Industrial Marketing Management 39 (2010) 1240–1249

Contents lists available at ScienceDirect

Industrial Marketing Management

How co-branding versus brand extensions drive consumers' evaluations of new

products: A brand equity approach
Ali Besharat ⁎
University of South Florida, College of Business Administration, Department of Marketing, 4202 E. Fowler Avenue, BSN 3403, Tampa, FL 33620-5500, USA

a r t i c l e i n f o a b s t r a c t

Article history: Current research into co-branding and brand extensions indicates that these marketing strategies benefit
Received 31 October 2008 firms, yet marketing literature examines the concepts only independently. This article reports the findings of
Received in revised form 3 July 2009 two studies, conducted among 256 students, that compare the effectiveness of co-branding versus brand
Accepted 2 December 2009
extension strategies. The comparison of these strategies, both individually and concurrently, considers
Available online 19 March 2010
consumers' attitudes, quality perceptions, and purchase intentions toward a new product (i.e., Bluetooth-
Keywords:
enabled sunglasses). The first study reveals that the presence of at least one high-equity brand in co-
Co-branding branding strategy suffices to leverage consumers' evaluations of a new product. However, the findings of the
Brand extension second study indicate no significant differences between co-branding and brand extensions in terms of
Brand equity consumer evaluations of an identical product.
Signaling theory © 2010 Elsevier Inc. All rights reserved.

1. Introduction
When Apple launched the iPhone, an Internet-connected smart
phone, its rivals in the cell phone industry worked furiously to introduce
similar prototype models and extend their product lines. Yet they faced
a considerable barrier, in that the iPhone's advanced features dramat-
ically increased phone users' expectations, to the point that meeting
such criteria was unrealistic in the short term. Nonetheless, Google,
which earns profits on the basis of advertising that appears on its online
products and services, partnered with T-Mobile to introduce a first-
generation cell phone, G1 in August 2008. The astonishing 1.5 million
preorders reported in the press indicate that T-Mobile made a smart
move to co-brand with Google to challenge the dominant iPhone.
Unstable environments, dynamic markets, intense competition, and
high costs to enter new markets force companies in various industries to
adopt nontraditional, innovative branding strategies, including co-
branding, in an attempt to exploit their existing brand equity (Desai &
Keller, 2002). This strategy usually provides a tool for differentiation
that leverages brands through the transfer of positive associations, such
as brand-quality, image, or awareness, from one brand to another
(McCarthy & Norris, 1999; Simonin & Ruth, 1998; Washburn, Till, &
Priluck, 2000). Co-branding strategy is defined when two brands jointly
appear on the logo and/or package of a new product (Grossman & Till,
1998). Because the ultimate goal of co-branding is to launch a new
product, it sometimes is referred to as a special case of brand extension
(Park, Jun, & Shocker, 1996). The financial benefits of extending a brand
⁎ Tel.: +1 813 974 6181 (Office); fax: +1 813 974 6175.
E-mail address: abeshara@usf.edu.
within and beyond an original product, such as lower cost advertising
and promotions, therefore usually generalize to co-branding strategy
(Tauber, 1988; Volckner & Sattler, 2006).
Various examples of co-branding span a wide range of product
categories. Kellogg, a cereal company, co-branded with ConAgra to
launch Healthy Choice cereals in 1994. Coca-Cola co-branded with
Diebels, a German beer producer, to market a new fruit beer called
Dimix in 1998. Nike and Apple jointly launched Nike Plus sport shoes,
which feature both brands' logos and a technology that enables
communication between the shoes and a runner's iPod. Even Ferrari,
the infamous Italian sport car manufacturer, and Fila, a sport goods
manufacturer, found the idea appealing and co-branded a series of
sport clothing and shoes from 2001 to 2004. Philips's cool skin electric
shaver is co-branded with Nivea, an international skin and body care
brand. This shaver dispenses Nivea for Men shaving lotion or gel.
No study empirically investigates how such combinations of brands,
with their unique brand equities, in co-branding strategy influence
consumers' judgments of the participating brands and the new product.
Prior studies instead study co-branding between one unknown brand
and one established brand or two well-known brands, without taking
into consideration the possible combinations of their brand equities
(Rao & Ruekert, 1994; Simonin & Ruth, 1998; Voss & Gammoh, 2004).
For example, if Sony, as a high-equity brand, intends to engage in co-
branding with another brand, is its best choice another high-equity
brand, like Nike, or would a low-equity brand produce similar consumer
judgments and market positioning for the new product? If there are no
significant differences in attitudinal factors, perceptions, and future
buying intentions, a high-equity partner such as Sony should select a
low-equity co-branding partner, because it would enjoy more power
during contract negotiations and likely more profit premiums.

