Creating brand identity: A study of evaluation of new brand names

Abstract

This research focuses on branding strategies for new products. Meaningful brand names (those that suggest a positive attribute or benefit of the product) are typically evaluated more favorably on overall liking and are easier to recall than non-meaningful brand names. The study reported here is the first such investigation of evaluation of brand names over repeated exposure. Of particular interest are differences in the extent to which meaningful versus non-meaningful brand names benefit from repeated exposure. Results of the study show that meaningful brand names continue to be evaluated more favorably than non-meaningful names, even after repeated exposure, but that evaluations of non-meaningful brand names may improve at a greater rate than those for meaningful brand names. The brand names also influence evaluations of quality and other product attributes. Implications of the research for branding strategies in both domestic and international markets are discussed.

Full text

Creating brand identity: a study of evaluation of new brand names
Chiranjeev S. Kohli*, Katrin R. Harich, Lance Leuthesser
California State University, Fullerton, United States
Received 21 October 2003; accepted 31 July 2004
Abstract
This research focuses on branding strategies for new products. Meaningful brand names (those that suggest a positive attribute or benefit
of the product) are typically evaluated more favorably on overall liking and are easier to recall than non-meaningful brand names. The study
reported here is the first such investigation of evaluation of brand names over repeated exposure. Of particular interest are differences in the
extent to which meaningful versus non-meaningful brand names benefit from repeated exposure. Results of the study show that meaningful
brand names continue to be evaluated more favorably than non-meaningful names, even after repeated exposure, but that evaluations of non-
meaningful brand names may improve at a greater rate than those for meaningful brand names. The brand names also influence evaluations of
quality and other product attributes. Implications of the research for branding strategies in both domestic and international markets are
discussed.
D2004 Published by Elsevier Inc.
Keywords: Branding; Brand equity; Name creation; Name testing; Brand image
1. Introduction
The brand name is regarded as an essential part of the
brand and the foundation for marketing communication
efforts (Aaker, 1992). Brand names not only help identify
the product (Friedman, 1985) but also acquire a rich set of
symbols and meanings embodied by products (Levy, 1978).
For successfully developed and managed brand names, the
associations that consumers make with the brand name
creates value, or brand equity, that can be a highly valued
asset to the company that owns the brand (Keller, 1991;
Meyers-Levy et al., 1994). For example, in 2003, Coca-Cola
was estimated to be worth $70.4 billion, McDonald’s $24.7
billion, and Marlboro $22.2 billion (Business Week, 2003).
Although substantial brand equity is built over long periods
of time, a carefully chosen new brand name can lend a
measure of inherent and immediate strength to the brand.
Building strong brands is becoming more difficult. There
has been an explosion in new products, as well as in the
ways to communicate with consumers, from cable channels,
to the Internet, to displays on mobile phones, and product
placements. The number of brands on US grocery shelves
has increased from 15,000 in 1991 to 45,000 in 2001. The
number of advertising messages to which an average US
resident is exposed has increased from 3000 in 1990 to 5000
in 2000 (Aufreiter et al., 2003). Rising above this clutter can
require a significant financial commitment. As these costs
continue to increase, companies need to get smarter about
branding. One cost effective way is to create a brand name
that is inherently strong. Creating a new brand name,
however, is not an easy task. A brand name should appeal to
the customer as well as facilitate trademark registration—
goals that can be conflicting. A case in point is the Internet
site, WebMD. The name immediately conveys medical
services on the Web. Though it is now a registered
trademark, it faced strong opposition from the Registrar of
the US Patent and Trademark Office. The Registrar
contested the name, claiming that it was merely a
combination of two descriptive words—Web and MD—
and as such was non-distinctive (Reiss, 2003).
0148-2963/$ - see front matter D2004 Published by Elsevier Inc.
doi:10.1016/j.jbusres.2004.07.007
* Corresponding author. Marketing Department, College of Business and
Economics, Fullerton, CA 92834, United States. Tel.: +1 714 278 3796;
fax: +1 714 278 7117.
E-mail address: ckohli@fullerton.edu (C.S. Kohli).
Journal of Business Research 58 (2005) 1506 – 1515

It is, then, not surprising that both marketing practitioners
and academics stress the importance of developing bgoodQ
brand names for new products, but are at a loss as to what
indeed constitutes a good brand name. In their efforts to
develop an effective name, marketers inevitably weigh the
use of a bmeaningfulQversus bnon-meaningfulQbrand name.
A meaningful name is one that conveys relevant information
about the product, or an important product attribute, or
establishes a connection between the product and the
product category (Keller et al., 1998). Marketing practice,
as well as theoretical and empirical research, has suggested
that meaningful brand names are easier to remember and
recall than non-meaningful brand names, and also that
meaningful brand names are generally preferred over non-
meaningful brand names (Klink, 2001; Kohli and Suri,
2000).
