Special Session Summary Branding Metrics: New Methods For Eliciting Consumer Brand Associations

Deborah Roedder John, University of Minnesota
Barbara Loken, University of Minnesota
[ to cite ]:
Deborah Roedder John and Barbara Loken (2002) ,"Special Session Summary Branding Metrics: New Methods For Eliciting Consumer Brand Associations", in NA - Advances in Consumer Research Volume 29, eds. Susan M. Broniarczyk and Kent Nakamoto, Valdosta, GA : Association for Consumer Research, Pages: 395-396.

Advances in Consumer Research Volume 29, 2002     Pages 395-396

SPECIAL SESSION SUMMARY

BRANDING METRICS: NEW METHODS FOR ELICITING CONSUMER BRAND ASSOCIATIONS

Deborah Roedder John, University of Minnesota

Barbara Loken, University of Minnesota

SESSION OVERVIEW

Understanding brand equity involves identifying the network of strong, favorable, and unique brand associations in consumer memory. Consumers might associate a brand with a particular attribute or feature, usage situation, product spokesperson, or logo. The focus of this session was on new methodologies for eliciting brand associations and brand association structures from consumers. The first two papers presented methods for uncovering the structure and interconnections between brand associations. The third paper presented a method for eliciting brand associations among young consumers.

The first paper, presented by Debbie Roedder John, reported on the development and testing of a technique to uncover brand association structures called BCM (Brand Concept Maps). This technique involves eliciting the interconnections between brand beliefs from consumers directly, instead of inferring what the connections might be between beliefs from qualitative elicitations or analytical procedures. Debbie reported on the application of this technique in a recent branding study with patients and non-patients of the Mayo Clinic.

The second paper, presented by Gerri Henderson, followed up on this theme of brand association networks, using a different qualitative technique (repertory grids) and a quantitative analysis technique (network analysis). Associative network structures for brands were modeled in the context of a branding experiment where subjects saw an advertisement for one of three types of brand extensions for a popular car brand. Results confirmed the usefulness of network analysis for identifying brand association structure and for assessing changes in brand associations as a result of brand extension activity.

The final paper, presented by Lan Nguyen, continued the theme of uncovering brand associations by presenting a new technique that can be used successfully with children. The technique, called the "Brand Collage," incorporates the idea of using collages to understand brand associations with a more concrete and structured task amenable to working with children. Children are asked to select a brand and make a collage about it using a set of interesting stimuli, such as pictures of hobbies and sports. For example, for hobbies, children are given pictures of Pokemon cards, Barbie dolls, musical instruments, and Star Wars paraphernalia. Lan presented the results of a study using the "Brand Collage" technique conducted with children from 3rd to 8th grade.

The discussion of all three papers was lead by Dawn Iacobucci. Dawn discussed the use of network analysis to study many of the interesting issues in branding structures. Questions from the audience precipitated a discussion of several additional topics, including the difficulty of aggregating individual brand association into a composite map or network of associations and the advantages and disadvantages of using collages versus other techniques for measuring children’s brand associations.

 

"IDENTIFYING BRAND STRUCTURES: A CONCEPT MAPPING APPROACH"

Deborah Roedder John, University of Minnesota

Barbara Loken, University of Minnesota

Kyeong-Heui Kim, University of Minnesota

Alokparna Basu Monga, University of Minnesota

Understanding brand equity involves identifying the network of strong, favorable, and unique brand associations in consumer memory (Keller, 1998). To aid in this analysis, a variety of methodologies have been used to capture the most important brand associations, including traditional free-association techniques (e.g., elicitation of beliefs), qualitative methods (e.g., focus groups, collages), and inferential quantitative techniques (e.g., MDS). Clearly, these techniques have been useful in understanding how consumers view brands, identifying brand strengths, and suggesting opportunities to leverage the brand into other product categories.

However, less progress has been made in developing methodologies for identifying the rich interconnections that exist between these brand associations. Most conceptualizations of brands assume a cognitive network structure, in which brand associations or beliefs are nodes that have direct linkages to the brand and may have additional linkages to other nodes (e.g., spreading activation model of belief structure, cf. Anderson 1983). Modeling these interconnections has proven difficult, although some recent advances in causal mapping techniques (e.g., Sirsi, Ward, and Reingen 1996), network analysis (Henderson, Iacobucci, and Calder 1998), and qualitative elicitation techniques (Zaltman and Coulter 1995) show promise for the branding area.

In this paper, we report on the development and application of a technique to uncover brand structures called BCM (Brand Concept Maps). This technique involves eliciting the interconnected brand belief structure from consumers directly, instead of inferring what the connections might be between brand beliefs from qualitative elicitations or analytical procedures. Eliciting the brand structure or map directly from consumers has several important advantages. Not only can we identify which brands beliefs are important to the consumer, but we can also identify interconnections between beliefs and uncover the intensity with which these connections are held. Additionally, we can aggregate individual consumer maps to uncover a composite brand structure map, which includes important core beliefs, their interconnections, and the intensity of these connections. While hypothetical examples of such brand maps are commonly included in textbooks and trade books on branding, there is no well-developed technique at this time for generating such maps.

Our presentation prvides an overview of the BCM technique and its application to a real branding situation. Specifically, we report on results using our technique for the Mayo Clinic brand name. Data was collected from Mayo Clinic patients and non-patients through one-on-one interviews during the fall of 2000. Patients were pre-recruited by phone, while non-patients were recruited via mall intercepts. A total of 165 consumers were interviewed. Using data from this sample, we discuss some of the analytical issues in developing aggregate brand maps and present the final brand maps for the Mayo Clinic brand. Using these maps, we discuss the advantages of the technique for brand management, including issues related to brand building, brand leveraging, and brand dilution.