0019-8501/$ – see front matter © 2010 Elsevier Inc. All rights reserved.
doi:10.1016/j.indmarman.2010.02.021
 A. Besharat / Industrial Marketing Management 39 (2010) 1240–1249
Furthermore, though researchers investigate the advantages and
disadvantages of brand extension and co-branding strategies, they
typically consider only one strategy, without assessing consumers'
evaluations of alternative strategies (Aaker & Keller, 1990; James,
2005). In particular, no work offers a paired comparison of co-
branding and brand extensions with identical new products. This
study addresses this gap by investigating whether co-branding
strategy, relative to brand extension strategy, significantly alters
consumers' attitudes, quality perceptions, and purchase intentions
toward a new product. For example, if consumers generally favor co-
branding to brand extensions for the same product, companies need
to seek partners that can help them leverage their brand equities by
boosting customers' evaluations in the market.
In the next section, this article describes the conceptual back-
ground and theories relevant to the study of co-branding and brand
extension strategies. A synthesis of previous research leads to a set of
hypotheses. The next section describes the research method and
presents the results, followed by a discussion of the findings with
some managerial implications and further research directions for co-
branding literature.
2. Defining the study scope
2.1. Brand extension definition
With brand extension strategy, companies use an established
brand to launch new products. This approach represents one of the
most frequently employed branding strategies (Aaker & Keller, 1990).
However, the failure rates of brand extensions in many fast moving or
high-tech industries are close to 80% (ACNielsen, 1999), which has
prompted extensive study of the potential determinants of brand
extension success (Bottomley & Doyle, 1996).
2.2. Co-branding definition
Because no globally accepted definition of co-branding exists
(Leuthesser, Kohli, & Suri, 2003), several other terms appear
interchangeable, including brand alliance, composite branding (Park
et al., 1996), ingredient or composite branding (Leuthesser et al.,
2003), multi-branding (DiPietro, 2005), and joint or dual branding
(Levin & Levin, 2000; Rao, Qu, & Ruekert, 1999; Rao & Ruekert, 1994).
Research that attempts to define co-branding (Abratt & Motlana,
2002; Baumgarth, 2004; Washburn et al., 2000) implicitly suggests
three criteria:
• Co-branding should be accompanied by a long-term agreement and
cooperation.
• The name of both brands should appear on the product, logo, or
product package.
• The primary objective is to launch a new product in a new or
existing market.
Previous research indicates the benefits of co-branding for
participating brands. For example, co-branding may facilitate the
transfer of positive brand associations from one brand to another.
McCarthy and Norris (1999) argue that co-branding also provides
quality signals to customers about a new product in the market, such
that quality perceptions of one partner brand influence quality
perceptions about another brand. Customers also perceive an average
quality host brand that partners with a high-quality brand more
favorably. Similarly, Park et al. (1996) suggest that two well-known
brands can achieve a better attribute profile when one of them
extends into a new product category. Levin and Levin (2000) further
assert that co-branding provides a legitimate context for influencing
impressions about the image of one brand through a transfer from the
second brand. When people encounter a co-branded product
1241
marketed by a well-known and an unknown brand, they tend to
assume the unknown brand shares values and images with the well-
known brand.
Research on co-branding also explores the antecedents and
moderators of favorable co-branding evaluations. For example,
Simonin and Ruth (1998) reveal that positive existing attitudes
toward each participating brand lead to favorable judgments about a
new product. They also identify product fit, or the extent to which the
product categories of the partnering brands appear compatible, and
brand fit, or the degree of consistency between the images of the
participating brands, as antecedents of co-branding evaluations.1 The
country-of-origin image of the participating brands also has a
significant impact on attitudes toward cross-border brand alliances.
That is, when customers perceive overall compatibility between the
countries of origin of the allied brands, they adopt a positive attitude
toward the new co-branded product launched internationally. Brand
familiarity may moderate customers' judgments about co-branding
strategy too. When customers are less familiar with partnered brands,
the effect of their prior attitudes decreases (Simonin & Ruth, 1998),
and more familiar brands contribute more to their co-branding
evaluations (Baumgarth, 2004).
Some researchers extend such findings to other contexts or end
users. Dickinson and Barker (2007) suggest that nonprofit organiza-
tions can ally with commercial partners in cause-related marketing to
acquire financial resources. Co-branding concept also may apply to
cultural and artistic products, such as those introduced by multiple
nonprofit art organizations (d'Astous, Colbert, & Fournier, 2007).
Bengtsson and Servais (2005) even examine how co-branding, as a
network of partnerships and trust, may enable industrial markets,
whose end users are industrial buyers rather than individual
customers, to increase their profitability.
2.3. Consumer-based brand equity
Customer-based brand equity receives extensive attention in
marketing literature (Keller, 1993), including investigations into the
effects on brand preferences, purchase intentions (Cobb-Walgren,
Ruble, & Donthu, 1995; Van Osselaer & Alba, 2000), and brand
partnerships (Rao et al., 1999; Rao & Ruekert, 1994; Washburn, Till, &
Priluck, 2004). There are many ways to measure the brand equity of a
company; some researchers assess it at the firm level, others at the
product level, and some at the consumer level. This study conceptua-
lizes brand equity from the customer's point of view and thereby
measures the existence and magnitude of any associations consumers
have with a brand (Keller, 1993). Two underlying dimensions—brand
awareness (recall and recognition of a brand) and brand image
(overall associations a brand has)—thus emerge to measure the latent
construct, consumer-based brand equity. Evidently, brands with high
levels of awareness and favorable and unique associations represent
high-equity brands (Keller, 2003; Yoo & Donthu, 2001).
2.4. Co-branding theories
Co-branding research has used various theories such as information
asymmetry (Voss & Gammoh, 2004; Rao et al., 1999), information
integration (Simonin & Ruth, 1998), concept combination (Levin &
Levin, 2000), associative learning (Washburn et al., 2004), and
associative network memory models (Samu, Krishnan, & Smith, 1999)
to explain underlying mechanisms for this strategy. The research
inquiries in this article can be explored using two theories: associative
learning (Shimp, 1991) and information asymmetry (signaling) theory
(Spence, 1974).
1
In a replication though, Baumgarth (2004) does not find a significant role of product fit
in co-branding evaluations.
1242 A. Besharat / Industrial Marketing Management 39 (2010) 1240–1249
2.5. Information asymmetry (signaling) theory
The basic tenet of information asymmetry theory is that different
