ZMET (Zaltman Metaphor Elicitation Technique)

Zaltman argues that human thought is largely image-based and metaphorical, and that much of what drives behavior is non-conscious, so asking people to state their feelings directly tends to produce shallow, rationalized answers. Having participants collect their own images first gives them a way to externalize feelings they cannot easily verbalize and provides concrete, personally meaningful artifacts to discuss. In the interview, talking about a picture, why it was chosen, what it has to do with the topic, naturally elicits metaphors and associations that a direct question would not. The images are thus a bridge to deeper cognition, which is why the pre-interview image-collection step is central to the method rather than optional.

A surface metaphor is a specific figurative expression a participant uses while telling the story of an image, for example describing a brand as 'a warm blanket' or 'an uphill climb.' A deep metaphor is one of a small number of universal, largely non-conscious structures that organize human thought, such as balance, transformation, journey, container, connection, resource, and control. In ZMET the interviewer probes the many surface metaphors people produce to identify the deep metaphors beneath them, so 'warm blanket' and several other expressions might all reflect the deep metaphor of container or connection. Identifying deep metaphors is what lets idiosyncratic stories be compared across participants and assembled into a shared mental model.

Not in a statistical sense. ZMET uses small, purposive samples and intensive interpretation, so its consensus map represents the structure of shared meaning among the participants, not projectable percentages for a population. Zaltman's claim is that even modest numbers of deep interviews tend to converge on a common set of deep metaphors and constructs, so the map captures widely shared mental models, but it remains a qualitative model. Treating the map's frequencies as market shares or as precise estimates is a misuse. The appropriate role of ZMET is to generate deep, well-grounded hypotheses about meaning and emotion, which can then be validated or sized with quantitative methods if projectable numbers are required.