Your market research is lying to you. Politely.

Ask a physician in a market research interview whether clinical evidence is the most important driver in their prescribing decision, and you'll get a confident yes. Watch what they actually prescribe, and a different picture emerges. This isn't because the physician is dishonest. It's because the question was wrong.

The gap between stated and revealed

Most pharma market research asks people to describe their own behaviour. People aren't very good at this. Not because they're unreliable — because a lot of behaviour is genuinely not accessible to introspection. A physician who starts a patient on Drug A rather than Drug B does so for a bundle of reasons: clinical, habitual, reputational, emotional, contextual. Asked to explain the choice, they'll construct a plausible story. The story is usually the clinical one, because that's the story they're trained to tell.

The clinical story isn't always wrong. But it's rarely the full picture, and it's rarely the part that's actually movable.

What to do instead

1. Ask about the specific patient, not the general case

"How do you decide between Drug A and Drug B?" invites a textbook answer. "Tell me about the last patient you started on Drug A — walk me through what happened in that consultation" invites a real one. The specific case is messier and harder to summarise. That's because it's true.

2. Compare stated and revealed in the same study

Ask physicians what drives prescribing. Then look at actual prescribing data from the same physicians. Where the two diverge is where the interesting insight is. The gap usually tells you what the physician believes about themselves, which is not the same as what's actually happening.

3. Watch the friction, not the preference

Stated preference research is cheap, so we run a lot of it. But behaviour is usually more shaped by friction than by preference. Which formulary tier? Which prior authorisation form? How many extra clicks in the EMR? How long does the patient counselling take? Preference tells you where the physician would like to end up. Friction tells you where they actually will.

The uncomfortable implication

If stated preference is a poor predictor of behaviour — and most of the evidence suggests it is — then a lot of the insight work that drives brand strategy is built on shaky ground. Not wrong, exactly. Just partial, in ways the team usually doesn't notice.

The fix isn't to throw out market research. It's to triangulate stated preference with revealed behaviour, to watch what people do as carefully as we listen to what they say, and to treat the gap between the two as the most interesting part of the data.

Behavioural research and insight