On the Relationship Between Motives and Purchase Decisions: Some Empirical Approaches

Citation:

Brian T. Ratchford and Richard Vaughn (1989) ,"On the Relationship Between Motives and Purchase Decisions: Some Empirical Approaches", in NA - Advances in Consumer Research Volume 16, eds. Thomas K. Srull, Provo, UT : Association for Consumer Research, Pages: 293-299.

Advances in Consumer Research Volume 16, 1989      Pages 293-299

ON THE RELATIONSHIP BETWEEN MOTIVES AND PURCHASE DECISIONS: SOME EMPIRICAL APPROACHES

Brian T. Ratchford, State University of New York at Buffalo

Richard Vaughn, Cone and Belding Communications, Inc.

The primary focus of this paper is on extending the general framework of the FCB Grid (Vaughn 1980, 1986, Ratchford 1987) to the study of imagery and emotion associated with brands. Specifically, we discuss two techniques, VIP and ICON, which have recently been developed by FCB (Foote, Cone and Belding) for determining brand imagery and emotion; an important feature of these techniques is that they are primarily nonverbal. We attempt to establish a link between these techniques and the grid model, and functional mode!s of attitudes. Finally, we discuss the potential application of these procedures to determining motives underlying brand purchase decisions.

While this presentation will focus extensively on various techniques developed and used at FCB, it is most emphatically not our intent to "sell" the techniques or the agency. Rather the intent is to provide an example of what practitioners are doing and find useful in the hope that this will stimulate and guide the thinking of academicians.

THE FCB GRID

In order to motivate our discussion of extensions of the FCB Grid model, we mat;e a brief presentation of the basic features of that model. Details are presented in (Vaughn 1980, 1986; Ratchford 1987).

The development of the grid model be,,an with the observation what traditional consumer behavior theories provided four basic explanations:

Economic motives with cost/utility forces driving pu. chase.

Responsive actions based on rote learning and habit formation

Psychological drives from deep emotions and subconscious desires.

Social goals reflected in peer-group imitation, role status and visibility.

Although interesting in the abstract, these theories never made much practical contribution to more effective advertising. They did, however, lead to the widely-known hierarchy of effects model ar,d eventually to other useful theorizing.

The FCB Grid can be seen as an extension of the hierarchy model to different types of purchase decisions. The grid postulates that the hierarchy varies depending on whether the decision is: (1.) high vs. low in involvement; (2.) thinking vs. feeling. Involvement, in its simplest form, is interest level or motivational intensity. Think/feel is left brain (logical, linear)/right brain (holistic, symbolic) information processing (see, for example, Holbrook and Moore 1981, Zajonc and Marcus 1982, Batra and Ray 1985 for a discussion of this distinction). Though the general idea that information processing varies with involvement and think/feel was not new, the basic insight of the grid model is to incorporate these into a simple planning framework. This can be described as in Figure 1.

This grid might be thought of as a simplified model of the consumer's mind space. To determine where specific purchase decisions fall on the grid, scales to measure a consumers' perceived levels of involvement and think/feel were developed; details are presented in Ratchford (1987). Using these scales, consumers' placement of a large number of purchase decisions on the grid has been studied in the U.S. and around the world. Especially because it acknowledges the "feeling" dimension of purchase decisions, the grid has proven very useful for stimulating advertising ideas in the visual, nonverbal realm and for challenging FCB to better understand and use emotion in strategic and creative development.

EXPANDING THE FEELING DIMENSION

While think and feel refer to modes of information processing in the grid model, these are closely related to underlying categories of motives. Thus if the purchase decision is based on a utilitarian motive (Katz 1960, McGuire 1976, Holbrook and Hirschman 1982) -- the need for performance of a function on one or more readily-defined attributes, information processing will be thinking (cognitive), stressing evaluation on performance-related dimensions. In this case, one might expect the traditional multi-attribute model to work well.

But there a number of motives which are likely to trigger "feeling" decisions -- decisions which require holistic, symbolic modes of information processing. In general, these motives would be what McGuire calls affective motives; affective motives stress the need to reach satisfied feeling states. From McGuire's classification system, and a reading of related literature, it is suggested that the major categories of "feeling" motives are as follows:

1. Ego gratification: the need to defend, enhance, and express one's basic personality (Vaughn 19803. This is an important component of McGuire's (1974, 1976) classification system of motives, and encompasses Katz' (1960) ego-defense (negative) and value expressive (positive) functions.