The dilemma faced by marketers, however, is that
meaningful names are inherently limited to the products
and product categories for which they have meaning,
whereas non-meaningful names can potentially be applied,
or extended, to any product or product category. The
increasingly global nature of many markets requires that
meaningful brand names be translated to achieve consistent
meaning, a problematic task that, even if successful, results
in a plurality of names for the same product. Clearly, then,
what constitutes a bgoodQname will depend on the priority
given to the brand naming objectives. If recognition, recall,
and positive affect are most important early in the
product’s life, a meaningful name is likely to be favored.
If flexibility and adaptability are given higher priority, then
a non-meaningful name is attractive. In the latter case it
may be preferable to accept a longer time horizon to
achieve brand recognition and value. Because both sets of
branding objectives are typically of high importance to
marketers, it would be extremely useful to have a sense of
the tradeoffs involved. That is what our research seeks to
accomplish. In the first such study of its kind, we examine
how consumer evaluations of meaningful and non-mean-
ingful brand names develop over repeated exposures to the
brand names, with a particular focus on differences
between the two.
2. Literature review and study objectives
The literature suggests that academics and marketing
practitioners look at brand names from very different
perspectives. Academic research has typically focused on
the relationship between brand name characteristics (such as
length, distinctiveness, suggestiveness, commonness, and
linguistic features) and factors such as liking, recognition,
and recall (e.g., Pavia and Costa, 1993; Heath et al., 1990;
Vanden Bergh et al., 1987). On the other hand, practitioners
(including the Patent and Trademark Office, trademark
attorneys, and marketing managers) are influenced to a great
extent by brand name classification that has been developed
by the courts. The courts have suggested a hierarchy of five
categories: generic,descriptive,suggestive ,arbitrary, and
coined, in ascending order of their potential for trademark
protection (Melton, 1979). According to this hierarchy,
generic names cannot be registered while coined names are
likely to get the strongest trademark protection. A generic
brand name is the general term used for the product; for
example, bsoap.QA descriptive brand name describes the
product (e.g., bLaser JetQfor a laser printer). A suggestive
brand name evokes the product’s benefit(s), such as
bDiehardQbatteries. An arbitrary brand name is a common
English word that has no apparent relation to the product
category (e.g., bArrowQshirts). Finally, a coined name is a
fictional word, inherently unrelated to any product or
product category. An example of a coined name is bEnronQ
(Cohen, 1986).
If trademark protection was the only concern, marketing
practitioners would prefer to use arbitrary or coined names
(that is, non-meaningful names), allowing brand name
meaning and brand value to accrue over time, based on
consumers’ associations with the name. This approach has
been referred to as the bJuliet Principle,Qafter Shake-
speare’s line bthat which we call a rose, by any other name
would smell as sweetQ(Collins, 1977). This strategy is
based on the idea that words have no intrinsic meaning,
but instead acquire their meaning only through the
associations that we make with them. Thus, if Milton
Hershey had decided to brew beer, and Adolph Coors had
decided to make chocolate, we would perceive the brands
bHershey’s BeerQand bCoors ChocolateQto be entirely
congruent.
The alternative strategy, dubbed the bJoyce PrincipleQ
after author and poet James Joyce, involves choosing a
brand name with desirable phonetic symbolism and mean-
ingful, positive connotations (Collins, 1977). Proponents of
this strategy seek to provide a bjump startQin establishing
favorable brand associations that, in turn, lead to high initial
brand preference, recall and recognition. For example,
bDieHardQbatteries, bMeow MixQcat food, and bSlimFastQ
diet drink all illustrate the Joyce Principle. Prior research
has identified several characteristics of meaningful brand
names that lead to enhanced attitudes towards the brand
(typicality, familiarity and congruence with product cat-
egory, to name a few). For example, Zinkhan and Martin
(1987) suggest that btypicalQnames (names that are typical
of the product category) are perceived more favorably than
batypicalQnames. Their research suggests that typical
brand names may be more successful because consumers
have preconceived notions about what a brand name for a
particular product should sound like. Along the same lines,
Peterson and Ross (1972) conclude that new brand names
that possess congruence with names of existing brands in
the product category may be favored over brand names
that are incongruent with existing brand names. Klink
(2001) imbedded semantic meaning into brand names and
found that consumers preferred products with meaningful
C.S. Kohli et al. / Journal of Business Research 58 (2005) 1506–1515 1507

brand names to products carrying less meaningful brand
names. Another study (Kohli and Suri, 2000) found that
meaningful names rate higher on overall liking than non-
meaningful names, and are easier for consumers to recall.
Thus, the marketing literature provides reasonably consis-
tent support for the Joyce Principle.
If a meaningful brand name can provide immediate
favorable associations, and advantages on such measures as
liking and recall, then this naturally raises the question of
how long it takes, if ever, for non-meaningful names to
bcatch up.QIn attempting to shed some light upon this
question, we draw upon previous research relating to
advertising repetition, mere exposure effects, halo effects,
and cognitive processing, to study how consumers might
respond in systematic ways to meaningful and non-mean-
ingful brand names.
