The Effect of Brand Names

Ivan Ross, University of Minnesota
ABSTRACT - The two papers composing this session are discussed. Neither study succeeds in addressing the issue of the importance of brand names. The Zinkhan and Martin paper intended to address this topic, but failed as a result of numerous conceptual and methodological problems. The paper by Rigaux-Bricmont did not address the issue of brand name, per se, because name was confounded with product. The discussant offers some comments on the importance of this topic.
[ to cite ]:
Ivan Ross (1982) ,"The Effect of Brand Names", in NA - Advances in Consumer Research Volume 09, eds. Andrew Mitchell, Ann Abor, MI : Association for Consumer Research, Pages: 478-480.

Advances in Consumer Research Volume 9, 1982      Pages 478-480

THE EFFECT OF BRAND NAMES

Ivan Ross, University of Minnesota

ABSTRACT -

The two papers composing this session are discussed. Neither study succeeds in addressing the issue of the importance of brand names. The Zinkhan and Martin paper intended to address this topic, but failed as a result of numerous conceptual and methodological problems. The paper by Rigaux-Bricmont did not address the issue of brand name, per se, because name was confounded with product. The discussant offers some comments on the importance of this topic.

COMMENTS ON PAPERS

The paper by Zinkhan and Martin, "The Attitudinal Implications of a New Brand Name," has severe conceptual and methodological problems.

The stated purpose was to study "methods used by consumers in responding to a new previously unknown, brand name." This is an important issue, and there has been previous research reported which has explored such methods (e.g. Peterson and Ross, 1972). However, the research reported here does not assess such methods, but rather deals with how subjects evaluated a new brand name in terms of perceived satisfaction along five product attributes.

The conceptual problem can be inferred from the authors' discussion of, "Theoretical Basis and Hypothesis Development," as well as from the, "Implications", sections of their paper. The gist of Hypothesis 1 and 2 appears to be that manufacturers or advertisers need not be concerned about the effect of brand name on behavioral intentions but should be concerned about the effect of brand names on attitudes. The conceptual problem, from this discussant's perspective, is that marketing or advertiser researchers would not apt recognize this distinction, and in fact, do not couch brand name selection in terms of this conceptualization at all. Rather, research will often be directed toward the "fittingness" of a hypothetical new brand name with the product category it is to "point to", as for example, in a free-association design. And, research may be directed toward whether or not a new brand name is perceived by prospects to aptly describe the attribute bundle embodied in the offering, both physical/sensory and/or as described in a "concept statement" or in advertising copy.

But, by definition, how can a new brand name evoke an "attitude" or a "behavioral intention", at least in terms of the way these terms are used in attitude/consumer research? Certainly, various expectations or beliefs may be evoked and measured, for example, does a brand name such as "Computron R-55" sound like a name which a "low price" brand would have? And, for whatever worth it might be, one could assess an "intention"; for example, assuming that price, programmability, etc. were "right", how likely is it that one would buy a Computron R-55? But it is obviously true that such an "intention" would be influenced by whether or not a person already owned a personal calculator, something not ascertained in this research.

There is no argument with the citations in this study concerning the importance of "personality" or "image" of a brand, but it is obvious that such attitudes are formed after experience (real or vicarious), advertising, word-of-mouth, and other such experiences occur. Brands do not have "images" when all that is known about them is the name alone, except in the "fittingness" sense previously discussed.

Finally, a conceptual problem is associated with the notion of "shifting" (attitudes/intentions) away from the neutral level, as stated in Hypotheses 1 and 2. Any brand name or other stimulus which is part of the product/service offering, when exposed to consumers, will have some "meaning", the affective or intentional component of which may range from plus to minus (affect) or high to low (intentions). One may empirically determine that such a stimulus results in a "neutral" affective evaluation and in a "low" or "zero" intentions rating.

But the particular evaluation such a stimulus receives on either dimension is clearly the result of the way it is described to the respondent (presumably, such a description beyond the brand name alone was provided, else how could consumers indicate the satisfactoriness of this "new" brand's price and other attributes), and therefore, the design is conceptually flawed since the general question of the impact of a brand name, per se, is not being studied, but rather the impact of a specific brand name, and as confounded by other information in addition to brand name, is the subject of the research. The very hypothetical brand name, "Computron R-55", chosen for this research is a descriptive name (rather than "coined" or "nonsense") and hence "carries" meaning into the study. The name sounds like other "real" brand names in this product category. For all we know, subjects may have assumed it was "real", or if they were in fact instructed that it was hypothetical, they may still have transferred it to associations springing from attitudes they had toward like-sounding brand names. And, second, what is being "shifted" and why "away" from "neutral"? What "attitude" is it that was "neutral" before being exposed to the brand name? There was no prior attitude, so nothing can be "shifted"!

