An Empirical Study of Lifestyle Correlates to Brand Loyal Behavior

Stephen M. Goldberg, University of Pennsylvania
ABSTRACT - This paper investigates a causal link between lifestyle and brand loyal behavior in the context of a beer market. The results indicate that lifestyle advertising themes may cause a negative response by those not favoring the promoted lifestyle.
[ to cite ]:
Stephen M. Goldberg (1982) ,"An Empirical Study of Lifestyle Correlates to Brand Loyal Behavior", in NA - Advances in Consumer Research Volume 09, eds. Andrew Mitchell, Ann Abor, MI : Association for Consumer Research, Pages: 456-460.

Advances in Consumer Research Volume 9, 1982      Pages 456-460

AN EMPIRICAL STUDY OF LIFESTYLE CORRELATES TO BRAND LOYAL BEHAVIOR

Stephen M. Goldberg, University of Pennsylvania

[Stephen Goldberg is a Ph.D. candidate at the Wharton School, University of Pennsylvania, Philadelphia, PA 19104. He would like to-acknowledge his appreciation to Scott Wart of the Wharton School for all his helpful comments.]

ABSTRACT -

This paper investigates a causal link between lifestyle and brand loyal behavior in the context of a beer market. The results indicate that lifestyle advertising themes may cause a negative response by those not favoring the promoted lifestyle.

INTRODUCTION

Virtually all marketers attempt to develop and maintain customer loyalty to the company's brands. What is not so clear is how marketers can foster the development of brand loyalty, and maintain it among consumer segments over time. Much past research has attempted to profile the brand loyal consumer for various kinds of products and services (Frank, 1967; Fraser and Wind, 1980, etc.). However, little research has attempted to relate consumer lifestyles to brand loyalty, in spite of the current emphasis on lifestyle variables as a basis for market segmentation and promotional objectives. This study addresses that issue -namely, to differences in consumer lifestyles provide a useful basis for predicting differences in brand loyalty?

A first consideration is a precise definition of both "brand loyalty" and "lifestyle." Despite the myriad of studies on brand loyalty, much confusion exists about its definition. Jacoby and Chestnut (1978) reviewed fifty-five definitions/measures employed in various research efforts Over several years and arrived at this definition of brand loyalty:

(it is) . . . the biased (i.e., nonrandom), behavioral response (i.e., purchase), expressed over time by some decision-making unit with respect to one or more alternative brands out of a set of such brands, and is a function of psychological (decision-making, evaluative) processes.

This definition is intuitively pleasing because it rejects the potentially spurious loyalty that is detected when using a "purchase runs" criterion (loyalty is present when an undetermined number of consecutive purchase of an item is made). It also posits that brand loyalty can exist for multiple individuals within some decision-making unit. Thus if the husband is directing the purchase of a particular product but the wife is performing the actual purchase, loyalty exists for the decision-making unit of both husband and wife, not just the wife alone.

Jacoby and Chestnut's definition provides a useful step beyond previous attempts; however, it does not elaborate about possible psychological processes which may underlie the formation ant/or maintenance of loyalty. This paper seeks to explore those processes; namely, the causal relationships between psychological processes inherent in different psychographic lifestyle patterns, and the degree of brand loyalty.

The practical concerns of marketers addressed here center around the issue of selection of promotional themes as they may relate to the creation or maintenance of brand loyalty.

In some markets, advertisers tend to focus on image themes directed at particular psychographic segments, rather than on physical attributes. For example, while the blue jean market might be broadly segmented into "nondesigner" jeans (e.g., Penney's, Levi's, etc.) and "designer" jeans (e.g., Calvin Klein, Jordache, etc.) at least brands in the latter category might be considered homogeneous in that they all focus on image aspects of the brands rather than on other attributes (e.g., physical characteristics). Similarly, the major national beers (as opposed to local, price beers or premium beers) often compete on the basis of lifestyle themes. One can conclude that in these and similar situations, marketers believe that purchase behavior (e.g., loyalty) results from linking use of the product to certain lifestyles. Moreover, the implicit assumption would seem to be that individuals with different lifestyles than those portrayed and appealed to will, at the very least, not react negatively.

In the following sections, brand loyalty and lifestyle, or psychographics, are briefly discussed; the methodology and variable section for the study is then described and results presented. Finally, conclusions and implications are drawn for the loyalty-lifestyle relationships and managerial ramifications discussed

BRAND LOYALTY

In an early experiment, Tucker (1964) established that brand loyalty develops over time. He concluded.

Whether one looks at the concentration of household purchases among brands at a point in time or at its stability over time, there is marked evidence that brand loyalty is a 'real' and reliable phenomenon.