Kohli and Suri (2000) investigated differences in
overall liking and recall of brand names across categories
following the Patent and Trademark Office classification
scheme discussed above, namely, descriptive,suggestive,
arbitrary and coined names. An analysis of their results
indicated that these categories could be collapsed into
two categories, meaningful (descriptive and suggestive
names) and non-meaningful (arbitrary and coined)
because there were essentially no differences in consumer
responses to descriptive versus suggestive names and
arbitrary versus coined names. Therefore, for the present
study, we have adopted the meaningful/non-meaningful
designation.
Our study extends the work of Kohli and Suri (2000)
by examining how evaluations of brand names develop
over time. In a 5-week, multiple-exposure context, we
explored consumer evaluations of four brand names, one
meaningful and one non-meaningful name in each of two
different product categories (all-purpose cleaners and flu
remedies). For each brand, we measured overall liking
(bOverall LikingQ) of the brand after one, three, and five
exposures to the brand. In addition to Overall Liking, we
also measured evaluations on: (1) perceived quality
(bQualityQ), (2) perceived performance on a desirable
attribute suggested by the meaningful brand name in the
category (bAttribute-SQ), and (3) perceived performance on
an attribute that, while desirable, was not suggested by the
meaningful brand name in the category (bAttribute-NQ). To
clarify these latter two measures, the bSQ(bNQ) designation
is based solely on the attribute being suggested (not
suggested) by the meaningful brand name in the category.
By definition, no attribute would be bsuggestedQby a non-
meaningful brand name, and so non-meaningful brand
names played no role in our choice of these two attributes.
To illustrate, CleanAll is the meaningful brand name in the
all-purpose cleaner category, and Attribute-S for the
category is bversatility for use on various surfacesQ(an
attribute suggested by CleanAll). Attribute-N for the
category is bhow gentle it is on your handsQ(a desirable
attribute, but not suggested by CleanAll). Table 1 shows
each brand name and its corresponding Attribute-S and
Attribute-N.
3. Hypotheses
Research on the mere exposure effect (Zajonc, 1968; for
reviews, see Van Beselaere, 1983) suggests that repeated
exposure to new stimuli tends to increase liking of those
stimuli (e.g., Brooks and Watkins, 1989). The mere
exposure effect has been attributed to a wide variety of
contexts and stimuli, including advertising effectiveness and
nonsense words (Zajonc, 1968). In their review of prior
research, D’Souza and Rao (1995) conclude that advertising
repetition should enhance the liking of brand names. The
authors’ work (1995) specifically found that advertising
repetition enhanced preferences for brands, which should be
very close to liking. We therefore hypothesize that:
H1a. Overall Liking of brand names will increase with
repeated exposure.
D’Souza and Rao (1995) found that while advertising
repetition influenced brand name preferences, it did not alter
the evaluation of other product attributes. The authors
suggest that repeated advertising exposure may affect brand
equity only, without influencing the evaluation of product
attributes. But these authors studied preferences for well-
known offerings in a mature category, hotels, and chose
only study participants who were regular users of the hotel
chains included in the study. Accordingly, these study
participants had substantial experience on which to base
their evaluations of product (service) attributes.
Our study involves hypothetical new brands, and all
brand and product related inferences made by our study
participants will be based on exposure to marketing
communication provided by advertisements in the mock
advertisements used in the study. In this context, the
assessments of individual product attributes are likely to
be bhaloedQby global evaluations of the brand names
because study participants will have no opportunity to
experience the product attributes directly. The halo effect
has received considerable attention in the marketing
literature (e.g., Dillon et al., 1984; Beckwith et al.,
Table 1
Brands and product attributes
Product
category
Meaningful
brand
Non-
meaningful
brand
Attribute-S Attribute-N
All-purpose
cleaner
Clean All Alcon Versatility
for use on
various
surfaces
How gentle
it is on
your hands
Flu remedy FluRelief Sequence How
effectively
it relieves
flu symptoms
How good
it tastes
C.S. Kohli et al. / Journal of Business Research 58 (2005) 1506–15151508

1978). The halo effect is an individual’s failure to
discriminate among conceptually distinct and potentially
independent attributes, leading to attribute ratings that
covary more than they otherwise would. A number of
sources of the halo effect have been identified in the
literature. In accordance with cognitive consistency theo-
ries, people strive to maintain a consistent set of beliefs
and attitudes because inconsistency in the cognitive
system is hypothesized to induce adverse psychological
tension. This tendency toward consistency manifests itself
as higher-than-actual correlations between attribute ratings
as individuals attempt to blevel outQdiscrepancies in their
belief structures (Beckwith et al., 1978). Another explan-
ation is that consumers often avoid active processing of
product information. Dubbed bcognitive misersQthey
avoid the intellectual effort of making attribute-by-attribute
evaluations and instead form overall affective impressions
(Burnkrant, 1976; McGuire, 1969; Petty et al., 1983). In
other cases, the halo effect has been observed when
respondents are unfamiliar with the product or product
category, and thus rely on global impressions (Cooper,
1981). Whereas the halo effect is viewed as an error in
connection with attribute measurement, seen from a
branding perspective it is a potential indicator of brand
name strength and brand equity (Leuthesser et al., 1995).