A final conceptual problem has to do with the authors' attempt to tie this research to the "attitude hierarchy" model. Of course, if the authors mean to say (and conclude) that it's true that an attitude change is more apt to occur after some intervening event than a behavioral change, then I would agree (but in so doing, setting aside passive learning issues), but this is no surprise, and in any event, is not a hierarchally relevant statement in the sense of "which comes first...attitude change or intentions change", since their research design has nothing to do with that question.

Given these conceptual concerns, the numerous methodological problems associated with the execution of the study may deserve little attention. Briefly, however, the following criticisms identify this discussant's major concerns.

With respect to the independent variables used to predict criterion measures (regarding Hypotheses 3, 4, 5 and 6), if it is true, as the authors suggest, that it was essential to the research that a product class be chosen that "fosters high levels of brand involvement or product interest", and if calculators do in fact satisfy that concern, then the four predictors, which really measure involvement and interest, would by definition generate range-curtailed scores. If one really believes these predictors are important in predicting attitudes and intentions, which is what the hypotheses specify, then one must choose an object or object which a priori would substantially vary in these measures. In any event, the reported R values of .27. .42. and .21 in Exhibit 5, are not "fairly substantial", and-one should always try to cross-validate, shrink, or in some other way imbue one's correlated error variance with humility. And multicollinearity remains a problem if any of the pairwise correlations among predictors is greater than .8 or .9 or so, and we do not see a correlation matrix.

Further, the authors do not define how they measured, "experience levels", but however measured, it would appear closely related to "number of purchases." In any case, it is a peculiar thing to measure. If "experience" is at issue,why not measure how much or often a person uses the product, in how many different applications, and so on. And, their measure of familiarity is explicitly a measure of cognitive complexity, which may be a manifestation of familiarity but is not familiarity itself.

The attitude measurement procedures are either unclearly specified or are incorrectly specified, at least according to Fishbein's model. First, behavioral intention is measured with respect to calculators in general while "Aact" is measured with respect to "a particular brand." I don't think you can do that and make sense out of results. Second, behavior intention is measured with respect to buying a calculator while "A-act" is measured with respect to using a particular brand. This is apples and oranges. Third, measuring attribute importance is not the same thing as measuring "consequences." The Fishbein measure of "consequence" has to do with the outcome of buying or using the calculator. Thus, little "a" in the model is no measured correctly. It should reflect an evaluation of ho good or bad the consequence is of performing the behavior, for example, how good or bad is making computation easier? Fifth, Bi should reflect whether an individual believes that performing the behavior (e.g. using the calculator) will lead to a particular consequence (e.g. making computation easier). Sixth, ai was measured in regard to calculators in general and EL regarding the specific new brand, "Computron R-55".

Further, as previously discussed, Figures 3 and 4 really don't fairly test Hypotheses 1 and 2, since beliefs, as the authors measure them, are really the subjects' perceived satisfaction with the new brand on the 5 attributes That none of the 5 attributes contains zero (the neutral midpoint) does not demonstrate that attitudes have been shifted anywhere. First, attitudes were not measured, jus beliefs were, and second, what shifted? What or who moved from what attitude level to what attitude level? Moreover that "attitudes" average around plus 1 on a minus 3 to a plus 3 scale does not seem very "positive". Finally, how can any score on a balanced (plus-minus) scale be compared with a score on a zero-ten intentions scale, with respect to the contrast between Hypotheses 1 and 2?

In sum, it is unclear what the reported research has to do with attitudes or new brand names, whatever else it may shed light upon.

The second paper, "Influences of Brand Name and Packaging on Perceived Quality", by Benny Rigaux-Bricmont, is rather straightforward, but a few comments are in order.

First, the design was not a very powerful one with respect to identifying the impact of regional differences in regard to taste preferences; 90 subjects per region are probably too few. And in any event, it would be important to have developed a rationale for why regional differences might have been expected to play a role. For example, were subcultural issues of concern, language groups, age, rural/ urban, or what? These factors would then have played a role in the sampling plan. and obviously, if regional differences in share was the major problem, a priori logic would cast more suspicion on "non-taste" factors such as differences in distribution strategy by store/chain type, pricing and promotion differences, and so on.

Second, a bit more information on the task confronted by the subjects would have been helpful. Why only women? Were varietal differences identifiable even in the"blinded" treatment by color, granule size, or other observable factors? Were no instructions given in regard to order of use and manner of preparation? Did the samples provided permit multiple preparation? Was it just the female respondent's preferences that were reflected in the subsequent questionnaire or might other householders have had input? Did, in fact, all respondents even prepare and caste all three brands before rendering their evaluation?