Consequently, much of the subsequent research concerning the question of brand loyalty has been directed toward developing an adequate measure of the phenomenon in the real world. Jacoby and Chestnut's (1978) review of thirty-three different behavioral indices of brand loyalty leads us to the conclusion that brand loyalty exists if it can be measured post hoc (i.e., it is an empirical phenomenon exclusively). Most of this kind of research specifies criteria by which loyalty obtains, rather than defining, a priori, what it is and then trying to measure it. This is intuitively unsatisfying; if brand loyalty is nothing more than repeat purchase behavior, it is strictly an empirical phenomenon and has limited applicability for marketing strategy. Day (1969) emphasized this point:

There is a difference between true or 'intentional' loyalty and 'spurious' loyalty associated with consistent purchase of one brand because (among other) reasons) there are no substitutes or because of a long series of deals 9 etc.

In Jacoby's words (1971), " . . . repeat purchase behavior is a necessary but nonsufficient condition for brand loyalty." The concept of brand loyalty must account for the motivational bases of repeat purchase behavior, both to contribute toward theory development and to provide direction for marketing action.

In an effort to link a priori motivations to subsequent repeat buying, researchers have explored attitudinal indices of brand loyalty. Consumer statements of loyalty (e.g., attitudes toward the brand) are taken to be indicants of brand loyalty. The critical problem here is that there is no direct tie to purchase; by these measures a person could be quite loyal but rarely buy the product (e.g., the consumer has most positive attitudeS toward car A but buys the less expensive brands). Without a direct link to behavior, it is virtually impossible to determine the functional relationship between the attitude toward the brand and the purchase of it. Again, this would be very unsatisfying to a marketer. In order for the concept to have practical use and contribute toward a viable theory. a measure of both attitudes and behavior is needed as a basis for inferring brand loyalty.

The final thrust of research in the brand loyalty area attempts to derive composite indices of brand loyalty which combine attitudes and behavior. Jacoby (1971) says,

. . . it seems reasonable to reserve the phrase 'brand loyal behavior' for instances where the purchase decision is the result of a process in which various alternative brands are psychologically (perhaps even physically) compared and evaluated on certain criteria and the 'optimal' brand(s) is (are) selected.

Several authors have sought to develop correlates of brand loyalty, with varied results. Carman (1972) reports positive results when looking at store loyalty, some personality variables, information sources, and miscellaneous others. However, the major explanatory variable is store loyalty--not surprising considering the number of brands available was not controlled for in the study. Nonetheless, the study does support the search for a causal factor to explain brand loyalty.

Frank (1967), in a series of articles dealing with this subject, had limited success in trying to develop indicants of brand loyalty for grocery products. In one study, Frank tried to segment the market based on consumption, degree of repeat purchase, store loyalty, extent of dealing, and some other variables. Re used a factor analytic approach to develop a brand loyalty dependent variable; however, Frank does not use any attitudinal indices. He regressed husband and wife scores on the Edwards Personality Preference Schedule, but concluded that brand loyal consumers cannot be profiled on the basis of these personality measures. In more recent years, psychographic measures generally have supplanted personality tests in attempts to relate personal characteristics to specific buying behavior (Wells, 1974, 1975). Consequently, it may be fruitful to look at these types of measures to see how well they explain brand loyalty.

PSYCHOGRAPHICS OR LIFESTYLE

Lifestyle and psychographics are commonly used as synonyms; however, there is a slight difference as noted by Hustad and Pessemier (1971):

(Psychographics) imply a broad range of general psychological and personality measures . . . Life Style . . . puts emphasis on attitudes and activities and also examines 'state' variables . . .

Wells (1974) elaborated:

. . . psychographics refer to comparatively heavy emphasis on generalized personality traits . . . 'lifestyle' . . . tends to focus on either broad cultural trends or on needs and values . . .

Of course, there is nothing to prohibit using many different types of statements within a large battery to cover attitudes, activities, "state" variables, personality traits, etc. The common procedure is to subject the responses to a factor analytic procedure to determine the underlying constructs of lifestyle (Kerlinger, 1964; Bruno and Pessemier, 1972).

METHODOLOGY OF THE CURRENT STUDY

A questionnaire was administered (via personal interview) to a random sample of over 1,000 Canadian beer drinkers as part of an overall market segmentation study. Respondents supplied several types of data which were used to develop a brand loyalty measure,lifestyle measures and segmentation criteria. In addition, each supplied consumption and demographic information.