2. Social acceptance: the need to be viewed favorably in the eyes of others. While this could be viewed as a component of ego gratification, it is discussed by Vaughn (1980) as social theory, and is a component of McGuire's classification system (affiliation, modeling). It therefore seems useful to list this as a separate category.

3. Sensory: the desire for pleasure to any of the five senses. While this might technically be labeled ego gratification, something like consuming a pizza need not be central to one's values. Nevertheless, this event involves a type of affective satisfaction, e.g., the taste feels great. Food items which go beyond utilitarian needs, music, pleasant scenes, all satisfy sensory needs.

Given their nature, gratification of the above three affective motives is likely to require "feeling" information processing: emotion, image, and holistic judgments are likely to be central. Consider, for example, the purchase of perfume, in which all three of the above motives are likely to play a role: the perfume may express characteristics of one's personality, must be acceptable socially, and obviously must have the proper smell (pleasant, sexy). Instead of carefully weighing the costs and benefits of levels of various attributes, the perfume buyer is likely to base his/her decision on a mixture of holistic evaluation of smell, brand imagery, judgments of how others will like the perfume, emotions conjured up by the various brands.

None of these is likely to be easily verbalized. Clearly, the standard practice of asking consumers to rate perfumes on a number of attributes, and scoring each brand on some average or sum of the ratings, sill not capture the decision process for choosing perfume or other feeling products very well. What is needed are techniques for determining how brands fare on the various dimensions of "feeling." Moreover, these techniques must be capable of preserving aspects of these judgments which are difficult to verbalize The two techniques, VIP and ICON, which are outlined in the following section of this paper, attempt to accomplish these objectives

VIP AND ICON

Overview

YIP, or Visual Image Profiling, has been used for a number of years. In VIP, stimuli consist of 100 photos of faces, 50 men and 50 women, which have been rated on an array of personality and lifestyle characteristics by a U.S. probability sample of consumers. In applications of VIP, respondents assign these to brands in a product class; the result is a profile of personality and lifestyle characteristics associated with each brand. To the extent that persons choose brands which are consistent with their self concept, which is implied by the ego satisfaction motive, brand choice should be driven by these personality and lifestyle characteristics.

ICON, or Image Configurations, is a newer technique which attempts to get at emotional content associated with brands. Stimuli consist of 60 photos of situations which have be n rated on emotional content by a sample of respondents. As in VIP, respondents assign these pictures to brands; the result is a profile of emotional associations for each brand. Again, it is postulated that these emotional associations will be related to brand choice. An important feature of both VIP and ICON is that respondents assign the pictures, rather than verbal rating scales, to brands: the intent is to get beyond the possible limitations of applying verbal methods to the study of "feeling" items.

Development and Application of VIP

While the nonverbal feature of this technique is of central importance, there is a need to know how consumers are likely to evaluate each of the photos used as stimuli. This is most readily done through conventional verbal techniques. For VIP, a development study was done in which each of the 150 photos (50 male and 50 female) was rated by a sample of 50 respondents on 99 personality and lifestyle items, demographics, and affinity and aspiration. The personality and lifestyle items were in the form of statements with the stem: "The person in the picture is the kind of person who..." Responses were on a three-point scale: "describes very well, describes somewhat, does not describe at all." Demographic

questions asked respondents to judge the age, income, education, occupation of the person in the picture. Finally, respondents were asked whether they share similar qualities with the person in the picture (affinity), and whether they feel hey would like to have qualities similar to those of the person in he photo (aspiration).

To determine lifestyle and personality traits, responses to the 99 items were factored (separate analyses were run for men and women), and additive scales were constructed .or each factor. For both men and women 16 factors were retained. Examples of both male and female factors are presented n Table 1.

Unfortunately, space limitations and the limitations of printing technology precluded us from including in this paper several examples of the VIP and ICON photos which were in our ACR presentation; interested readers may obtain a longer version of this paper which contains these examples from the senior author.

The female picture included in our presentation was rated by respondents as an "acceptance seeker", "sociable", "organized intellectual", age 40, income $35,000, college educated, white collar, managerial occupation. By contrast, the male picture which we showed was rated as "hard driving", "individualist", "pragmatist', having a violent temper, cynical and disbelieving. Altogether, he's a tough customer, who might easily be part of a brand's personality profile.