parties in a transaction hold different information regarding the
transaction (Spence, 1974). A highly reputable brand might help mitigate
the information difference between the seller and customer by sending
quality signals (Rao & Ruekert, 1994). The brand signal acts as an
indicator that reduces the likelihood of a bad outcome for the buyer
(Montgomery & Wernerfelt, 1992), because the threat of future monetary
and reputation losses usually prevent high-equity brands from behaving
opportunistically and producing low-quality products. Otherwise, con-
sumers will punish the brand by choosing not to repurchase, engaging in
negative word of mouth, or even boycotting the brand.
Rao and Ruekert (1994) propose that brand alliances also provide
credible marketplace signals. Even when an unknown brand partners
with a highly reputable one, consumers can assume the overall quality
of the product, because the presence of the second brand name signals
sufficient quality. Empirical support for this theory comes from, for
example, Levin et al. (1996), who find that well-known, branded
ingredient enhances product evaluations for both unknown and well-
known brands. In their investigation of two types of ingredient
branding strategies—branding the target attribute ingredient with a
new name as a self-branded ingredient (e.g., Tide with its own
EverFresh bath soap) or with an established, well-respected name as a
co-branded ingredient (e.g., Tide with Irish Spring bath soap)—Desai
and Keller (2002) find that the co-branded ingredient, relative to a
self-branded ingredient, evokes better quality perceptions.
2.6. Associative learning theory
Animals make connections among various events that take place in
their environment (Shimp, 1991). Both instrumental learning and
classical conditioning represent examples of associative learning
(Shimp, 1991). In the former, the connection may include a reward,
whereas in classical conditioning, the connection is temporal. These
theories can help explain brand equity effects. For instance, buyers can
be conditioned to form favorable impressions about a constituent
unknown brand in co-branding when they see it allied with a high-
equity partner (Levin, Davis, & Levin, 1996). If learning occurs during
this stage, people tend to associate their perceptions and emotions
about a well-established brand with a second brand. With instrumental
learning, consumers develop positive feelings toward a particular brand
because of their positive past experiences with the brand.
3. Hypotheses
3.1. Consumer judgments of co-branding strategies
According to the signaling theory, familiarity with a high-equity,
well-reputed brand provides a consumer with favorable information
about the desired quality and should prompt a positive consumer
attitude (Van Osselaer & Alba, 2000). Van Osselaer and Alba (2000)
emphasize that consumers base their predictions on product cues, so
they are accurate to the extent that consumers have properly learned
the relationship between the cues and performance. Existing research
also shows that people may judge a brand based upon external cues
such as brand equity when the brand's attributes are presented
ambiguously (Levin & Levin, 2000). By making an alliance with
another high-equity brand, a new high-equity partner can build
positive associations that lead to an enhanced attribute profiles (Park
et al., 1996).
Attitudes and beliefs about a co-branded product in the market are
mainly formed by the integration of customers' attitudes and evalua-
tions about each participating brand in the alliance (Simonin & Ruth,
1998). Ultimately, companies, by having co-branding strategy with
well-known brands, are able to integrate original brand attitude with
their existing brand attitudes. This integration process is similar to any
other sponsorship or celebrity endorsement literature (Dickinson &
Barker, 2007). Therefore, co-branding strategy that combines two high-
equity brands should evoke a better value perception because the
combination adds additional information (Washburn et al., 2004). On
the other hand, low-equity brands, according to an instrumental
learning theory, will not provide enough information to people to
leverage their attitude. Simonin and Ruth (1998) also show that the
attitude toward a co-branded product is accessed in the memory by
sufficiently strong associations and cues from each of the constituent
brands. If the association links in the memory are not strong enough or
do not exist, customers are not able to leverage their judgment about co-
branding strategy when no positive additional information is accessible
(Raaijmakers & Shiffrin, 1981). In other words, adding a low-equity
brand to co-branding strategy will not help the other brand to improve
the customers' judgments. Accordingly, for any particular product,
people should hold a greater positive attitude toward co-branding
strategy having two high-equity brands than co-branding strategy
having a low-equity brand paired with a high-equity brand.
H1a. Consumers' attitudes for a product having two high-equity
brands (high-high) are significantly greater than consumers' attitudes
for the same product produced by a high-equity brand and a low-
equity brand (high–low).
H1b. Consumers' attitudes for a product having two high-equity
brands (high–high) are significantly greater than consumers' atti-
tudes for the same product produced by a low-equity brand and a
high-equity brand (low–high).
According to the associative learning theory, well-known and
familiar brands typically have developed a series of associations in the
consumers' minds. When a brand is allied with an almost unknown or
less familiar brand, people tend to extend the associations' network of
the first brand to the second partner (Washburn et al., 2004). Since
customers always expect high-equity brands to deliver on their
promises regarding the characteristics and quality of their products,
customers perceive it to be very uncommon for a high-equity brand to
threaten its reputation and put its brand value into risk by delivering a
low-quality product. Therefore, consumers typically use brand values
of the participating brands as an information source to make their
judgments about a co-branded product, especially when the attri-
butes of a new product are not testable before a product trial (Rao
et al., 1999). In other words, when customers recognize a high-equity
brand in co-branding strategy, they may use it as a simple decision
heuristic to choose their option among other potential alternatives
(Janiszewski & Van Osselaer, 2000).
Rao and Ruekert (1994) believe that one of the major objectives of
co-branding strategy is to leverage the brand value of the participat-
ing brand(s). That is, co-branding strategy implies a synergistic
alliance by which the sum is greater than the constituent elements.
Therefore, the higher the aggregated brand equity of the partnering
brands in co-branding strategy is, the stronger the signal about the
quality of a new co-branded product will be. Accordingly, one should
expect that the strength of quality cues associated with a co-branded
product having two high-equity brands will be greater than the
strength of quality signals being sent out by a high-equity brand
paired with a low-equity partner.
H2a. Consumers' quality perceptions for a product having two high-
equity brands paired together (high-high) are significantly greater
than quality perceptions for the same product produced by a high-
equity brand and a low-equity brand (high–low).