Brand Loyalty Measure

Following Jacoby, a Brand loyalty measure was constructed which links attitudes toward the favorite (or usual) brand and purchase behavior. Specifically:

EQUATION

[The attitude statements which subjects evaluated on semantic differential scale (coded O to 6) reflecting their attitudes toward their favorite brand were:

An average quality beer 1 2 3 4 5 6 A very high quality beer

I don't like it 1 2 3 4 5 6 I like it

A beer that doesn't have a crisp refreshing taste 1 2 3 4 5 6 A beer that has a crisp refreshing taste

Hard to digest1 2 3 4 5 6 Easy to digest

Harsh 1 2 3 4 5 6 Smooth

A beer not brewed more carefully than others 1 2 3 4 5 6 A beer brewed more carefully than others

A beer that would not be completely acceptable to all my friends 1 2 3 4 5 6 A beer that would be completely acceptable to all my friends

A watery beer 1 2 3 4 5 6 A full-bodied beer]

This measure's two components represent indices which interact to form the construct of brand loyalty. Management indicated repeat-purchase percentage was about 75% which was similar to the sample value of 73%. The mean loyalty here was only 67% for the sample--conceptually, attitudes seem to attenuate behavior.

Both components of this composite measure have the desirable property of having a zero to one range. In addition, each one is unambiguous in its assignment of its share of loyalty--the greater the value, the more loyal a person is. Consequently, the product of the two components is also on a zero to one scale, where one denotes complete loyalty and zero indicates no loyalty whatsoever. Practically speaking, one would not expect to see a zero score for any respondent in this study given s/he has specified, and is evaluating, her/his favorite brand. However, if a person specified a brand which is purchased only about half the time and s/he is only lukewarm toward it (i.e., marked a three or averaged three on all semantic differential scales). then the brand loyalty value would be only 0.25.

Lifestyle Measures

In contrast to what one might consider a relatively unique measure of brand loyalty, the lifestyle measures employed in this study were rather ordinary. A battery of over sixty psychographic statements were self-administered by the respondents. The battery used was a subset of one of the more or less standard ones used with the addition of several unique statements.

Unfortunately, there are two potential problems with measures of this kind: first, the interpretation of factors derived from the analysis is subjective; second related to the first--is that no standard set of lifestyles exists; e.g., results are idiosyncratic to the particular data set analyzed. However, due to the vast number of published psychographic studies, Bruno and Pessemier (1972) were able to examine common factors and loadings on factors across studies. Table l shows how the factors obtained from this study compare with those highlighted by Bruno and Pessemier. One can see that there is a fair congruence between factors derived in this study and factors derived in other studies. Hence, there is reasonable confidence in the validity of the lifestyle constructs derived here.

TABLE 1

A PARTIAL LIST OF VARIABLES WHICH LOAD TOGETHER; AND FACTOR LOADINGS FOR THIS STUDY (G) AND THOSE FOUND BY PESSEMIER & BRUNO (PB) (IF AVAILABLE)

In this study, the factor solution was limited to the extraction of ten factors whose combined accounted-for variance was 38.8% (at the upper end of the range--30 to 44%--found by Bruno and Pessemier in their comparison). These lifestyle constructs were judgmentally identified as:

Health Concerned                 Cosmopolitan Sophisticates

Swingers                               Passive Sports Enthusiasts

Social Responsibility             Outdoors Enthusiasts

Conservatives                       Constant Feeders

Independent Leaders             Drinkers

Appendix A provides a complete description of each lifestyle construct.

Market Segmentation

While lifestyle or psychographics are often used as segmentation criteria, this would provide little new information. There are many potential segmentation criteria available, two of which had primary appeal--demographics and attitudes toward beer in general. Management felt that product-use segments would provide the most information regarding the beer drinking market, thus demographics were discarded and a post hoc segmentation scheme (Green and Tull, 1978) was adopted. Consequently, a Howard-Harris cluster analysis was conducted using attitudes toward beer; a five cluster solution was derived (based on the decrease in pooled within-groups variance). Given the centroids of each group, the segments were defined as follows:

Beer is just another beverage- These people do not attach any special import to beer or beer consumption

Beer as a These people drink beer as a substitute for hard liquor for a variety of reasons

Negative toward beer- These people drink beer but it's difficult to ascertain just why

Hard core beer drinkers--These people drink beer regularly and often as a reward

Proud social beer drinkers- These people like beer and are proud to drink it on social occasions

A demographic profile of the market segments in shown in Table t. Segments two, four, and five show the highest weekly consumption, tend to be predominantly male, and are older than the sample population in general. Segments three and five are the most educated and have the highest incomes.