.When respondents apply the pictures is brands in a category, examining the collage of pictures applied to a given brand often provides strong insights into the personality associations of a brand. For example, one collage, which was associated with one brand of a commonly purchased package goods product, showed blue-collar, rather rough-and-tumble males. In comparison, the profile of a competing brand showed young, modern mothers, and family settings

TABLE 1

EXAMPLES OF MALE AND FEMALE FACTORS FROM VIP PICTURE RATINGS

While it is possible to derive a score for each brand on each personality dimension by adding the scores on each dimension for each picture assigned to the brand, collages such as those discussed above tend to have the most practical value for creative work: the visual information contains a richness which is not readily portrayed in verbal scales. At the same time, the visual information can be subject to misinterpretation; the verbal scales are useful for providing guidance. But, aside from this, the verbal scales are secondary in importance to the visual information provided in the brand personality profiles.

Development and Ap plication of ICON

Except that emotions rather than personality factors are involved, the procedure used for developing ICON is similar to that for VIP. Thus an assessment of the emotional content of the 60 ICON pictures was made by having each of the pictures rated on 75 emotion words by a sample of 50 respondents. The emotion words were obtained from a thorough study of the extensive literature on typologies of emotion in psychology: at least 38 sources were consulted, and an attempt was made to cover all of the major categories of emotion delineated in the literature. For each picture which they saw, respondents were asked to "X" the boxes for all words that describe how each picture makes you feel.

In analyzing emotional associations for each picture, we looked at both the pattern or individual words assigned to the picture, and at general emotion categories associated with each picture. The latter were constructed by applying the VARCLUS procedure in SAS to the word assignment data, where the observations were pictures by subjects. Twelve categories of emotion, which are broadly consistent with other classifications in the literature, emerged: Joy, Displeasure, Tenderness, Enthusiasm, Confusion, Attraction, Sympathy, Depression, Serious, Quiet, Fear/Hate, Skeptical. Additive scales of the words comprising each of these Categories were constructed.

TABLE 2

ALTERNATIVE INDEXES OF EMOTION FOR TWO BRANDS OF TOOTHPASTE

As with VIP, the main intent of associating words with pictures is to provide guidance about the meaning of the pictures. As stated above, i.e main application of ICON is to determine which pictures consumers associate with brands, and thereby to obtain evidence about emotions associated with brands. As with VIP, it is generally most instructive to look at the collage of pictures associated with competing brands.

As an example, consider two photos which a sample of consumers most often associated with a major brand of toothpaste (call it Brand A). It would be evident from looking at the photos that an underlying emotional content of joy and tenderness is associated with this brand. Conversely, it would be evident from looking at the photos of a competing brand (Brand B) that emotions associated with this brand are tranquil, quiet, relaxed, sad, sentimental.

As with VIP, visual examination of the collage of pictures assigned to a brand is likely to be most helpful in stimulating thinking about advertising strategy. However, it is also possible to develop a score for each brand on words or emotions from the picture assignment data. For each word k, a score for each brand j might be computed as:

S_kj = SUM_i (w_kip_ij)

where w_ki is the proportion of assignment of word k to picture i, p_ij is the proportion of assignment of picture i to brand j, and the sum is over the i = 1...60 pictures. Similarly, for each generic emotion m, a score for each brand j may be computed as:

T_mj = SUM_i (e_mip_ij)

where e_mi is score of picture i on emotion m (proportion of the time words comprising that emotion are assigned to that picture), and everything else is as above.

At this point one might ask why it isn't easier and better to simply have respondents rate brands on words directly rather than have them assign pictures which have been rated on words. While we may not have a complete answer to this question at this time, it is clear that using pictures will not necessarily give the same results as having brands rated on words directly. In the study of toothpaste brands discussed above, for example, it was found that the highest scoring words describing Brand B were tranquil, quiet, relaxed, sad, sentimental when pictures were used, and the scoring method outlined above was applied; when Brand B was rated on words directly, the highest scoring words were satisfied, secure, protected, nostalgic, pleased.

To continue the example, Table 2 presents indexes of emotion for toothpaste Brands A and B. For pictures, these were constructed by first scoring, each brand on each emotion as outlined above. Then the scores were divided by the grand mean of all scores across emotions and brands. This gave an index unadjusted for propensities to assign a particular emotion or brand. The indexes in Table 2 were obtained after removing the effects of emotion and brand by dividing the unadjusted indexes by the product of overall indexes for the corresponding emotion and brand: these "double centered" indexes may be interpreted as measuring degree of association between brand and emotion. The procedure for constructing indexes from words assigned directly to brands was the same, except that raw scores were simply the proportion of the time words associated with that emotion were assigned to that brand.