H2b. Consumers' quality perceptions for a product having two high-
equity brands paired together (high–high) are significantly greater
 A. Besharat / Industrial Marketing Management 39 (2010) 1240–1249
than quality perceptions for the same product produced by a low-
equity brand and a high-equity brand (low–high).
One of the major differences between a high-equity and a low-equity
brand is the fact that a company with a high brand value can charge
premiums prices for its products, without losing customers (Keller,
2003). As it was expressed previously, brand values can be used as
information cues for consumers' judgments and preferences (Aaker &
Keller, 1990). As a result, customers are typically more inclined to pay
high premiums for branded products as opposed to unknown brands,
even if they are unaware of the product attributes. Co-branding is
argued to have a synergistic role on the brand strategy of a company
(Rao & Ruekert, 1994). That is, the brand equity of co-branding strategy
is greater than the sum of to the brand equities of its constituent brands.
Alternatively, Washburn et al. (2004) show that the brand equities of
various co-branding strategies fall in order from highest to lowest as
follows: high–high, high–low (or low–high), and low–low.
Simonin and Ruth (1995) argue that the attitude toward co-
branding strategy is an important determinant of the price judgment
about the co-branded product. Therefore, the higher the attitude
toward co-branding strategy is, the greater the judgment about the
price of a co-branded product will be. Accordingly, co-branding
strategy that has two high-equity brands provides additional
information and evokes better value perceptions about a product
than co-branding strategy with a lower brand equity combination,
given the same product.
H3a. Consumers' purchase intentions for a product having two high-
equity brands (high–high) are significantly greater than purchase
intentions for the same product produced by a high-equity brand and
a low-equity brand (high–low).
H3b. Consumers' purchase intentions for a product having two high-
equity brands (high–high) are significantly greater than purchase
intentions for the same product produced by a low-equity brand and a
high-equity brand (low–high).
However, low-equity brands cannot induce favorable attitudes
and quality perceptions (Van Osselaer & Alba, 2000) and price
judgment (Simonin & Ruth, 1995). Washburn et al. (2004) believe
that a low-equity brand may serve as a discounting cue, which may
cause consumers to rely less on the information retained from it.
Therefore, consumers likely exhibit lower attitudes, quality percep-
tions, and purchase intentions in response to a pairing of two low-
equity brands, because they lack any positive signals about product
characteristics and quality.
H4. Consumers' attitudes for a product having two low-equity brands
(low–low) are significantly less than attitudes for the same product
produced by two high-equity brands (high–high).
H5. Consumers' quality perceptions for a product having two low-
equity brands (low-low) are significantly less than quality perceptions
for the same product produced by two high-equity brands (high–high).
H6. Consumers' purchase intentions for a product having two low-
equity brands (low–low) are significantly less than purchase intentions
for the same product produced by two high-equity brands (high–high).
3.2. Consumer evaluations of co-branding versus brand extensions
Simonin and Ruth (1998) argue that preexisting attitudes influence
overall evaluations of brand alliances, because consumers integrate new
information into their existing beliefs. Because of a synergy effect, the
partners in co-branding strategy may each provide additional informa-
1243
tion, increase product associations, or enhance the overall signal about
the presence of attributes or unobservable quality (Rao et al., 1999; Rao &
Ruekert, 1994). Furthermore, as Hillyer and Tikoo (1995) posit in their
conceptual paper, strong brand equity combinations might enhance
brand evaluations in an alliance because people process more positive
information about the brands. This rationale suggests that in co-branding
strategy, companies are able to provide customers more information
than when they launch a new product independently. The additional
information may serve as a positive decision cue to customers (Van
Osselaer & Alba, 2000). In addition, Washburn et al. (2004) argue that co-
branding strategy increases the brand equity of each participating
brands compared to that of each brand prior to the alliance. Therefore,
co-branding strategy enhances the value perception and improves the
price judgment about each constituent brands after the pairing (Simonin
& Ruth, 1995). Integrated information and signals from co-branding
strategy, regardless of the value of each partner brand, may evoke more
favorable attitudes, quality perceptions, and purchase intentions toward
the new product than would brand extension strategy.
H7. Consumers' attitudes toward a new product in co-branding
strategy are greater than attitudes for each constituent brands
launching the same product independently.
H8. Consumers' quality perceptions toward a new product in co-
branding strategy are greater than quality perceptions for each
constituent brands launching the same product independently.
H9. Consumers' purchase intention toward a new product in co-
branding strategy are greater than purchase intentions for each
constituent brands launching the same product independently.
4. Research methodology
4.1. Participants
The combined participants for both experiments consisted of 256
undergraduate business students, who received extra credit in a
marketing course offered by a public, southern U.S. university. The
participants ranged somewhat in age (18–22 years 76.1%, 23–26
15.3%, 27–30 3.5%, and 35+ years 2.7%, no response 2.4%), with a
mean age of 22.8 years and a standard deviation of 2.9. They were
distributed nearly equally in gender (male 45.9%, female 54.1%). Only
3 subjects left the questionnaire completely blank and were excluded
from the sample. In addition, very few cases involved missing data
(less than 1%); in these cases, the mean substitution method assigned
the mean of the group to those with missing variables.
4.2. Pretests
To determine the appropriate product for a hypothetical co-branding
strategy, various combinations were tested (e.g., swimsuit and watch,
sport shoes and vitamin products, DVD players and GPS devices, and
sunglasses and cell phones). In each combination, respondents (n = 42)
initially read a story describing how two product categories have
formed a new product. Then, they saw the new product on an
anonymous website to get a sense how the new product looks. Subjects
were told that because of some privacy issues, they were not allowed to
see prices and brands of each new product in the study. For instance, in a
sunglasses and cell phone scenario, the experiment indicated that cell
phone technology had paved the way for Bluetooth technology, while
sunglasses design and quality had increased to allow the technology to
be embedded into sunglasses. Then, they were asked to assess the
features and the design of the product on a pre-designed website. On the
website, respondents were able to rotate the product and read some
descriptive information about the basic attributes of the product.
1244 A. Besharat / Industrial Marketing Management 39 (2010) 1240–1249
Subjects in the pretest ranked their perception about the degree of the