TABLE 2

DEMOGRAPHIC PROFILES OF SEGMENTS - INDEXED AS % OF TOTAL SAMPLE

Research Questions

Given the tremendous amount of lifestyle advertising which exists in this market, the question of whether brand loyalty can be predicted based on lifestyle seemed to be quite relevant. If none of the lifestyles were to prove significantly related to loyalty, it would indicate that managers ought to pursue other advertising themes for their products.

One possibility, however, is that any specific effect may be significant to one market segment but not to others. Hence, the effect would be undetectable when looking at the market as a whole. In addition, a specific effect could be lost if it is positive in one segment and negative (of roughly the same magnitude) in another segment. Consequently, a second research question was posed--do lifestyle effects vary across different attitude segments? Since most campaigns are targeted at specific segments, knowledge of lifestyle patterns among drinkers provides considerably more insight for the marketer.

The regression results can be seen in Table 3. In examining the intercepts (which are, in a sense, the mean brand loyalty for the segment), there are some intuitively satisfying results. Overall, the beer market exhibits very high brand loyalty an average of 0.67. While adjusted R-square is not high, eight of ten lifestyle coefficients are significant and their direction was in accord with management's feelings. [Since the factor analysis accounted for only a portion of the variance in the psychographic data, and a substantial amount of within-groups variance from the cluster analysis remained, it would be expected that a multiple R2 from regression would be relatively low, in the neighborhood of 0.05 to 0.15. Wells (1971) also noted that with data of this kind, a high proportion of accounted-for variance is not likely to occur:

When expressed as product-moment correlations, the relationships between AIO items and products or media are low--often around 0.2 and seldom higher than 0.3 or 0.4. Thus, they do not 'explain the variance' very well, even when put together in a prediction equation . . . It should be remembered, however, that the variance 'explained' is the variance in the behavior of individuals, not the variance in the average behavior of groups, so a product-moment correlation of 0.2 is deceptively small.]

TABLE 3

SIGNIFICANT REGRESSION COEFFICIENTS FOR LIFESTYLE FACTORS (p = .10)

Since the regression for each segment is significant and the test for differences in coefficients across groups is also significant, the hypothesis that lifestyle effects differ across segments is supported. Indeed, a look at the mean loyalty- for the segments shows marked differences. As might be expected, the negative toward beer segment is the least loyal on average. Interestingly enough, the hard core segment is not the most loyal; in fact, it is just at the population average. Perhaps these people are somewhat more committed to drinking beer than they are to a particular brand, The most loyal segment is the beer as a substitute segment. It appears once these people find a substitute, they are loyal to it.

The results shown in Table 3 indicate that the composition of significant effects in each segment differs; in fact, the only lifestyle with an effect in all segments is the leader lifestyle (in the negative direction as would be expected from the overall leader coefficient). Thus, there is additional support found that lifestyles are differentially loyal depending upon the segment under consideration.

Two lifestyles, swingers and outdoors enthusiasts, show some of the most interesting results. The change in sign of the regression coefficients across segments demonstrates the potential peril of lifestyle themes in advertising. A strategy aimed at outdoors enthusiasts might build loyalty in the hart core segment, but these people tend to be less loyal in the social segment--a segment which shows higher consumption than the hard core one.

Intuitively, one would think sports enthusiasts would have a propensity toward loyalty across all segments. But this is true in only two segments--the substitute and negative ones. Considering all the sports oriented beer advertising directed at the hard core segment, it appears much money is being wasted.

A marketer devising a strategy to build or maintain loyalty for his brand would obviously seek to concentrate on the substitute and social segments for two reasons. First, average loyalty in these segments is highest; and, second, consumption is highest there too. In the substitute segment, though, appeals based on lifestyle themes may have little effect given the few positive coefficients and the high loyalty already present.

The social segment offers an advertiser two highly positive lifestyle effects- swingers and conservatives. Nevertheless, execution of a campaign will require care since three of the lifestyles have negative coefficients in this segment.

Additional Analyses

Among others, Fraser and Wind (1980) have examined demographic measures along with attitudinal measures in a predictive model of behavior. The question of whether the addition of demographic variables would affect the resuLts of the earlier analysis was investigated. While the number of significant demographic variables differed across segments, changes in the number and value of significant lifestyle coefficients (which were the primary interest) were negligible- the mean sum of squared deviations was 0.0001. However, the variance accounted for in each segment's regression equation improved substantially:

TABLE

The test for differences across groups was still significant. Thus, it can be concluded that the effect of lifestyle is stable though the addition of other explanatory variables may improve predictive power o, the overall model.

CONCLUSION

It should be obvious that brand loyalty is an important concept to a marketer. From there, it is not a great leap to say that the dimensions of brand loyalty and the ability to predict loyalty should be understood as completely as possible.

