the relevant real-world behavior can actually be observed — existing purchase patterns, actual usage data, real A/B test results from a live pricing or feature change. Stated preference remains valuable for exploring reactions to genuinely new concepts, understanding stated reasoning and motivation behind choices (which revealed preference data alone can't explain), and situations where observing real behavior isn't yet possible. The strongest consumer research designs typically triangulate — using stated preference to understand motivation and explore new concepts, and revealed preference wherever real behavioral data is available to validate or correct what the stated data suggested.
What this means for how consumer research gets designed and interpreted
- Treat a gap between stated preference and actual sales or usage data as expected and informative, not as evidence respondents were being dishonest
- Prefer revealed preference data whenever a real behavioral alternative exists and can be measured
- Use stated preference deliberately for what it's actually good at — new concepts, motivation, and reasoning — rather than as a general-purpose substitute for behavioral data
- Design purchase-intent research, wherever possible, to include a real trade-off (a real price shown, a real commitment asked for) rather than a costless hypothetical, since this narrows the gap between stated and revealed results
Neither method is more "honest" than the other — they're answering different questions, and the practical skill is knowing which one the decision in front of you actually requires.