compatibility between each set of product categories forming all four
new products. The categories of sunglasses and cell phones categories
were chosen to construct a new fictitious product, Bluetooth-enabled
sunglasses, in this particular study. The results of the first pretest
indicated that this combination seemed realistic to the subjects and that
the participants perceived the two product categories to have good
product fit (Simonin & Ruth, 1998).
A second pretest served to classify each product category
according to its consumer-based brand equity and thereby construct
distinct co-branding combinations. The participants in this pretest
(n = 36) evaluated the brand equity of eight brands from each
category, using nine-item measures drawn from Yoo and Donthu's
(2001) study, which employ seven-point Likert scales. The items
come from a validated instrument in the literature, but this study tests
their reliability again to confirm their relatedness to the brand equity
construct (Cronbach's alpha = .836). The items and scales used to
measure the brand equity in this study are listed in the Appendix A.
Participants used the brand equity scales to indicate the value of each
brand and revealed Ralph Lauren (mean = 5.24) has the highest brand
equity in the sunglasses category, whereas Blackberry (mean = 5.08)
ranks first in the cell phone category. Police (mean= 2.88) and Hitachi
(mean = 3.19) are the lowest equity brands in the sunglasses and cell
phone categories, respectively. Therefore, four different co-branding
combinations appear in the experiment pertaining to the launch of a
new fictitious product, Bluetooth-enabled sunglasses: Ralph Lauren–
Blackberry (high–high equity), Ralph Lauren–Hitachi (high–low),
Police–Blackberry (low–high), and Police–Hitachi (low–low).
4.3. Measures
4.3.1. Consumer attitudes
The operationalization of this variable uses four items, each
measured with seven-point semantic deferential scales. The items
come from Maheswaran and Sternthal (1990). In support of the
reliability of the interitem test, the Cronbach's alpha (.811) exceeds
.70, the standard cut-off point. The items and scales used to measure
this construct appear in the Appendix B.
4.3.2. Product quality perceptions
Four items, each measured on a seven-point semantic differential
scale, assess product quality perceptions. Lim, Darley, and Summers
(1994) provide the items, which produce a Cronbach's alpha of .643,
below the .70 standard but still acceptable for an evolving construct
such as quality perceptions. An additional analysis of this latent
construct to consider the possibility of hidden constructs, using
principle component analysis (i.e., items excluded from the reliability
measure one at a time) reveals no improvement in the Cronbach's
alpha. The Appendix B lists the items that measure this construct.
4.3.3. Purchase intentions
To measure this variable, the study uses three items, each
measured on a seven-point semantic differential scale, derived from
existing literature on purchase intentions (Baker & Churchill, 1977;
Kilbourne, 1986). The Cronbach's alpha (.866) indicates the items
provide good indicators of purchase intentions. Again, the items and
scales appear in the Appendix B.
5. Experiments
5.1. Study 1
The first study investigates the effects of different levels of brand
equity combinations on consumers' attitudes, quality perceptions, and
purchase intentions for a new product. It employs a 2 (sunglasses
brand equity: high or low) × 2 (phone brand equity: high or low)
between-subjects factorial design, such that the participants (n = 84)
are randomly assigned to a treatment condition (for each condition,
n = 21). In each treatment, the first brand belonged to the sunglasses
category, whereas the second brand pertained to the cell phone
category. Participants in every treatment condition hear that two
participating brands have decided to introduce high-tech Bluetooth-
enabled sunglasses, which will enable users to listen to their favorite
music on their cell phone (or any other Bluetooth-enabled devices)
while they are wearing the sunglasses. Then, subjects were able to see
the product's image showing paired brands' names on the product as
well as read the general descriptive information about the new
product in an anonymous website. The price of the co-branded
product was $100 and was kept fixed for all the treatments.
After reading the treatment-specific scenario, participants
responded to a set of questions designed to measure their attitudes,
quality perceptions, and purchase intentions toward the new co-
branded product. As a check on the manipulations, each participant
also responded to a set of questions developed to measure the brand
equity of the co-brand and the partner brands independently. These
brand equity measures come from Yoo and Donthu (2001) and
employ a seven-point Likert scale. Finally, the respondents completed
the last section of the questionnaire to indicate their demographic
factors (e.g., age, gender, race, major, income) and general habits in
terms of listening to music and wearing sunglasses (e.g., how often
they listen to music and wear sunglasses, how likely they would be to
listen to music while wearing sunglasses).
5.2. Statistical results
The results confirm that the manipulation is successful. That is, the
brand equity ratings of co-branding fall in order from highest to
lowest as follows: high-high (mean = 5.86), high–low (mean = 5.04),
low–high (mean = 4.93), and low–low (mean = 3.22).2 Also, the
constituent brands, selected to represent high brand equities in both
product categories, show higher brand equities than partner brands,
assumed to have low brand equities (sunglasses: mean difference= 2.29,
p b .05; cell phone: mean difference= 2.47, pb .05). Because the objective
of this experiment is to detect significant differences among various co-
branding strategies for variables such as attitudes, quality perceptions,
and purchase intentions, appropriate statistical analysis tools include
separate analyses of variances (ANOVAs) or a MANOVA (multivariate
analysis of variance). MANOVA's data assumptions are satisfied, in that
the dependent variables are moderately correlated, the data are
multivariate normally distributed for each treatment, and the equality
of the variance/covariance matrices (Box's test) is available for all
treatments.
The statistical results reveal significant differences in the means
across the treatments (F = 24.3, p b .05). To determine which
dependent variable causes these mean differences, three separate
analyses of covariance (ANCOVA) use each dependent variable as the
model's dependent variable, such that the remaining dependent
variables provide the covariates for that model. 3
No significant mean differences occur in consumers' attitudes in the
high–high versus high–low (mean difference= .234, p N .05) or the
high–high versus low–high (mean difference = .576, p N .05), meaning
that it fails to support H1a and H1b. That is, consumers' attitudes toward
a product branded by two high-equity brands are not significantly
different than those for a product that combines a high-equity brand
with a low-equity brand (whether high–low or low–high). Table 1 lists
the results of the multiple comparisons for the attitude construct.
2
The result of this manipulation test also confirms the findings of a study by Washburn
et al. (2004).
3
The Bonferroni method is chosen to adjust the confidence level for multiple
comparisons.
 A. Besharat / Industrial Marketing Management 39 (2010) 1240–1249 1245