Table 2 indicates that there is a reasonable amount of agreement between indexes constructed from pictures and words for Brand A; the correlation coefficient between the two sets of indexes for Brand A is .67. But, for Brand B, there is almost no agreement between the two sets of indexes: the correlation between them is only .27, and is not significant. For pictures, there are above average scores for depression and displeasure for Brand B; these negative associations do not come out for words applied directly. Conversely, Brand B is well above average on tenderness for words, but well below average on this emotion for pictures.

FIGURE 1

THE GCB GRID

While reasons for similarities/differences in emotional profiles when pictures vs. words are assigned to brands must be the subject for further study, we conjecture that differences may result because pictures get at emotions which are not easily expressed verbally.

Applications of VIP and ICON

Perhaps the most important point to make about VIP and ICON is that they are eminently practical tools which have been used with considerable success: the advertising community badly needs tools which attempt to go beyond the limitations of purely verbal responses. While both techniques have been used in quantitative studies, they have been used most extensively in small-sample qualitative work where the general intent is to develop strategies which may lead to fruitful creative ideas. VIP can be used to profile brand users or dimensionalize personality issues wherever and whenever they may be creatively useful. ICON is primarily a diagnostic tool that generates soft, exploratory data for creative development. It provides emotional range and nuance identification by providing brand feeling differences and similarities.

The two techniques have been integrated into a planning system called Consumer Insights Mapping. In the first stages brands and category attributes are plotted in the involvement and think/feel spaces of the Grid depicted in Figure 1. Then VIP and ICON are used to add personality and emotional depth. The result is a thorough mapping of knowledge of the category which becomes the first step in the advertising planning process.

POTENTIAL APPLICATIONS TO DETERMINING MOTIVES

The techniques outlined above are meant to get at perceptions of brands on dimensions cf personality or emotion. The techniques are related to the major dimensions of feeling outlined above: VIP would seem most directly related to ego gratification: expression of one's personality; emotional associations uncovered through ICON' might be related to any or the three categories of feeling - ego, social, sensory - outlined above. The next step is to establish a link between these perceptions and overall evaluations of brands (attitudes), and consequent brand choices.

In the case of VIP, the link would appear to be well established. rr,us there is a well established body of theory which postulates a relationship between self image and brand image; this body of theory motivated the development of VIP in the first place. In essence this body of theory postulates that consumers will choose brands which are congruent with either their actual or ideal self image. If so, knowledge of personality characteristics associated wi!h a brand should help one to predict what types of consumers still choose it. While work in this area is plagued by the usual definitional and measurement problems, empirical evidence generally supports the postulated link between brand image and self image (Sirgy 1982). Thus VIP can be linked to purchase decisions and choice.

While there doesn't seem to be any elaborate body of theory linking the emotions revealed by ICON to brand choice, this is probably not needed: it seems obvious that positive or negative emotions can be associated with the consumption of alternative brands; and that these emotions can motivate brand choice. Havlena and Holbrook (1986) demonstrate that respondents can readily associate emotions with consumption experiences.

The more interesting question would be to determine the relative influence of brand personality as revealed by VIP, brand emotion as revealed by ICON, and other utilitarian attributes on brand choice. This would say something about the relative importance of image congruence, emotion, and utilitarian factors as motives. Such evidence would provide valuable guidance for advertising: one would like to stress the factors which are most salient to overall evaluations. Moreover, this evidence would provide a means of validating the placement of individual products on the FCB Grid: one would expect the utilitarian facto: s to be most important for "thinking" products, the personality and emotional factors to be most salient for "feeling" products.

FIGURE 2

RELATION BETWEEN BRAND PREFERENCE AND DETERMINANTS

Lutz (1978) provides a valuable clue about how to evaluate the relative contribution of the various thinking and feeling factors to one's overall evaluation of a brand. Lutz argues that the expectancy-value framework implicit in the standard multiattribute model used in marketing (Wilkie 1986, p.460), can be used to measure attitude functions, e.g., their motivational basis. One could thus measure attitude toward a brand as the sum of expectation that the brand performs a particular function (expectation that it is congruent with one's personality, provides a particular emotional experience, performs a given utilitarian function) weighted by the importance of that function. The weights would thus indicate which motives were most influential for a particular product category .