In our study, the use of fictitious brands is an advantage
because it affords us a more bpureQfocus on the effects of
brand name alone, which is our objective. Because we
anticipate that Overall Liking will increase with repeated
exposures, and because we expect the halo effect to be
operating, we further hypothesize that:
H1b. The evaluation of product attributes (Quality, Attribute-
S, and Attribute-N) will increase with repeated exposure.
Findings from earlier studies showed that meaningful
names were evaluated significantly higher on overall liking
than non-meaningful names. Consistent with H1, we expect
this pattern to hold over repeated exposures. Further, because
we expect the halo effect to be operating, we expect that this
pattern will extend to product attributes as well. Therefore:
H2a. Overall Liking will be evaluated more favorably for
meaningful brands than for non-meaningful brands.
H2b. Quality, Attribute-S, and Attribute-N will be eval-
uated more favorably for meaningful brands than for non-
meaningful brands (although the difference may be smaller
for Attribute-N).
We now address the issue of relative improvement in
ratings for meaningful versus non-meaningful brands, over
repeated exposures. The primary reason for choosing a
meaningful brand name is to embed relevant product
information into the brand name itself, where it is expected
to receive top-of-mind processing. Under these circum-
stances, repeated exposures should have less incremental
effect in shaping respondents’ evaluations, compared to non-
meaningful brand names, because information on which to
base the evaluation is immediately available in the brand
name itself. We expect the initial advantage of meaningful
brand names to be strongest on Overall Liking, Quality and
Attribute-S, because these measures should be perceived by
respondents as closely related to the benefit suggested by the
brand name. In other words, beyond halo effect, the brand
name provides relevant information respecting these meas-
ures. On the other hand, evaluations on Attribute-N are likely
to be influenced by halo effects only, regardless of type of
brand name. We therefore hypothesize that:
H3a. On repeated exposure, evaluations of Overall Liking,
Quality and Attribute-S will show greater improvement for
non-meaningful brands than for meaningful brands.
H3b. On repeated exposure, there will be no significant
difference in the improvement of evaluations of Attribute-N
between meaningful and non-meaningful brands.
4. Methodology
The two target product categories included all-purpose
cleaners and flu remedies. Manipulation checks were
conducted on the brand names. First, the face validity of
the names (belonging to the respective categories) was
tested using three independent judges. The appropriateness
of the names was also tested by asking respondents to rate
the brelevance of the name to the product categoryQon five-
point scales. Means for manipulation checks were 4.42
(CleanAll/meaningful) and 1.76 (Alcon/non-meaningful) in
the all-purpose cleaner category, and 4.30 (FluRelief/mean-
ingful) and 1.64 (Sequence/non-meaningful) in the flu
remedy category. Both mean differences were significant
at the 0.01 level.
Respondents were asked to rate the meaningful and non-
meaningful brand names in each category on Overall
Liking, Quality, Attribute-S, and Attribute-N. Overall
Liking was assessed by asking respondents to brate how
much you likeQeach brand. The remaining measures were
obtained by asking respondents to brate how well you think
the product will perform on the attributes listed.QThe
attributes listed for each brand were boverall qualityQand the
two attributes that appear in Table 1. All measures were
obtained using five-point scales.
We conducted the study with approximately 60 under-
graduate students at a major university. All students were
enrolled in one of two summer school business classes. At
the beginning of the summer semester, students were told
that they would participate in a study on how people process
written information and print advertisements. For that
purpose, a mock magazine had been created. The mock
magazine, Campus Watch, included articles relating to
world, business, marketing, and campus news. Emphasis
was placed on topics that were likely to be of interest to
C.S. Kohli et al. / Journal of Business Research 58 (2005) 1506–1515 1509

college students majoring in business. All mock magazines
contained a variety of advertisements, namely those
promoting the brands under investigation plus several
distracter ads. The advertisements had a standard format,
i.e., all ads included a pictorial and displayed the brand
name, and a tagline prominently emphasizing the attribute
suggested by the meaningful brand (Attribute-S). One group
of students (half the students in each class) was exposed to
copies of Campus Watch containing the ads that promoted
the two meaningful brand names under investigation (Clean
All and FluRelief), while the other group of students was
exposed to copies of Campus Watch containing the ads
promoting non-meaningful brand names (Alcon and
Sequence). Other than the brand name itself, ads for the
meaningful and non-meaningful brands in each product
category were identical.
