Understanding Brand Awareness: Let Me Give You a C(L)Ue!

ABSTRACT - Despite the importance of brand awareness to brand choice, consumer researchers have given little attention to developing an understanding of awareness as a construct. Focusing on brand awareness in memory-based situations (i.e., where the brands must be brought to mind), this paper reports on a qualitative research project that explores how brands come to mind in a variety of choice situations. A subsequent quantitative study shows that associates identified in the qualitative research act as cues, and that brand awareness varies significantly depending on the cues that are salient.

Citation:

Stephen J.S. Holden (1993) ,"Understanding Brand Awareness: Let Me Give You a C(L)Ue!", in NA - Advances in Consumer Research Volume 20, eds. Leigh McAlister and Michael L. Rothschild, Provo, UT : Association for Consumer Research, Pages: 383-388.

Advances in Consumer Research Volume 20, 1993      Pages 383-388

UNDERSTANDING BRAND AWARENESS: LET ME GIVE YOU A C(L)UE!

Stephen J.S. Holden, Ecole Superieure des Sciences Economiques et Commerciales (ESSEC)

ABSTRACT -

Despite the importance of brand awareness to brand choice, consumer researchers have given little attention to developing an understanding of awareness as a construct. Focusing on brand awareness in memory-based situations (i.e., where the brands must be brought to mind), this paper reports on a qualitative research project that explores how brands come to mind in a variety of choice situations. A subsequent quantitative study shows that associates identified in the qualitative research act as cues, and that brand awareness varies significantly depending on the cues that are salient.

INTRODUCTION

Brand awareness is a much neglected construct and deserves considerably more attention due to its central importance in brand choice. From an empirical point of view, numerous researchers have shown that brand awareness measures are powerful predictors of consumer choice behavior (e.g., Axelrod 1968; Haley and Case 1979; Nedungadi and Hutchinson 1985). Furthermore, from a conceptual or theoretical point of view, brand awareness has been recognized as preceding and necessary to brand evaluation (Howard and Sheth 1969; Holden and Lutz 1992; Nedungadi 1990). That is, evaluation is a process of selection from the set of alternatives which are evoked.

This paper outlines some concepts for understanding brand awareness, reports research which explores and tests notions arising from that framework and concludes by noting some of the theoretical and practical implications of this expanded understanding of brand awareness.

BRAND AWARENESS AS CUED MEMORY

Brand awareness is typically measured by recall or recognition (Rossiter and Percy 1987). When talking of situations where the brands are not present (the focus of this paper), the appropriate measure is recall measured by presenting a product category and asking respondents to recall brands from that category. For situations where the brands are present, researchers typically take a recognition measure in which they present the brand name and ask respondents whether they know of the brand.

Rossiter and Percy (1987) make an important observation that measures of recognition of the brand name may not necessarily reflect the recognition process that takes place in the choice situation. They note that in the choice context, brand awareness may be mediated by recognition of one or more of a number of elements of the product, for example, the package, the colors, the brand logo, etc. Furthermore, measures of ease of recognition (Alba and Hutchinson 1987) may be more appropriate.

Similarly, one may ask the question of whether measures of brand recall by product category are reflecting the process of brand evocation in the choice process. Such measures assume that consumers' memory is organized by category, and perhaps more importantly, that choice is a process that utilizes these categories. Accordingly, researchers have viewed the category production task in which subjects generate exemplars in response to a category cue as the most relevant experimental paradigm for understanding product recall (Alba, Hutchinson and Lynch 1991). However, advances in categorization research and renewed attention to the way in which consumers actually make choices suggest that there is a need to replace (or at least modify) this paradigm.

First, the role of situation in influencing brand evocation is important. Generally, the influence of situation on choice has been considered to be mediated by an influence on the weighting of attributes in the multi-attribute model (e.g., Miller and Ginter 1979). However, research from the categorization literature (Roth and Shoben 1983) has suggested that contextual factors can change the graded structure of a category. Given the relationship between graded structure (typicality) and recall (Nedungadi and Hutchinson 1985), it might be hypothesized that contextual factors would also affect recall. Ratneshwar and Shocker (1991) have shown that situations influence the recall of products (or product variants) in the overall category of snack foods.