One might estimate the weights using the general framework outlined in Figure 2. The VIP and ICON techniques might play a central role in this process. In Figure 2, brand personality, emotions associated with the brand, and brand attributes determine an overall evaluation, which might be measured in terms of a preference rating, or infer cd from brand choice. The attributes also may have an indirect effect on preference through possible effects on brand personality and emotions. While the framework in Figure 2 might easily be extended to incorporate influences of advertising and other exogenous factors on perceptions of attributes, personality, and emotions, these extensions are not necessary for the study of the basic relation between motives and overall preferences.

Measurement would follow from ratings of alternative brands by a sample of consumers. Conventional attribute ar,d preference measurements could be obtained. In addition, consumers would apply the VIP pictures to both the brands and to their own personality. This would provide the basis for calculating a congruence measure for each brand: one of the measures suggested by Sirgy (1982) might N used. The congruence measure might be derived by adding the scores associated with each picture on major personality/lifestyle dimensions across pictures assigned to brands or oneself. Conversely, it might be obtained from examining the frequency of matches of pictures assigned to oneself and a given brand. Similarly. respondents would be asked to assign the ICON pictures to each brand. As outlined a&;e, scores on major dimensions of emotion can be calculated for each consumer from the picture assignments.

For each brand and consumer this would result in a preference measure, ratings on a number of functional attributes, a measure of congruence between brand and self, scores on major dimensions of emotion. Using data pooled across consumers and brands, the relation between these variables outlined in Figure 2 might be estimated as a recursive system of equations. The regression coefficients would then identify the importance of each of the determinants of preference.

CONCLUDING COMMENTS

This paper has focussed on approaches that have been found to be of value to advertising practitioners. By definition these approaches suggest directions which these practitioners would find a fertile ground for future development. The major themes running through our discussion have been the need to get at "feeling" aspects of purchase motives and decisions, and the need to go beyond conventional verbal techniques to get at aspects of these feeling decisions which are difficult to verbalize. We believe that further efforts in these directions, and also in determining the relative influence of thinking vs. feeling factors in various purchase decisions. could have great practical value.

REFERENCES

Batra, Rajeev and Michael C. Ray (1935), "How Advertising Works at Contact," in Linda Alwitt and Andrew C. Mitchell, eds., Psychological Processes and Advertising Effects, Hillsdale, N.J.: Lawrence Erlbaum.

Havlena, William J. and Morris B. Holbrook (1986), "The Varieties of Consumption Experience: Comparing Two Typologies of Emotion in Consumer Behavior," Journal of Consumer Research, 13 (December), 794-104.

Holbrook, Morris B. and Elizabeth Hirschman (1982), The Experiential Aspects of- Consumption: Consumer Fantasies, Feelings and Fun," Journal of Consumer Research, 9 (September), 132-140.

Holbrook, Morris B. and W.L. Moore (1931), "Feature Interactions in Consumer Judgments of Verbal vs. Pictorial Presentations," Journal of Consumer Research, 8 (June), 103-113.

Katz, Daniel (1960), "The Functional Approach to the Study of Attitudes," Public Opinion Quarterly, 24 (2), 163 -204.

Lutz, Richard (1978), "A Functional Approach to Consumer Attitude Research," in H. Keith Hunt, ed., Advances in Consumer Research, Volume >, Ann Arbor: Association for Consumer Research, 360-369.

McGuire, William J. (1974), "Psychological Motives and Communication Gratification," in J.G. Blumler and E. Katz, eds., The Uses of Mass Communications: Current Perspectives on Gratifications Research.

McGuire, William J. (1976), "Some Internal Psychological Factors Influencing Brand Choice, Journal of Consumer Research, 2 (March), 302319.

Ratchford, Brian T. (1987), "New Insights About the FCB Grid," Journal of Advertising Research, 27 (August/September), 24-38.

Sirgy, M. Joseph (1982), "Self Concept in Consumer Behavior: A Critical Review, " Journal of Consumer Research, 9(December), 287-300.

Vaughn, Richard (1980), "How Advertising Works: A Planning Model," Journal of Advertising Research, 20 (5), 27-33.

Vaughn, Richard (1986), "How Advertising Works: A Planning Model Revisited," Journal of Advertising Research, 26 (1), 57-66.

Wilkie, William L. (1986), Consumer Behavior, New York: John Wiley and Sons.

Zajonc, R.B. and H. Marcus, "Affective and Cognitive Factors in Preferences," Journal of Consumer Research, 2. 123-131.

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