Table 1
Pairwise comparisons of means for brand attitudes, quality perceptions, and purchase intentions.

Treatments comparison Mean difference Standard error Significance

Consumer Quality Purchase Consumer Quality Purchase Consumer Quality Purchase

attitude perception intention attitude perception intention attitude perception intention

Ralph Lauren-Blackberry versus Ralph Lauren-Hitachi 0.234 0.015 1.063 0.227 0.261 0.379 1.000 1.000 .033
(High-High) (High-Low)
Ralph Lauren-Blackberry versus Police-Blackberry 0.576 0.154 1.244 0.226 0.269 0.379 0.077 1.000 0.009
(High-High) (Low-High)
Ralph Lauren-Blackberry versus Police-Hitachi 0.973 1.080 1.304 0.233 0.272 0.382 0.043 0.037 0.006
(High-High) (Low-Low)
Ralph Lauren-Hitachi versus Police-Blackberry 0.341 0.169 0.321 0.219 0.254 0.380 0.740 1.000 1.000
(High-Low) (Low-High)
Ralph Lauren-Hitachi versus Police-Hitachi 0.194 0.538 0.382 0.217 0.248 0.371 1.000 0.200 1.000
(High-Low) (Low-Low)
Police-Blackberry versus Police-Hitachi 0.147 0.369 0.061 0.223 0.255 0.384 1.000 0.916 1.000
(Low-High) (Low-Low)