In addition and somewhat belatedly, consumer researchers have begun to recognize the importance of motives in consumer choice. Rossiter and Percy (1987) noted that ads should link brands to the category need in order to be effective. In a similar vein, Warshaw (1980) noted that brand attitudes are better able to predict behavior if intentions towards the product category are known.

More generally, a person's motives may be satisfied by products from different categories. Studies on categorization have distinguished between taxonomic categories (e.g., product categories) and goal-derived categories such as "things to take on a camping trip" (Barsalou 1985). One of the features of a goal-derived category is that elements may be drawn from a range of taxonomic categories. The notion that decisions may be made between alternatives which are noncomparable alternatives (e.g., Johnson 1988) implies evoked sets which comprise brands drawn from different product categories. In addition to being noncomparable, brands in an evoked set which is goal-derived may be complementary, or at least not competitive, as in "things to take on a camping trip." The evaluation task is therefore one of selecting one or more of the items depending on other constraints (e.g., financial).

Both the influences of context (or situation) and motives are captured by a model (Holden and Lutz 1992) that suggests that brand awareness is a function of the cues that are salient in the retrieval situation. The model proposes that the brand be considered to be a central node in memory with links to it from various other nodes representing situations, benefits, attributes, product category and other brands (see Holden and Lutz 1992, p. 105). Any of these nodes is posited to have the potential to act as a cue to the brand, the success of the cuing being dependent on the associative strength between the cue and the brand.

It is further posited that the links to the brand may be direct or indirect. For instance, a situation may directly evoke a particular brand while other situations may lead to thought of some benefit which in turn leads (directly) to the evocation of a brand.

Research Objectives

The research reported in this paper is a preliminary examination of some of the principles that arise from this model.

1) Examine nature of associates salient at choice

- categories or types of cues

- direct and indirect cues

2) Examine nature of evoked sets

- evidence for goal-derived categories

- basic level of recall (Rosch 1975)

3) Test influence of associates on brand awareness

- variation in evoked set formation by cue

A qualitative project focuses on the first two objectives and subobjectives while a quantitative study focuses on the third objective.

QUALITATIVE STUDY C NATURE OF CUES AND EVOKED SETS

An exploratory qualitative project was undertaken to examine the nature of brand associates salient at choice and the nature of evoked sets. The qualitatitive approach is in line with the call from a number of researchers for "discovery-oriented" studies (e.g., Lutz 1991; McGuire 1983).

Based on the theory of spreading activation (Collins and Loftus 1975), the method utilized presumes that increased activation of some node in response to a cue will be reflected in an association task or a generation task. That is, subjects asked to list thoughts in response to a given cue word will list those words corresponding to nodes which are relatively strongly associated with the cue word.

Method

Subjects saw one of four situations for each of three product-types. The three product-types (and four products for each product-type) were drinks (weekend party, in class, breakfast, studying); snacks (missed breakfast, watching TV, road-trip, after a workout) and restaurants (late night, weekday lunch, dinner and Sunday brunch). Interviewed on a one-on-one basis, subjects considered all three product-types in a random order. They were provided with one, randomly selected situation cue for each of the three product-types. A total of 15 University of Florida students from the undergraduate marketing fundamentals course were interviewed. The sample comprised eleven females and four males. All subjects considered one of four situations for each of the three product-types. The interview was recorded to ensure "top-of-mind" responses were captured.

In each interview, situation associates were measured first. Subjects were exposed to three situations and were asked to list whatever thoughts came to mind to each. A second measure, similar to benefit laddering (Reynolds and Gutman 1988) provided evocation cues. After generating situation associates, subjects were asked to reconsider each situation in turn and to generate products coming to mind within the target product-type. Following evocation for all three product-types, subjects were asked to consider the evoked sets they had generated for each situation and were questioned in more depth about why the group of brands was evoked. The subsequent probes aimed at exploring the ladder of what was presumed to be cues were similar to probes used in benefit laddering (see Reynolds and Guttman 1988).