The study was conducted over a period of 5 weeks.
During the first meeting of each week, the Campus Watch
magazines were distributed to the students, who were
instructed to look through the magazine as they would
normally. Students took 5 to 10 min to browse through the
magazines. In the second class meeting of each week (2
days later), questionnaires were distributed.
In weeks 1, 3, and 5 (after one, three and five exposures
to Campus Watch), students rated several brand names
(those under investigation plus a few distracters) on Overall
Liking. They then rated products in a variety of product
categories, including the two target product categories and a
few distracter categories, on the three product attributes
Fig. 1. Sample advertisements.
Table 2
Brand evaluations
Measure Type of Name Mean rating
a
Week 1 Week 3 Week 5
Overall Liking Meaningful 3.92 4.18 4.20
Non-meaningful 2.55 2.90 3.10
Quality Meaningful 3.90 4.10 4.08
Non-meaningful 3.02 3.31 3.53
Attribute-S Meaningful 4.22 4.22 4.06
Non-meaningful 2.97 3.26 3.40
Attribute-N Meaningful 2.63 3.02 3.02
Non-meaningful 2.40 2.29 2.66
a
On five-point scales. Higher numbers reflect a more favorable rating.
C.S. Kohli et al. / Journal of Business Research 58 (2005) 1506–15151510

(Quality, Attribute-S and Attribute-N). In each category,
students were asked to rate a single brand name (either
meaningful or non-meaningful). Finally, students were
asked to rate the magazine itself (for example, on overall
liking and on breadth of coverage, readability and rele-
vance). All measures were obtained using five-point Likert-
type scales, with a higher rating reflecting a more favorable
evaluation.
To maintain credibility of the cover story, the measure-
ments taken in weeks 2 and 4 referred to articles in Campus
Watch. Similar to the procedure described above, students
provided ratings of the articles and the magazine itself.
While the print ads in Campus Watch were always the same
for each of the two groups, the articles were updated each
week. To avoid order and position biases, the ads were
rotated, i.e., placed on varying pages in the magazine (see
Fig. 1 for sample advertisements).
5. Results
All hypotheses were tested using a repeated measures
general linear model with type of brand name (meaningful
versus non-meaningful) as the between subjects factor. The
total of 109 usable repeat-measure observations was
distributed approximately equally by type of name (51
observations for meaningful names and 58 observations for
non-meaningful names). Table 2 shows the means of all
measures obtained in the study; plots of those means
appear in Figs. 2–5.
Between-subjects effects (i.e., overall differences in
evaluations of meaningful versus non-meaningful names)
were tested by examining the significance of differences in
averaged dependent measures (refer to Table 3 ). Within-
subjects effects (i.e., differences in evaluations over repeated
exposures) as well as within-subjects by between-subjects
interactions (i.e., differences in effects of repeated exposure
depending on type of brand name) were tested by examining
the significance of linear contrasts of the dependent
measures (refer to Table 4).
Since the primary logic for hypotheses _Hlt85360098[H1b
and H2b was the halo effect of Overall Liking affecting
evaluations of Quality, Attribute-S and Attribute-N, we
repeated the analyses using Overall Liking as a covariate. If
a halo effect is indeed operating, the significance level is
likely to drop when the covariate is used. The results from
these analyses are reported in the respective rows in Tables 3
and 4.
5.1. Overall liking (H1a)
We hypothesized that Overall Liking would increase with
repeated exposure. The results indicate that this is indeed the
case with a highly significant linear contrast (Overall
Liking, p=0.00; please refer to Table 4 and Fig. 2). Thus,
hypothesis H1a was supported.
5.2. Evaluation of product attributes (H1b)
We hypothesized that the evaluations of product
attributes would increase with repeated exposure. H1b
was partially supported. The results indicate support for H1b
on Quality and Attribute-N ( p=0.00 and p=0.00, respec-
tively; please refer to Table 4 and Figs. 3 and 5) but very
weak support for Attribute-S ( p=0.11; refer to Table 4 and
Fig. 4). This latter result is discussed below in connection
with the repeated exposure effects on meaningful versus
non-meaningful brands. When Overall Liking was used as
a covariate, the results were generally consistent with
Table 3
Between-subjects effects
Source Measure Mean square FSig. With overall liking as a covariate
Mean square FSig.
Type of name Overall Liking 126.69 42.68 0.00
Quality 44.44 24.32 0.00 0.32 0.62 0.43
Attribute-S 74.48 31.44 0.00 1.20 1.61 0.21
Attribute-N 15.81 7.12 0.01 1.51 1.07 0.30
Fig. 2. Mean ratings on overall liking
1
.(
1
Measured on scales of 1 to 5.
Higher numbers reflect a more favorable rating.)
C.S. Kohli et al. / Journal of Business Research 58 (2005) 1506–1515 1511

earlier findings, but there is a drop in significance level as
expected.