Similarly, no significant difference emerges for the quality percep-
tions in the high–high versus high–low (mean difference = .015,
p N .05) nor high–high versus low–high (mean difference = .154,
p N .05). Unlike the predictions of H2a and H2b, quality perceptions
for products with two high-equity brands do not differ from those for a
product with one high-equity and one low-equity brand (see Table 1).
However, in support of H3a and H3b, significant differences for the
mean values of purchase intentions are observed. As Table 1 illustrates,
people are more willing to buy a co-branded product having two high-
equity brands than to buy the same product having only one high-equity
constituent brand (high-low condition: mean difference = 1.063,
p b .05; low–high condition: mean difference = 1.244, p b .05).
The tests of H4, H5, and H6 involve comparing the low–low brand
equity condition with the high–high combination for the three
variables of interest. These tests also rely on sets of ANCOVAs for
each variable. Customers' attitudes differ significantly for the high–
high versus low–low treatments, and, specifically, they express
significantly less attitudes toward low brand equity combinations
(mean difference = .973, p b .05). So H4is supported. A significant
difference emerges for quality perceptions in support of H5 (mean
difference = 1.08, p b .05). Consumers' quality perceptions of a
product that combines two low-equity brands are significantly
lower than those for a product having two high-equity brand
combinations. Finally, the mean values of purchase intentions differ
significantly between low–low brand equity combinations and high–
high brand equity combinations (mean difference = 1.304, p b .05).
This result supports H6: A product that consists of two low-equity
brands induces significantly lower purchase intentions than does a
product with two high-equity brands (see Table 1).
6. Discussion
Consumer attitudes do not necessarily change for different forms
of co-branding strategy, having at least one high-equity brand. This
intuitive finding can be explained with a connectionist model
(Janiszewski & Van Osselaer, 2000). Janiszewski and Van Osselaer
(2000) argue that there are two distinctive types of models to explain
brand associations. The first model, Direct Association (DA) model, is
based on the spreading-activation model (Keller, 1993) in which
declarative knowledge is represented as a network of nodes
connected by links in the memory. This model assumes that each
cue can gain predictive values that are independent and additive. The
second model, Least Mean Square (LMS) connectionist model,
assumes that cues compete to acquire predictive values. Unlike a DA
model, the LMS model explains why some additional associations
bundled with a current set of associations may not enhance (or even
sometimes be detrimental to) the positive predictive cues of an
existing brand. When one of the partner brands in co-branding has a
high brand equity, brand names in co-branding strategy may compete
to gain predictive cues. Hence, the associations of one constituent
brand in the joint branding may inhibit the access to any possible
associations of the second ally. That is, when one of the constituent
brands is well-known, the synergy effects may not be applicable to co-
branding strategy. In other words, the inclusion of only one high-
equity brand in co-branding strategy has a ceiling effect and provides
sufficient predictive cues to construct consumers' attitudes. This is a
compelling reason why H1a and H1b are rejected.
On the other hand, low-equity brands do not provide enough
predictive cues to compete. Therefore, a DA model may explain the
effects of two low-equity brands in co-branding strategy. Simonin and
Ruth (1998) believe that attitude toward a co-branded product, when
associations are accessible from the memory, is the outcome of the
additive information acquired from each partner brand. Nevertheless,
a low–low equity co-branding decision may still fail to provide strong
brand associations to customers to construct a favorable attitude.
Therefore, addition of a low-equity brand to another low-equity
partner does not evoke a synergy effect and ultimately does not
change the customers' attitudes toward a new product. This notion
provides a promising rationale in support of H4.
In general, consumers do not expect high-equity brands to produce
low-quality products, even if a firm has allied with a low-equity brand.
Van Osselaer and Alba (2000) believe that there is a direct link between
the learning and the brand equity. That is, high-equity brands predict
product quality for consumers even without learning about the
attributes' information. Similarly, Van Osselaer and Janiszewski (2001)
by using an adaptive learning model explain how brand equity may
influence the quality perceptions about a product. According to them, an
adaptive-learning model leads to a blocking effect. That is, customers
cease any additional information processing after they find a justifiable
cue (e.g., high brand equity) predicting the quality. Therefore, findings
of the current research suggest that having at least one high-equity
brand in co-branding strategy provides sufficient predictive cues to
customers to judge the quality of the new product. This justification
perhaps explains why H2a and H2b are rejected.
However, low-equity brands may not provide additional information
to enhance the overall signal to the consumer about the presence of
attributes or unobservable quality associated with them (Rao et al., 1999).
Specifically, when attributes of a new product are not testable before a
product trial, an experienced product, people do not have any information
source to judge the quality of a co-branded product (Washburn et al.,
2004). H5 is supported because having two high-equity brands in co-
branding strategy can significantly enhance the strength of quality signals
compared to the pairing of two low-equity brands.
Customers are typically more willing to pay higher prices for
branded products than for unknown brands (Keller, 2003). Brands
generally serve two objectives. First, they can be used as extrinsic cues to
represent an aggregate of information about a product to infer quality
perceptions (Rao & Ruekert, 1994; Keller, 1993). Second, brands
1246 A. Besharat / Industrial Marketing Management 39 (2010) 1240–1249
positively influence customers' internal reference prices. Biswas and
Blair (1991) note that customers generally construct an internal
reference price based on their past experience with a stimulus. One
element of past experience is recognition of a brand name. According to
the associative learning theory, even when customers have not had any
sort of direct experience with a product, merely exposure to a brand
name gives them a certain degree of familiarity (Shimp, 1991).
Grewal, Krishnan, Baker, and Borin (1998) similarly show that
brand values positively influence customers' internal reference prices.
That is, the more positive the brand reputation is, the higher the
buyers' expectations about the price would be. Washburn et al. (2004)
show that the brand equities of various co-branding strategies fall in
order from highest to lowest as follows: high–high, high–low (or low–
high), and low–low. Therefore, by integrating the findings of Grewal
et al. (1998) and Washburn et al. (2004), one should expect that the
price judgment of customers about a new co-branded product will
also fall in order from highest to lowest as follows : high–high, high–
low (or low–high), and low–low. Grewal et al. (1998) indicate that
there is a positive relationship between value perception and purchase
intention of a product. Since the price of the co-branded product was
identical for all the experiment treatments ($100), the value
perception of respondents as well as their purchase intention for a
given product should also comply with the same order. This is a
legitimate reasoning why hypotheses H3a, H3b, and H6 are supported.
6.1. Study 2
This experiment attempts to test H7, H8, and H9 empirically by
determining whether co-branding strategy might be more successful
than brand extension. It compares the effectiveness of these two
alternative strategies for the variables of interest (attitudes, quality
perceptions, and purchase intentions) for an identical product. The
experiment consists of four, between-subjects, randomized designs,
each with three levels. The first level features co-branding, whereas
the two remaining levels correspond to the individual constituent
brands that that form co-branding strategy. Participants (n = 130)
were randomly assigned to a single brand condition to evaluate a new
product (Bluetooth sunglasses). The questionnaire booklet contains
the same scenario as in Study 1, but with single constituent brands.
6.2. Statistical results
A global MANOVA test contrasts each treatment, that is, both co-
branding and the single constituent brands. Similar to the Study 1, the
assessment of the MANOVA assumptions shows that the dependent
variables are multivariate normally distributed for each treatment,
and equality of the variance/covariance matrices is available for all of
the treatments.
Insignificant results of the global MANOVA test (F = 2.217, p N .05)
show that none of the dependent variables significantly differ between
co-branding and corresponding brand extensions. Specifically, co-
branding, regardless of the brand equity combination, does not improve
consumers' attitudes toward the new product compared with brand
extensions by each of the constituent brands (F = 3.615, p N .05). This
finding contradicts H7: Consumers' attitudes toward a new product are
just as favorable in co-branding strategy as in brand extension strategy,
regardless of the level of brand equity. Similarly, the results do not
support H8, such that consumers' quality perceptions in co-branding
strategy do not differ significantly from their quality perceptions in
brand extension strategy, regardless of the level of the brand equity
(F = 4.127, p N .05). Finally, co-branding, regardless of the level of brand
equity, does not increase customers' purchase intentions relative to
brand extension by each individual brand, so H9 must be rejected
(F = 2.119, p N .05). Table 2 shows the detail results of the mean
comparisons between co-branding and individual brand extensions for
the attitudes, quality perceptions, and purchases intentions constructs.
7. Discussion
The extant literature suggests that co-branding strategy may
improve a firm's existing product mix (Desai & Keller, 2002), help the
firm to enter into new markets (Abratt & Motlana, 2002), add new
attributes to current brand associations (Park et al., 1996), or reduce
total cost of advertising (Samu et al., 1999). However, findings of this
study counter-intuitively show that this strategy compared to the
brand extension strategy does not leverage the consumer attitudes,
quality perceptions, and purchase intentions toward a new product.
Least Mean Square (LMS) connectionist and Direct Association (DA)
models (Janiszewski & Van Osselaer, 2000) can explain why co-
branding strategy under various brand equity combinations does not
have any advantage over the brand extension strategy in terms of
customers' perception and evaluations of a new product.
The LMS connectionist model asserts that additional associations
for a second partner brand added to the current set of associations
may not boost the positive predictive cues of an existing brand.
Therefore, some of new information or quality associations shared
with a second brand in co-branding strategy may be blocked
(Janiszewski & Van Osselaer, 2000). A high–high equity co-branding
strategy may provide the same amount of predictive cues to
customers as brand extension by either constituent brand may.
Similarly, when consumers' evaluations of a high–low (or low–high)
equity co-branding strategy are compared to those of each partner
brand independently, the ceiling effect may occur (Van Osselaer &
Janiszewski, 2001). That is, customers will stop processing new
information after a certain cue predicts an outcome of an entity.
Hence, when people see a high-equity brand allied with a low-equity
brand, they base their processing merely on the high-equity partner
because it provides sufficient predictive cues about the outcome.
The findings also suggest that a low-equity brand, by itself, is not
less favorably evaluated than co-branding strategy with a high–low
equity combination. This fact can be explained by the current body of
literature identifying the disadvantages of adopting co-branding
strategy. For instance, Janiszewski and Van Osselaer (2000) in their
fourth study show that co-branding strategy may sometimes dilute (or
hurt) the constituent brands' associations. Similarly, Farquhar (1994)
argues that brand alliances create asymmetries that risk diluting brand
associations. Hence, combining a high-equity brand with a low-equity
brand may inhibit some positive associations that would have been
activated by the high-equity partner. Consequently, the interference
by one partner brand's associations with associations of another ally
may lead to a negative synergy, whereby the sum of the parts is less
than the sum of individual effects. In other words, a low-equity brand
may serve as a discounting cue in co-branding strategy, which may
cause consumers to be less inclined to appreciate the value of a high-
equity brand paired with a low-equity one (Washburn et al., 2004).
Various studies have shown that the strength of brand associations
depends on the extent to which predictive cues are accessible from
the memory (e.g. Keller, 1993). Therefore, if brand's associations are
retrievable from the memory, they may be processed independently
and additively (Van Osselaer & Janiszewski, 2001; Janiszewski & Van
Osselaer, 2000). However, Van Osselaer and Alba (2000) believe that
low-equity brands do not provide enough predictive cues to
customers. Therefore, even if the direct association (DA) model is
involved in the evaluation of a low–low equity co-branding decision,
still the lack of strong brand associations may not form a greater
positive evaluation toward a new product compared to brand
extension by either low-equity constituent brands.
Overall, two strategies, co-branding and the brand extension, are
equally perceived in terms of consumers' attitudes, quality perceptions,
and purchase intentions. Therefore, the value of brands pairing in co-
branding strategy does not imply that possibly a more favorable
judgment about a new product will be formed for co-branding versus
the brand extension. These findings will help managers who dominantly
 A. Besharat / Industrial Marketing Management 39 (2010) 1240–1249 1247