Results and Discussion

This section is organized along the lines of the formulation of the research questions rather than the dependent measures. As the data are qualitative, verbatims from the qualitative data are used to support observations made and conclusions drawn.

Associates Salient at Choice. An illustration of the situation associates and the evocation cues generated by four people in response to one situation is provided in the Table 1. More general conclusions are made in the following remarks. It will be noted in Table 1 that brand associates (i.e., situation associates and evocation cues) have been classified into four different types: situations, benefits, attributes and products (including category labels and brands). This is the categorization suggested by Holden and Lutz (1992). It should be noted that the classification of some responses was difficult suggesting a need for further research and specification of cue types.

Overall, the data gave strong evidence for the existence of direct and indirect cues. Provision of the situation cue alone (i.e., without any product-type prompts) led to numerous mentions of brands and products. For example "breakfast" elicited mentions of brands and/or products from a number of subjects: "Carnation Instant Breakfast," "Special K and bananas," "toast or Pop-Tarts, or dishing out money for Krystals," and "coffee, newspaper." Simililarly, "weekend party" elicited numerous mentions of "beer." All situations describing a meal such as breakfast, dinner and lunch tended to be more likely to directly elicit brands of restaurants and drinks (e.g., Coke).

As might be expected given the notion of a direct cue, closer examination of situations as direct cues revealed that subjects seemed to be unable to articulate the links between the situation and the products or brands brought to mind. For instance, one subject who was asked to imagine after a workout responded that the first thing that came to her mind was "foodCI wanna eat." When probed on why that was the case, she responded "I don't know. I think it's just a mental thing," and later suggested that she ate after working out "because [of] habit."

Evidence for situations acting as indirect cues came from two sources. First, some responses represented "ladders" of associates from the situation cue to benefits to brands: "After working out? Like lots ofCrest. Eating the right foods, with the right vitamins, right nourishment and stuff. Always have like some proteinCmeat like chicken or beef, with rice for starch, and always take some fruit afterwards. And I would always take Amino 1000s which is just lots of protein." This subject appears to be progressing from the situation cue, through a number of benefit associates (e.g., healthy) and attribute associates (e.g., vitamins, protein and starch) to finally reach specific products (e.g., chicken, beef or rice).

A second, perhaps more compelling form of evidence for indirect cuing from situation cues is the way in which many subjects were able to clearly take the situation cue, and through the provision of self-generated subcues, provide different evoked sets for sets of subcues which could be considered more direct cues. For instance, one subject's response after being asked what snacks came to mind in a situation after a workout responded as follows: "If I am in a healthy mood, I'll have a yoghurt, or fruit or something. If I'm in a bad mood, chocolate. Chocolate and Coke." In this case, the motive or benefit sought appears to be something healthy and something uplifting respectively.

A somewhat surprising observation was that some cues tended to directly evoke brands or products even though those products or brands were not preferred or even liked. The result is surprising as it runs counter to the generally accepted notion that more preferred brands are recalled earlier (e.g., Nedungadi and Hutchinson 1985). One subject asked to list what drinks came to mind when she thought of weekend parties mentioned a brand she liked, and then, in the same response, evoked a product at the same time as she indicated she did not like the product: "Bartles and James. At this point, I don't like beer." Another subject cued with weekday lunch (and no product-type) responded "Cafeteria food. Yuck." It seems in both cases, that these are alternatives that are very closely associated with the situation even though the products themselves are negatively evaluated.

Nature of Evoked Sets. It was clear that the evoked sets generated had a number of characteristics of goal-derived categories. For example, one subject asked for products that came to mind when thinking about studying volunteered "Colored pens. I always used colored pens a lot to try and keep me awake.... Other things that I think of is A+ notes.... I think of sitting down with diet Coke and pretzels." This list might be designated as those things necessary to study. In terms of the more direct cue, the colored pens, diet Coke and pretzels were all identified by the subject (without prompting) as assisting in the objective of "staying awake."

TABLE 1

SITUATION ASSOCIATES AND BRAND CUES-AFTER A WORKOUT

In addition, it may be noted in the above example, and in other responses provided by subjects, that they were comfortable mixing mentions of brands and products. Another subject listing what came to mind when she thought of breakfast mentioned "Raisin Bran, bananas, toast, hash browns and chocolate milk."