Third, a more thorough diagnosis of taste and other sensory parameters might have been undertaken if these factors were seriously considered to play a role in overall evaluation, for example, appearance before and after brewing, bitterness, richness, aftertaste, the detection of additives, and so on.

Fourth, the rationale for the specific package manipulation employed is not clear Why not just two levels, one with brand name on and one with brand name off an otherwise plain package? That would have sufficed to evaluate the role of brand names alone. And then, to address the role of alternative "full tress" (regular) packaging, various levels of that factor should have been crossed with the "name " factor.

Fifth, I don't think the author can flatly conclude that the taste of Brands 1 and 2 are more popular than the taste of Brand 3, because this was not the case in the unidentified treatment. It was only true in the two "brand identified" treatments. Incidentally, this finding seems to me sufficient to discard a "taste" explanation of share differences, whether regionally or nationally. Further confirmation of the lack of respondents' ability to identify varietal differences could have been accomplished by the addition of experimental groups who got each brand but in the "wrong" regular packages. Again, the hypothesis would be that one would discover a package main effect but no main effect or interaction with variety (brands).

Sixth, I was surprised that the data indicated that when brand name was known, and evaluations were related to brands actually bought, that the highest ratings for one' own brand were so low...31% for Brand 1, 25% for Brand 2, and 24: for Brand 3. Even given that ties were excluded, one might expect much higher preference levels for one's own brand, especially in a three-brand choice set. Does this suggest that some people had simply never tasted these other brands? Perhaps a product sampling strategy is called for.

Seventh, and a trivial point, I'm not sure that the least significant differences test employed in the analysis is appropriate since, as equivalent to a multiple t-test, the alpha levels have been inflated. Tukey or Sheffe tests may have indicated fewer differences of import.

In sum, although the study may have been useful in attempting to solve a "real world" problem confronting a coffee processor, it was designed to test primarily whether specific brands are more or less preferred, not whether specific brand names are more preferred. That is, names were confounded with actual product. To fully investigate the role of brand name, a design with names and products crossed would have been required.

THE IMPORTANCE OF BRAND NAMES

Although neither paper really succeeds in addressing the general question posed in this session, "Are brand names important?", the issue remains one worthy of future research efforts.

Certainly, many product names serve to provide explicit positioning for the product (for example, Zest, Gleem, Clearasil, Handy Wrap, Excedrin, etc.) either by suggesting the identity of the product class/function, or by implying some more narrow point-of-difference or benefit. In these cases, the aptness of the brand-name serves as an important anchor for other aspects of the product dress such as package design, performance statements on package or in advertising, product shape/style, and "actual" product (physical/sensory attributes). As the consumer learns these other attributes, the mere mention or sight of the brand name serves to evoke this constellation of other cognitions.

But most brand names are not "suggestive" at all and are initially without product-relevant meaning. Winston, Jantzen, and Granada, for example, come to have meaning as a result of marketing communications efforts of the firm. Whatever meaning they had initially bears little or no relationship to the meaning they come to have as products in the marketplace. In these cases it seems clear that no "name fittingness" research was done nor was any "positioning" intended by the brand name alone.

Research might well be directed at the question of whether the initial "meaning" or lack of it has any bearing upon the speed with which consumers come to learn that name as having "secondary meaning", that is, as being associated with a product/service from a particular source, or if source anonymous, being a "particular" thing with its own identity as separate from other products/services or from the "meaning" of the name before its marketplace-existence. Measures of "meaning" might be taken before market introduction along parameters perhaps not yet well understood, but probably including "fittingness", "distinctiveness" (from other brand names in the product/service category), and "mood" (e.g. affect, intensity, etc.).

Measures beyond "speed of learning" (as a specific source) such as ultimate market success are apt to be difficult to relate to brand name "goodness", per se. Post hoc analysis is too easy, of course, but it is apt to be true that as many "good" as "bad" brand names succeed and fail in the marketplace. If there is any relationship with ultimate success, it is likely to be the result of the fact that companies which expend effort to select the "right" name are also more apt to put appropriate muscle behind the product than those who select "bad" brand names.

Yes, folks, even Edsel might have become a highly regarded brand name had that particular product offering been introduced at a better time and with a different constellation of product attributes.

REFERENCES

Peterson, Robert A. and Ross, Ivan (1972), How to Name New Brands," Journal of Advertising Research, 12, pp. 29-34.

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