Journal of Marketing Research Scholarly Insights are produced in partnership with the AMA Doctoral Students SIG – a shared interest network for Marketing PhD students across the world.
Textual data analysis is a powerful tool for gleaning insights from the wealth of information generated across various consumer and market contexts. However, traditional automated approaches often struggle to fully capture the nuanced, context-dependent ways in which meaning is expressed, for example, when it comes to the use of figurative language such as metaphors.
Metaphors such as “elephant in the room” and “heart of gold” are an inherent part of our everyday language. But surprisingly, tools to account for these linguistic phenomena in communications remain scarce. To address this challenge, a recent Journal of Marketing Research study introduces a novel methodology termed “metaphor-enabled marketplace sentiment analysis” (MEMSA), which combines the breadth of automated text analysis with the depth of qualitative, sociocultural interpretations.
In the study, researchers Ignacio Luri, Hope Jensen Schau, and Bikram Ghosh demonstrate how MEMSA can leverage the sentiment-conveying potential of metaphors to uncover a deeper, more contextual understanding of market discourse. By linking metaphors to sentiments across different domains, MEMSA enables researchers to scale qualitative insights to larger, “big data” textual samples.
The authors’ innovative use of metaphor analysis within an automated text analysis framework represents a significant advancement in consumer and market research. The managerial implications of this research are far-reaching, as they hold the potential to inform marketers of more effective value proposition development, targeting, and communication strategies by accounting for the prevalence of metaphors in the marketplace, especially when it comes to consumers’ opinions and affective responses. In addition, the insights can support more informed policymaking decisions that better address the needs and concerns of consumers. This deeper insight into consumer sentiment can directly inform the development of more effective and responsive policies that cater to public needs and concerns.
The authors point out that qualitative researchers are very aware of the prevalence of metaphors in the marketplace, as there are millions of ways to say things, and the choice of every word can be strategic. Therefore, methodologies such as MEMSA help to gain a better understanding of the marketplace.
To explore this innovative approach further, we sought additional insights directly from the authors. Our conversations with the researchers shed more light on the practical applications and methodological contributions of the MEMSA framework.
Two major points from our author interview (below):
- Metaphors permeate consumer language, providing valuable insights into sentiments and attitudes.
- The study provides a detailed, step-by-step guide for developing dictionaries that capture the nuanced and elaborate usage of text, extending beyond the words’ literal meaning.
Q: How can researchers effectively leverage the sentiment potential of located metaphors and scale insights to analyze big textual data?
A: At the heart of MEMSA is the reality that we express all kinds of complex mental processes, including emotion, in the form of words. Truly understanding the meaning and associations of words is a challenge because meanings change and are so culturally bound and context-dependent. However, once we create a link between words and mental processes, we open the door to big data insights. Metaphors are the perfect linguistic pattern to exploit this association between words and mental processes because they are so rich and the association itself has been so thoroughly documented. MEMSA added a way to employ these associations to scale insights into the level of big textual data.
Q: To what extent can MEMSA be combined with traditional qualitative and quantitative methods to understand and interpret textual communications?
A: We see MEMSA as complementing, not replacing, qualitative insights while being a quantitative methodology in itself. In our opinion, the topic of human emotion is complex and inherently qualitative.
Q: What are the benefits and limitations of MEMSA compared to the traditional qualitative and quantitative approaches?
A: MEMSA inherits the active researcher role of qualitative methodologies, which can be its biggest strength and a significant barrier depending on the situation. Other quantitative approaches such as traditional sentiment analysis will continue to have a place by virtue of being out-of-the-box ready solutions that can be applied to any dataset by applying a pre-made dictionary using some software or model. In our opinion, recent technological developments will push Natural Language Processing toward being more contextually aware, blurring the line between qualitative discourse analysis and automated text analysis. Metaphors have been the classical example of being too subjective, human, and contextually dependent to be studied with big data approaches, but we hope MEMSA and future approaches will change that.
Q: What practical benefits do marketers gain from understanding nuanced market sentiments in terms of crafting, targeting, positioning, and communicating value propositions effectively?
A: Metaphors direct thinking and emotion, thus driving behavior. A data-driven understanding of metaphorical frames for a given topic can have great value for marketers. For example, any educator knows the value of metaphors in “shedding light on” a complex topic, which has implications for consumer education, completely new product introduction, or brand extensions. As our paper clearly illustrates, these metaphors can educate in a way that harnesses or redirects emotions. Think about AI anxieties and how the “copilot” metaphor has been used all over the branding of Generative AI products and services. The copilot metaphor not only educates consumers on how to use the technology but also preventively deflects concerns about human substitution and AI-driven job loss.
Think about AI anxieties and how the “copilot” metaphor has been used all over the branding of Generative AI products and services. The copilot metaphor not only educates consumers on how to use the technology but also preventively deflects concerns about human substitution and AI-driven job loss.
Q: How do you envision the potential use of metaphors in exploring psychological processes beyond affect, such as cognitive processes or behavioral intentions?
A: Critical Metaphor Analysis has long used metaphors to explore many topics beyond affect, including power dynamics, stereotypes, or learning. We think that the role of metaphors in learning and sense-making of complex topics, such as AI or climatic change will be a productive path for social researchers.
Q: Can you explain the process of developing and validating metaphor dictionaries for analyzing textual data? What are some of the main challenges that practitioners may face when performing this type of work?
A: The process shares much with the development of any other theoretical dictionary. A theoretically relevant concept is identified (a frame, ideology, attitude, belief, stance, etc.), words and expressions that serve as linguistic signposts of that concept in discourse are collected, and the word list is validated. A strength of dictionary methods is their transparency and face validity, but it is easy to forget how challenging constructing a good dictionary is, or how much back and forth between data and instrument development is required. In our paper, we provide some practical guidance for dictionary development and validation which we believe have been lacking.
Q: Apart from the consumer debt conversations discussed in the paper, in what other contexts do you anticipate metaphors being prevalent, thus offering significant strategic insights through MEMSA in understanding sentiments embedded in textual communication?
A. Consumer debt was a good fit to illustrate our methodology because it is an everyday topic but also highly abstract, complex, and emotion laden. Many important social dynamics, such as immigration or gentrification, are equally prone to metaphorical thinking, and MEMSA could provide great insights into it. A timely topic that we have dedicated our attention to recently is artificial intelligence. The very concept of AI is metaphorical by necessity; we have no other way to understand or talk about it! However, this does not stop people and companies from using a range of different metaphors to speak of AI, sometimes strategically, in ways that are ideologically and emotionally loaded.
Read the Full Study for Complete Details on the MEMSA Method
Reference: Ignacio Luri, Hope Jensen Schau, and Bikram Ghosh (2023), “Metaphor-Enabled Marketplace Sentiment Analysis,” Journal of Marketing Research, 61 (3), 496–516. doi:10.1177/00222437231191526
Manaswini Acharya is a doctoral student in marketing, Texas Tech University, USA.
Christian Arroyo Mera is a doctoral candidate in marketing, University of South Florida, USA.