This mixing of the mentions of brands and products reflects Rosch's (1975) notion of "basic levels" of categorization. However, it also represents the limit of categorization for some subjects. At least some subjects made the comment that "I don't know the brands of things." It seems plausible that this lack of brand knowledge may reflect that subcategorization and branding is not particularly important to consumersCthat is, a banana is adequate whether it is Chiquita or some unknown brand. On the other hand, brands like Coke enjoyed very high recall being mentioned virtually whenever drinks were under discussion.

It was noted that subjects did not always use the category labels that might be given to a category by a marketer. One example was that some subjects mentioned Coke; however, their subsequent comments and responses indicated that they were referring to the category of sodas. A similar example was the mention of Pop-Tarts which sometimes referred to the Kellogg's brand, and sometimes to the category of toaster pastries.

When discussing snacks, many mentioned the category of "junk food." The attributes of the category were not altogether clear. Despite this, subjects did not generally appear to feel a need to articulate the nature of the category implying that there was a shared understanding about the nature of junk food. Similarly, many restaurants were grouped as fast-food, which in the words of one subject, is where "you can go through a drive-through."

TABLE 2

CHI-SQUARED ANALYSIS OF EVOCATION BY CUE

Finally, some subjects indicated that they were not sure what to call some categories of products. One subject recalling brands of potato chips said "Planters for their kind of things, but not for potato chips." Perhaps if forced on the issue he could have given a label to the category of "their kind of things," but clearly it was not a label that was readily accessible.

QUANTITATIVE STUDY C CUING INFLUENCES ON BRAND AWARENESS

The quantitative study tested the hypothesized cuing influence of salient associates on brand evocation. The method and results are summarized here; some of the elements of the design and method that are not relevant to this paper have been omitted (see Holden 1992 for more complete details).

Method

Utilizing a small selection of the cues found in the qualitative pretest, the brand evocation study examined the hypothesis proposing that evoked sets vary as a function of the associates provided at retrieval. Subjects were asked to list brands coming to mind for one of five cues for each of the three product-types (drinks, snacks and restaurants). The five cues, nested in product-type, were three benefit cues and two situation cues identified in the qualitative pretest: drinks - breakfast, party, relaxing, thirst quenching, socializing; snacks - watching TV, mid-afternoon, filling, limited time, nutritious; restaurants - Sunday lunch, with family, different, healthy, romantic.

The brand evocation measure was a questionnaire requiring subjects to generate three evoked sets. Subjects generated an evoked set for one of the five cues associated with each of the three product-types. The order of the product-types in each questionnaire was randomized. Subjects were assigned to cues on a random basis with the constraint that the cues for each product-type were of the same type. That is, if one of the cues for one of the product-types was a situation cue, then the other two cues selected for the other two product-types were also situation cues.

The brand evocation measure was administered to four large groups (classes) of undergraduate students in marketing to give a total sample of 116 students. Subjects were allowed 20 seconds to generate each evoked set, and were then required to move on to generate the next. The brand evocation measure also included a measure of indirect recall: after generating an evoked set for each cue ("direct recall"), subjects were instructed to return to each cue and to list up to three brands for each mention of a product or product category rather than a brand ("indirect recall"). There was no time limit imposed on subjects for completion of the indirect recall task.

Results and Discussion

All brands and products evoked were coded and entered as data into the analysis. As was expected, there was a mixture of products and brands in subjects' recall, and this was preserved in the coding of the data. That is, evocation at the "basic level" (Rosch 1975) forms the level at which analysis was conducted.

In an analysis similar to that conducted by Ratneshwar and Shocker (1991), a chi-squared analysis of the frequency of recall of the five most frequently recalled brands (products) for each of the five cues for each product-type was conducted. The analysis revealed that there was a significant difference in evocation by cue for all three product-types (see Table 2).