5.3. Evaluations for meaningful versus non-meaningful
names (H2a and H2b)
We hypothesized that the overall liking of meaningful
brand names, as well as the evaluation of attributes of
products with meaningful brand names, would be more
favorable than the corresponding evaluations for non-
meaningful brand names. The results indicate strong
support for this hypothesis with differences on Overall
Liking, Quality and Attribute-S significant at p=0.00, and
Attribute-N at p=0.01 (please refer to Table 3 and Figs. 2–
5). The differences were smaller for Attribute-N, as
expected. When Overall Liking was used as a covariate,
the differences between meaningful and non-meaningful
brands on Quality, Attribute-S, and Attribute-N became
insignificant. This lends further support to the premise of a
halo effect.
5.4. Repeat exposure effects for meaningful versus non-
meaningful names (H3a and H3b)
Hypothesis H3a predicted that evaluations of Overall
Liking, Quality and Attribute-S for non-meaningful brands
would show greater improvement with repeated exposure
than the corresponding measures for meaningful brands.
The results show partial support for hypothesis H3a. The
exposuretype of name interactions for both Quality and
Attribute-S were significant ( p=0.04 and p=0.00, respec-
tively). Recall that, overall, evaluations of Attribute-S did
not show significant improvement with repeated exposures
(hypothesis H1b above). This finding, together with the
significant exposuretype of name interaction for Attribute-
S (see Table 4 and Fig. 4) suggest that repeated exposure has
a positive influence on evaluations of Attribute-S only for
non-meaningful brand names. The halo effect was not
expected to have a differential influence on the evaluations
over repeated exposures for the two different types of brand
names. Thus, the use of Overall Liking as a covariate did
not alter the results above.
Fig. 3. Mean rating on quality
1
.(
1
Measured on scales of 1 to 5. Higher
numbers reflect a more favorable rating.)
Fig. 4. Mean ratings on attribute-S
1
.(
1
Measured on scales of 1 to 5. Higher
numbers reflect a more favorable rating.)
Table 4
Within-subjects effects
Source Measure Mean square FSig. With overall liking as a covariate
Mean square FSig
Exposure Overall Liking 9.26 19.72 0.00
Quality 6.53 17.49 0.00 1.06 2.85 0.09
Attribute-S 1.02 2.63 0.11 0.05 0.11 0.74
Attribute-N 5.74 11.25 0.00 1.74 3.35 0.07
ExposureType of Name Overall Liking 1.04 2.22 0.14
Quality 1.57 4.22 0.04 1.75 4.69 0.04
Attribute-S 4.69 12.10 0.00 5.36 12.38 0.00
Attribute-N 0.24 0.47 0.49 0.19 0.37 0.54
C.S. Kohli et al. / Journal of Business Research 58 (2005) 1506–15151512

The lack of support for Attribute-S on H1b and partial
support for Attribute-S on H3a may be because of the
unexpected turn taken by evaluations on Attribute-S for
meaningful names (see Fig. 4). This may be a chance
occurrence. Further replication is needed to get a more
definitive conclusion on this.
Contrary to hypothesis H3a, the exposuretype of
name interaction for Overall Liking was not significant,
indicating that improvement in overall liking with repeated
exposure is not significantly different for meaningful versus
non-meaningful brands. Finally, Hypothesis H3b, which
predicted that improvement in evaluations of Attribute-N
would not be significantly different for meaningful versus
non-meaningful brands, was supported ( p=0.49). Results
appear in Table 4.
6. Discussion
The study provides important insights into how brand
names are perceived and evaluated. The first key finding is
that the evaluation of both meaningful and non-meaningful
brand names improves with repeated exposure. The excep-
tion to this is, for meaningful brand names, ratings of product
attributes suggested by the brand names do not improve with
repeated exposure. It appears that in this case, attribute-
related information is immediately and easily accessible
from the brand name itself. These are important findings,
because they point to the limitations of branding research
that almost exclusively employs a single-exposure context.
The improvement in the evaluations of meaningful and
non-meaningful brand names indicates support for the mere
exposure phenomenon as well as the halo effect. This
parallels prior research findings on brand equity and
indicates that consumers draw inferences about the product
from the brand name, and that these inferences go beyond
purely affective responses to the name itself. Another key
finding is that the affect of a brand name carries over to
perceptions of quality as well as attributes that are suggested
by the brand name, and even to attributes that are not
suggested by the brand name.
In general, the findings suggest that non-meaningful
brand names may be evaluated less favorably than mean-
ingful brand names even after repeated exposure to the
names. However, we observed significantly greater
improvement for non-meaningful names than for mean-
ingful names, lending some support to the Juliet Principle.
We recognize that it may be argued that repeated exposures
benefit non-meaningful brand names more than meaningful
brand names, simply due to a bceiling effect.QGiven that
meaningful brand names are liked fairly well, there is less
room for ratings to go up because respondents are already
using the upper end of the scale. However, we believe that
this can only be a partial explanation, because the trend of
flattening evaluations observed for meaningful brand names
is consistent, even where there is substantial room for
improvement (ratings on Attribute-N barely cross 3 on a
five-point scale).
