There are themes, experiences and concepts that words cannot capture and express as instinctively and vividly as visual images. Verbal probing, alternately, may not be enough to invoke respectively what goes in the minds of consumers (e.g., thoughts and feelings, beliefs and associations). However, the combination of visual images, especially photographs, with the words of consumers for explicating their associated thoughts and feelings can generate the meaningful input sought. Images can help to trigger visceral, automatic and associative responses. Hence, visual images may be employed to elicit the verbal expressions of consumers’ thoughts and feelings, perceptions and beliefs, in their own words, with regard to a topic, concept or experience.
Uses of photos or other pictorial images for elicitation may differ on some important dimensions. Two primary dimensions are: (1) using images produced or selected for display by the researchers or images produced and presented by the consumers participating in a study; (2) images that explicitly depict objects (e.g., products) or scenes (e.g., store interior spaces) as the topic of interest vis-à-vis any images that may be implicitly related in consumers’ minds to the topic of interest, concrete or more abstract (e.g., a brand, an experience, self image). In the latter case, a visual image is used as a metaphor because it represents in a different form (e.g., borrowed from another domain or context) an attribute or propensity associated with the topical object or concept.
A well-known and appraised methodology in this field is Zaltman Metaphor Elicitation Technique (ZMET) developed by professor of marketing Gerald Zaltman (Harvard Business School). In ZMET participants bring to interviews photos and other images of their own creation, whose visual content is connected with or reflects on the topical target of research in their minds. The images are employed by the interviewer to hold a conversation and elicit the associated thoughts and feelings of each participant in his or her words, encouraging respondents to open-up through the visual images. The data collection and analytic approach of ZMET is qualitative.
Steven Naert, Global Solutions Leader at Ipsos, presented recently a new approach developed in the research firm, a revised quantitative adaptation of visual metaphor elicitation (“Pictures Speak Louder Than Words“, GreenBook (MR Association): Blog, 1 October 2020 [*]). The methodology ties the utilisation of photo images as metaphors for elicitation of verbal responses with advanced analytic methods that can be applied to larger amounts of data collected in a survey.
Key aspects of the approach described by Naert are highlighted, and commented upon, below:
A large set of pictures was constructed by the research team. The photo images were selected through a systematic approach with intention to represent various values and benefits. The pictures can be divided into two subsets: positive and negative. Respondents are first asked to choose up to five pictures that most closely express how they feel about a product; next they are asked to explain what these pictures represent for them (open-ended free-text responses).
The complete set includes 200 positive pictures and 200 negative pictures, 400 in total. It is understood, however, that a subset of more relevant pictures would be presented in a particular study (e.g., positive or negative). The advantage of using images created by participants (as in ZMET) is that the images are more likely to represent the associations most personally relevant, intuitively evoked in their minds with no external direction. Yet, using a pre-selected set of pictures, common to all participants as the source set of images to choose from, allows the researchers to match and compare responses associated with same or proximate images shared. In this research, the underlying values or benefits of the picture referred to should further match the interpretation of the picture by respondents in their evoked thoughts or feelings.
The methodology relies on elicited verbal answers-explications composed by respondents, more in line with a qualitative approach. As Naert acknowledges, responses to structured survey questions (e.g., especially based on rating scales) can be too limiting, unable to capture the genuine associations, thoughts and feelings of participants. It could probably be even more limiting than verbal answers probed without visual image elicitation. On the other hand, a survey is expected to produce a large amount of text to review, sort and evaluate. To cope with this challenge Ipsos employed text (mining) analytic tools to extract key themes (defined by Naert as “the main verbal constructs people associate with each picture”). Therefore, Ipsos introduces visual metaphor elicitation research to the new generation of analytic techniques (i.e., machine learning, Natural Language Processing). It would not provide the depth of insights that a qualitative approach is likely to produce; it may also not detect elaborate meanings of responses as a human reader-coder, but it can create new opportunities to apply and test the visual image elicitation approach with larger samples of consumers.
Nevertheless, a gap seems to persist between the original set of values and benefits underlying the pictures and the verbal explications to them provided by respondents. It remain unclear how the researchers verify that responses to a picture fit with the meaning intended (‘cross-validation’); where possible, the two types of information (pre-defined and generated by a consumer-respondent) may be augmented. Naert suggests that the researchers “review the key themes from the text analytics in conjunction with the visuals themselves and make any adjustments” (italics added) — this explanation seems to correspond to the issue raised here, but it leaves the matter rather vague (i.e., how are they actually matched).
In virtue of generating a structured information format from the original unstructured verbatim data, it is possible to apply more quantitative analytic methods on the basis of extracted themes coded (e.g., factor and cluster analysis). In particular, Naert offers the possibility of identifying segments of consumers-respondents based on the themes they share (e.g., values, needs or benefits they seek, aspirations and goals in life). This is demonstrated in an example of a study on women in Japan, South Korea and China (see the figure showing profiles of women with similar views in the Striver Segment — exemplar Values purportedly relate to the original underlying values of pictures chosen, and Expressions illustrate the responses of survey participants, apparently of matching themes).
Approaches of photo-elicitation, and visual metaphor elicitation as a more sophisticated form, create opportunities for revealing richer, deeper and more candid portrayals of associations and meanings, cognitive and affective, that consumers uphold with respect to products and services, brands and companies. The quantitative adaptation approach proposed by Ipsos to elicitation with pictures as visual metaphors adds a novel possibility for application of this research approach in marketing practice. Expanding the possibilities for using visual images in marketing and consumer research, and encouraging creative applications of them, is congratulated.
[*] A version of the article of Naert is also available as a PDF document that can be accessed on a website of Ipsos. The document includes some additional visual illustrations.