In addition to the overall analysis, a comparison of each cue with each other cue within each product-type revealed that there were significant differences in recall between every pair of cues except three (see Table 2). The recall of drinks in response to "socializing" was not significantly different from recall cued by "party." The recall of snacks in response to "filling" and "limited time" were significantly different from each other, but neither of the evoked sets generated for these two cues was significantly different from recall to the cue, "mid-afternoon."

The lack of differences in recall in the three cases described are perhaps not surprising. Given the student population from which the sample was drawn, the benefit of socializing and the situation of party are likely to be very closely related. Indeed, it implies that at a party, a major benefit sought is socializing. Perhaps the result seen in snacks can be explained in a similar manner. The situation of wanting a snack in the middle of the afternoon is perhaps associated with limited time and/or a desire for something filling, and hence, recall of brands to the cue, mid-afternoon, is not significantly different from the recall cued with filling or limited time even though the two latter cues lead to significantly different recall.

Clearly, the recall data support the findings of Ratneshwar and Shocker (1991) and extends their finding by showing that recall varies by situation cue across a range of product-types. In addition, the data provide clear support for the notion that evoked sets vary by the cuing benefit. Interestingly, the only non-significant differences in recall appeared in the comparison of situation cues with benefit cues. One interpretation of these cases is that the situation cues are operating indirectly via the benefit cues; evocation would not be expected to be significantly different between benefit cues and situation cues which evoke those same benefits. That is, the situation operates as an indirect cue and the benefit operates as a more direct cue.

Given the basic result showing support for the model of brand awareness being a function of the cues present at brand retrieval, it might be possible to extend the present research to alternative measures of brand awareness. As the basic construct in the model is one of associative strength, the model would predict that measures of brand accessibility based on response time (see Fazio 1990) should be related to brand accessibility. Preliminary results (Holden 1992) supporting this contention are reported briefly in the following.

Subjects in a response time experiment were asked to state whether each word appearing on a computer screen was a real word or a nonsense word (i.e., a lexical decision task). The major manipulation was the word-prime (cue) that immediately preceded the presentation of the target brands with the hypothesis being that stronger associative strength between the cue and the brand being reflected in faster response times for identification of the brand. The cues were a selected from the set used and reported here in the brand evocation measure. Comparing the average response times for brand identification with the frequency of brand evocation in response to the same cue, a strong correlation was found. The relationship between the two measures supports the notion of brand awareness as a function of the accessibility of the brand in response to the cue.

CONCLUSIONS AND IMPLICATIONS

The qualitative research found brand associates of various types that appeared to be operating as cues to brand retrieval. It was possible to classify cues as suggested by Holden and Lutz (1992), but some borderline cases suggested a need for clarifying and perhaps exploring further the types of brand associates that might operate as cues at brand retrieval.

Evidence for direct and indirect cuing was found in the qualitative research. However, when an indirect measure of cuing was incorporated into the brand evocation measure, it did not increase the relationship between brand evocation and brand verification. An improvement in the relationship would be expected given that the verification measure implicitly measures both direct and indirect cuing.

Both the qualitative research and the brand evocation measure found evidence to support the notion that evoked sets are best described as goal-derived categories where the basic level of recall might be a brand and/or a product category. The elements within the sets were often from different product (taxonomic) categories, and in the qualitative research at least, it was clear that the items included were often complements rather than substitutes.

The analysis of brand evocation clearly showed an influence of cues on evoked set formation. In addition, the probability of evocation was found to be highly correlated with the average response time for identification of the brand when presented after the corresponding cue.

The results have a number of interesting implications for marketers. First, the measurement of brand recall in response to a product category cue tends to ignore the significant differences in evocation across situations and across benefits. Marketers targeting a particular usage segment (Srivistava 1980) or benefit segment would get a more appropriate measure of their "awareness" relative to others by measuring evocation in response to the appropriate situational or benefit cues.

Second, the consideration of brands (and product categories) appears to be potentially manipulable if communication strategies take into account the cues that facilitate retrieval of the marketer's brand. By placing cues known to facilitate retrieval in the choice situation, the probability of brand evocation is increased, and thereby increases the probability of choice of the brandCwithout evocation, the probability of choice is zero.