In this study, it was shown that lifestyle has a differential effect on brand loyalty for different consumer segments. In addition, the lifestyle effect was stable with the inclusion of other explanatory variables. While the change in lifestyle effects was minimal, one area for further research would be to investigate the possible interaction effects of demographics and lifestyle on brand loyalty.

Since this study was limited exclusively to the beer market, another fruitful area of research would be to see if people with different lifestyles exhibit differing loyalty patterns with respect to other consumer goods (such as coffee, breakfast cereals, and other frequently purchased products). Such an investigation would be useful to marketers in planning their communications strategies.

APPENDIX A: LIFESTYLE DESCRIPTIONS

A. Health Concerned--A major part of one's life is spent on activities surrounding the maintenance of the body and the mind, i.e., exercise, grooming, entertainment, etc. These people pursue activities which maintain their-physical condition and their surroundings.

B. Swingers--Life revolves around social activities. These people are adventurous and behave in ways which enhance this characteristic.

C. Social Responsibility--An important part of life is wrapped up in promoting general welfare and uncovering potentially harmful actions by others. These people are concerned with identifying major social problems and leading the charge to solve them.

D. Conservatives--In some ways, these people are living in the past. Life revolves around dependence on input from friends, media, etc. These people do not change often and are relatively proud of it.

E. Independent Leaders--Life is a series of decisions; decisions made for others or for oneself. These people are busy influencing what others do, at work or at leisure.

F. Cosmopolitan Sophisticates--Life is a series of intellectual, worldly, fashionable pursuits. These people are doing all the "right" things. There is an acceptable model for a person to which they aspire.

G. Passive Sports Enthusiasts--These people live for sports events and news concerning them. Another appropriate term would be armchair quarterbacks.

H. Outdoors--A major part of life involves active outdoor hunting, fishing, camping, etc. This is the predominant avocation.

I. Constant Feeders--Life is one big meal. These people are overweight and tend to stuff themselves at every opportunity.

J. Drinkers--These people are very concerned with alcohol ingestion. They will drink any alcoholic beverage at almost any time of the day; they drink frequently.

These ten dimensions are chosen specifically because they represent a wide range of (hopefully) nonoverlapping lifestyles.

REFERENCES

Bruno, Albert, and Pessemier, Edgar, (1972), "An Empirical Investigation of the Validity of Selected Attitude and Activity Measures," in Proceedings of the Third Annual Conference, Association for Consumer Research.

Carman, James, (1972), "Correlates of Brand Loyalty: Some Positive Results," Journal of Marketing Research, 7.

Day, George, (1969), "A Two Dimensional Concept of Brand Loyalty," Journal of Advertising Research, 9.

Frank, Ronald E., (1967), "Is Brand Loyalty a Useful Basis for Market Segmentation," Journal of Advertising Research, 7.

Frank, Ronald E., (1967), "Correlations of Buying Behavior for Grocery Products," Journal of Marketing. 31.

Frank, Ronald E., DougLas, Susan and Polli, Rolando, (1968) "Household Correlates of Brand Loyalty for Grocery Products," Journal of Business, 41.

Fraser, Cynthia and Wind, Yoram, (1980), "Physical and Social Psychological Anxiety as Correlates of Purchase Behaviors" Working Paper #80-004, The Wharton School, University of Pennsylvania.

Green, Paul E. and Tull, Donald, (1978), Research for Marketing Decisions, 4th ed. Englewood Cliffs, N.J.: Prentice-Hall.

Hustad, Thomas and Pessemier, Edgar, (1971), "Segmenting Consumer Markets with Activity and Attitude Measures," Institute Paper 298, Purdue University.

Jacoby, Jacob, (1971), "Brand Loyalty: A Conceptual Definition," in Proceedings of the American Psychological Association. vol. 6.

Jacoby, Jacob and Chestnut, Robert, (1978), Brand Loyalty Measurement and Management. New York: John Wiley.

Kerlinger, Fred, (1964), Foundations of Behavioral Research New York: Holt, Rinehart & Winston.

Tucker, W. T., (1964), "The Development of Brand Loyalty," Journal of Marketing Research, 1.

Wells, W. D., (1974), "Life Style and Psychographics: Definitions Uses and Problems," in Life Style and Psychographics, ed. W. D. Wells. Chicago: American Marketing Association.

Wells, W. D., (1975), "Psychographics: A Critical Review," Journal of Marketing Research, 12.

Wells, W. D. and Tigert, Douglas, (1971), "Activities, Interests, and Opinions," Journal of Advertising Research, 11.

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