 

"USING NETWORK ANALYSIS TO UNDERSTAND BRANDS"

Geraldine R. Henderson, Howard University

Dawn Iacobucci, University of Arizona

Bobby J. Calder, Northwestern University

Millions of marketing dollars are spent each year around the world, publicly and privately, to develop and support brand names. Nothing is more important to brand managers than the ability to measure and understand consumer brand associations, the responses that are evoked when consumers think about brands.

In this paper, we model brand associations as an associative network, consistent with recent conceptualizations of how consumers organize brand information and perceptions. Associative networks represent consumer brand knowledge as links of associations among "nodes," or units of information. Although the idea of brand associative networks is well accepted, rarely are these networks elicited and modeled explicitly. Here, we do so by using methods of network analysis to provide the analytical direction and techniques to identify the associative network structure.

To describe our technique, and its usefulness for identifying brand associative networks, we present the results of a branding experiment. We elicited brand associations about "Porsche" using a qualitative method called the repertory grid. Participants began by naming a number of salient car brands. Groups of three brands were compared at a time, using a procedure called triadic elicitation (Shaw 1981), where respondents are asked in what way two brands are alike and the third differs. This procedure produced a set of brand associations, which were then subjected to network analysis to measure three properties of network structure: (1) centrality, which measures how important each node is in the network; (2) cohesion, which groups the nodes that are most interconnected; and (3) equivalence, which groups the nodes according to their similarity vis-a-vis their connections to the other nodes.

This elicitation and analysis technique was used prior to and after subjects saw an advertisement for one of three types of brand extensions for Porsche. "Model X" was described as an economy car (low price and good mileage), a family car (trunk space, child car seats, four-wheel drive), or a typical Porsche ("everything you ever thought a Porsche to be"), which served as a control group. Network maps were constructed for the "before" condition to understand the brand associations and linkages for Porsche. These maps were compared to the "after" condition to assess any changes in brand structure for the first two types of brand extension and to examine stability of the maps for the third type of brand extension.

Results confirmed the usefulness of network analysis for identifying brand association structures. Measures of centrality identified the most important associations, while measures of cohesion and equivalence identified several changes after subjects were exposed to the first and second brand extensions. Additionally, the results point to the relative stability of these brand association networks. We discuss these results, and the network measures employed here, in terms of important branding issues. In particular, we discuss how the different measures used in this study can provide direction for issuesrelated to brand positioning, co-branding, and brand dilution.

 

"ASSESSING BRAND EQUITY IN CHILDREN: A METHOD FOR ELICITING BRAND ASSOCIATIONS"

Lan T. Nguyen, University of Minnesota

Deborah Roedder John, University of Minnesota

Current research in branding stresses the need to understand the associations and beliefs that consumers associate with brand names (Keller 1993; Aaker 1996). To answer this call, researchers have identified a number of techniques that are helpful in eliciting consumers’ brand associations, including a variety of quantitative and qualitative techniques (Keller 1998).

Despite the success of these techniques, they are less than satisfactory when used for interviewing children. Many of the qualitative techniques require a level of abstraction and verbal articulation that is rare for young children. It is not uncommon for young children to respond to direct questioning about a brand in terms such as "I just like it!" Adolescent consumers, who are old enough to articulate feelings, often say not much more than a brand is "cool." Even methods that rely less on verbalization can be problematic. For example, asking children to draw pictures of a brand often results in a faithful rendering of the package or container, which fails to reveal the types of underlying brand associations they have.

In this paper, we present a qualitative technique for eliciting brand associations with children that alleviates many of these difficulties. The technique, which we call the "Brand Collage," incorporates the idea of using collages to understand brand associations (see Zaltman and Coulter 1995) with a more concrete and structured task amenable to working with children. Asking children to construct collages for a brand or product category can be a demanding task, especially for younger children. Cutting out pictures, words, or drawing visual images holds little interest for some children, as does the process of constructing the collage from these elements. Further, the time involved in constructing such collages is well beyond the attention span and interest level of many young children.

Our technique involves asking children to select a brand name and imagine that the brand is a person. We then instruct children to make a collage representing "what the brand name would be like if it were a person.". Instead of requiring children to find elements (pictures, words, concepts) to place on the collage, they are given a set of words/pictures representing different personality traits, personal descriptions, hobbies, sports, and TV show characters. For example, for the "personal description" category, children are given choices such as "kid," "teenager," "girl," and "boy." The choices for all five categories are mounted on five Post-It boards, which allow children to simply pick off the chosen pictures for their collage. The collage board is also a Post-It board, which allows children to easily put their words and pictures on the collage, move them around, or delete them as their collage develops.

We present the results of a study using the "Brand Collage" technique recently conducted with children from 3rd to 8th grade. Our experience thus far has been that the technique is well understood by children in this age group. The technique uncovers some interesting associations that children have with well-known brand names such as Abercrombie & Fitch, Nike, and McDonald’s. For example, McDonald’s is a kid (patron) or a teenager (service staff), is fun and happy, likes Barbie and Matchbox cars (because they were in Happy Meals), and likes basketball and football (because they are popular like McDonald’s). Our presentation concludes with a discussion of the benefits of the "Brand Collage" technique and some future directions for further development and testing.

----------------------------------------