Finally, and more subtly, a marketer could identify cues that are salient in the choice situations in order to link the brand to those cues in marketing communications. To the extent that the marketer is successful in modifying the consumer's memory, salience of the cues in the choice situation will lead to the marketer's brand coming to mind. As this is a memory effect, it seems likely that it will be at least somewhat enduring. Hence, given that the probability of brand choice is likely to increase as a function of the increase in the probability of evocation (Nedungadi 1990), this notion suggests a perspective on brand equity that derives from brand awareness.

REFERENCES

Alba, Joseph W. and J. Wesley Hutchinson (1987), "Dimensions of Consumer Expertise," Journal of Consumer Research, 13 (Mar), 411-454.

Alba, Joseph W., J. Wesley Hutchinson and John G. Lynch, Jr.(1991), "Memory and Decision Making," in Handbook of Consumer Theory and Research, eds. H.H. Kassarjian and T.S. Robertson, Englewood Cliffs, NJ: Prentice Hall, 1-49.

Barsalou, Lawrence W. (1985), "Ideals, Central Tendency and Frequency of Instantiation as Determinants of Graded Structure in Categories," Journal of Experimental Psychology: Learning, Memory and Cognition, 11 (4), 629-653.

Fazio, Russell H. (1990),"A Practical Guide to the Use of Response Latency In Social Psychological Research," Research Methods in Personality and Social Pscyhology, 11, 74-97.

Haley, Russell I. and Peter B Case (1979), "Testing Thirteen Attitude Scales for Agreeement and Brand Discrimination," Journal of Marketing, 43 (Fall), 20-32.

Holden, Stephen J.S. and Richard J. Lutz (1992), "Ask Not What the Brand Can Evoke; Ask What Can Evoke the Brand," Advances in Consumer Research, 19, 101-107.

Holden, Stephen J.S. (1992), "Brand Equity Through Brand Awareness: Measuring and Managing Brand Retrieval," Doctoral dissertation, University of Florida.

Howard, John A. and Jagdish N. Sheth (1969), The Theory of Buyer Behavior, New York, NY: John Wiley and Sons.

Johnson, Michael D. (1988), "Comparability and Hierarchical Processing in Multialternative Choice," Journal of Consumer Research, 15 (Dec), 303-314.

Lutz, Richard J. (1991), "Editorial," Journal of Consumer Research, 17, March.

McGuire, William J. (1983), "A Contextualist Theory of Knowledge: Its Implications for Innovation and Reform in Psychological Research," Advances in Experimental Social Psychology, 16, 1-47.

Miller, Kenneth E. and James L. Ginter (1979), "An Investigation of Situational Variation in Brand Choice Behavior," Journal of Marketing Research, 16, February, 111-123.

Nedungadi, Prakash (1990), "Recall and Consideration Sets: Influencing Choice Without Altering Brand Evaluations," Journal of Consumer Research, 17 (Dec), 263-276.

Nedungadi, Prakash and J. Wesley Hutchinson (1985), "The Prototypicality of Brands: Relationships with Brand Awareness, Preference and Usage," Advances in Consumer Research, 12, 498-503.

Ratneshwar, S. and Allan D. Shocker (1991), "Substitution in Use and the Role of Usage Context in Product Category Structures," Journal of Marketing Research, 28 (Aug), 281-295.

Reynolds, Thomas J. and Jonathan Gutman (1988), "Laddering Theory, Method, Analysis, and Interpretation," Journal of Advertising Research, 28 (Feb), 11-31.

Rosch, Eleanor (1975), "Cognitive Prepresentations of Semantic Categories," Journal of Experimental Psychology, 104, 192-233.

Rossiter, John R. and Larry Percy (1987), Advertising and Promotion Management, New York, NY: McGraw Hill.

Srivistava, Rajendra K. 1980), "Usage-Situational Influences on Perceptions of Product Markets: Response Homogeneity and its Implications for Consumer Research," Advances in Consumer Research, 7, 644-649.

Warshaw, Paul R. (1980), "Predicting Purchase and Other Behaviors from General and Contextually Specific Intentions," Journal of Marketing Research, 17 (Feb), 26-33.

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