Ideally, we would like to have extended the time, and
number of exposures, in the study but we were constrained
by the availability of our respondents. Although it goes
beyond conclusions supported directly by the findings of the
study, it can be inferred that a continuing trend of faster
improvement in evaluations of Quality and Attribute-S for
non-meaningful brand names would eventually close the
bgapQwith ratings for the meaningful names. This provides
an interesting opportunity for future research that could
provide a better sense of the trade-off between the
adaptability and flexibility gained by using a non-mean-
ingful brand name versus the greater investment of time and
resources needed to develop it.
Contrary to what we expected, evaluations on overall
liking for non-meaningful brand names did not improve at a
faster rate than for meaningful brand names. However, an
examination of the results shows that, for meaningful
names, ratings at exposures two and three held essentially
constant, whereas ratings for the non-meaningful names
followed a nearly linear, increasing trend (refer to Table 2
and Fig. 2). While this interaction was not statistically
significant at the 0.05 level, the results nonetheless lean in
the direction of support for the Juliet Principle. As above, a
greater number of exposures would have been likely to shed
more light on the question. Finally, as we expected, ratings
on Attribute-N for non-meaningful brand names did not
improve at a faster rate than for meaningful names. For
Attribute-N, neither type of brand name provides any
relevant information, so any improvement in these ratings
should be attributable to halo and mere exposure effects
only. This appears to have been the case.
Fig. 5. Mean ratings on attribute-N
1
.(
1
Measured on scales of 1 to 5. Higher
numbers reflect a more favorable rating.)
C.S. Kohli et al. / Journal of Business Research 58 (2005) 1506–1515 1513

7. Managerial implications
First, the results show that meaningful brand names may
be favored by brand managers with promotional budgets too
limited to support and build a brand name over extended
periods of time. Not only are meaningful brand names
preferred initially, but the halo effect of a meaningful brand
name also appears to carry over to the evaluation of product
attributes. This could be particularly advantageous in
settings, such as supermarket aisles, where the product’s
ability to grab attention would provide an important
competitive advantage.
In contrast, for companies with larger promotional
budgets and longer time horizons, non-meaningful names
are certainly a viable alternative. Non-meaningful brand
names are more flexible than meaningful names, enabling
managers to build any desired image for the brand, and
importantly, permitting them to change the brand’s image
over time. Additionally, non-meaningful brands are likely to
achieve stronger trademark protection. Furthermore, non-
meaningful brand names lend themselves to product line
extension strategies. As an example, bFruitopiaQis limited to
fruit drinks, whereas bGatoradeQcan be extended to any
beverage.
Research on international branding suggests that few
brands are truly global, that the use of global branding
depends on the product category, and that the majority of
companies adapt products to local markets while striving to
be global in order to achieve economies of scale (e.g., Still
and Hill, 1984). The meaningful/non-meaningful distinction
is particularly important in this context. It is virtually
impossible to create standardized, global brand names that
are meaningful. Meaningful names must be adapted to fit
local markets in order to convey the intended meaning.
Unilever’s bSnuggle,Qfor example, is bKuschelweichQin
Germany, bCoccolinoQin Italy, and bMimosinQin France.
All of these names convey the same connotations in their
respective languages (Dibb et al., 1994). However, adapting
brand names in this manner is costly and is wasteful if fewer
resources could have built the desired image using a
standardized, non-meaningful name.
The improvement in evaluations of brand names on
repeated exposure suggests qualified support for the Juliet
Principle. That is, even though a non-meaningful brand
name may be at a disadvantage early in its life, based on our
research it appears that repeated exposure to the brand may
yield incremental gains in brand evaluations that exceed
those for meaningful names. However, this area requires
further investigation before a stronger conclusion can be
offered. Still, even if extensive research were to suggest that
meaningful brand names provide a long-term advantage in
terms of favorable name associations, there are likely to be
many situations in which the flexibility and adaptability of
non-meaningful names would dictate their use. This under-
scores the strategic importance of carefully assessing the
context in which the brand will be used.
8. Limitations and directions for future research
The reactive arrangements—the artificiality of the blab
settingQand the absence of a physical product—may have
led to stronger effects of the brand name as opposed to what
may be expected in the breal world.QNonetheless, the study
does provide insights into the way consumers evaluate
brand names and the nature of inferences they draw from
them. Our research design allowed us to focus pointedly on
the effects of brand name, dissociated from its underlying
product. In the future, however, the effects of repeated
exposure should be examined in more realistic settings, with
actual products, and over a greater number of exposures.
The challenge will be to control for product characteristics
other than brand name so that valid inferences can be made
about the effects of brand name alone.
All the products chosen for this study could be classified
as frequently purchased consumer goods. The study
included both low involvement (all-purpose cleaner) and
high involvement (cold and flu medication) products.
However, the generalizability of the findings may still be
somewhat limited, and further research with other products
would be of value.
References
Aaker DA. Managing brand equity: capitalizing on the value of a brand
name. New York (NY)7The Free Press; 1992.
Aufreiter NA, Elzinga D, Gordon JW. Better branding. McKinsey Q 2003;4
(available online at http://www.mckinseyquarterly.com).
Beckwith NE, Kassarjian HH, Lehmann DR. Halo effects in marketing
research: review and prognosis. Adv Consum 1978:465 – 7.
Brooks III JO, Watkins MJ. Recognition memory and the mere exposure
effect. J Exp Psychol 1989;15(5):968 – 76.
Burnkrant RE. A motivational model of information processing intensity. J
Consum Res 1976;3(June):21 – 30.
Business Week. Brands in an age of anti-Americanism. Bus Week 2003;
69 – 78 [August 4].
Cohen D. Trademark strategy. J Mark 1986;50:61 – 74.
Collins L. A name to conjure with. Eur J Mark 1977;11(5):340– 63.
Cooper WH. Conceptual similarity as a source of illusory halo in job
performance ratings. J Appl Psychol 1981;66(3):302 – 7.
Dibb S, Simkin L, Yuen R. Pan-European advertising—think Europe, act
local. Int J Advert 1994;13:125 – 36.
Dillon WR, Mulani N, Frederick DG. Removing perceptual distortions in
product space analysis. J Mark Res 1984;21(May):184 – 93.
D’Souza G, Rao RC. Can repeating an advertisement more frequently than
the competition affect brand preference in a mature market? J Mark
1995;59(April):32 – 42.
Friedman M. The changing language of a consumer society: brand name
usage in popular American novels in the postwar era. J Consum Res
1985;11(March):927– 38.
Heath TB, Chatterjee S, Russo France K. Using the phonemes of brand
names to symbolize brand attributes. Proc Am Mark Assoc 1990;38 – 42.
Keller, K.L. 1991. Conceptualizing, Measuring, and Managing Customer-
Based Brand Equity. Marketing Science Institute; Report No. 91-123,
October: Cambridge, MA.
Keller KL, Heckler SE, Houston MJ. The effects of brand name
suggestiveness on advertising recall. J Mark 1998;62(1):48 – 56.
Klink RR. Creating meaningful new brand names: a study of semantics and
sound symbolism. J Mark Theory Pract 2001;9(2):27 – 34.
C.S. Kohli et al. / Journal of Business Research 58 (2005) 1506–15151514

Kohli SC, Suri R. Brand names that work: a study of the effectiveness of
different types of brand names. Mark Manag J 2000;10(2):112– 20.
Leuthesser L, Kohli SC, Harich KR. Brand equity: the halo effect measure.
Eur J Mark 1995;29(4):57 – 66.
Levy SJ. Marketplace behavior—its meaning for management. New York7
AMACOM; 1978.
McGuire WJ. The nature of attitudes and attitude change. Gardner Lindzey,
Aronson Elliot, editors. The handbook of social psychology, vol. 3.
Reading (MA)7Addison-Wesley; 1969. p. 136 – 314.
Melton CA. Generic term or trademark? Confusing legal standards and
inadequate protection. Am Univers Law Rev 1979;29:109 – 33.
Meyers-Levy J, Louie TA, Curren MT. How does the congruity of brand
names affect evaluations of brand name extensions? J Appl Psychol
1994;79:46 – 53.
Pavia T, Costa JA. The winning number: consumer perceptions of alpha-
numeric brand names. J Mark 1993;3:85 – 98.
Peterson RA, Ross I. How to name new brands. J Advert 1972;12:29 – 34.
Petty RE, Caccioppo JT, Schumann D. Central and peripheral routes to
advertising effectiveness: the moderating role of involvement. J
Consum Res 1983;10(September):135 – 46.
Reiss S. Annual review. Trademark Report 2003;93(May–June):652 – 4.
Still RR, Hill JS. Adapting products to lesser developed markets. J Bus Res
1984;12:51 – 61.
Van Beselaere N. Mere exposure: a search for an explanation. In: Doise W,
Moscovici S, editors. Current issues in European social psychology,
vol. 1. Cambridge7Cambridge University Press; 1983.
Vanden Bergh B, Adler K, Oliver L. Linguistic distinction among top brand
names. J Advert Res 1987;39 – 44 [August/September].
Zajonc RB. Attitudinal effects of mere exposure. J Pers Soc Psychol
Monogr 1968;9(2):2.
Zinkhan GM, Martin CR. New brand names and inferential beliefs: some
insights on naming new product. J Bus Res 1987;15:157 – 72.
C.S. Kohli et al. / Journal of Business Research 58 (2005) 1506–1515 1515