Table 2
Direct comparison of co-branding strategy to the brand extension based upon mean of consumers' attitudes, quality perceptions, and purchase intentions.

Treatments Comparison Mean Difference Standard Error Significance

Consumer Quality Purchase Consumer Quality Purchase Consumer Quality Purchase

attitude perception intention attitude perception intention attitude perception intention

Ralph Lauren-Blackberry Ralph Lauren (High) 0.360 0.430 0.726 0.241 0.214 0.387 1.000 0.112 0.193
(High-High) Blackberry (High) 0.034 0.032 0.353 0.236 0.216 0.386 1.000 1.000 1.000
Ralph Lauren-Hitachi Ralph Lauren (High) 0.220 0.234 0.032 0.197 0.179 0.346 0.796 0.580 1.000
(High-Low) Hitachi (Low) 0.264 0.043 0.231 0.200 0.184 0.323 0.570 1.000 1.000
Police-Blackberry Police (Low) 0.317 0.269 0.373 0.169 0.168 0.289 0.192 0.338 0.597
(Low-High) Blackberry (High) 0.355 0.045 0.726 0.164 0.166 0.275 0.098 1.000 1.000
Police-Hitachi Police (Low) 0.433 0.149 0.187 0.180 0.204 0.354 0.054 1.000 1.000
(Low-Low) Hitachi (Low) 0.202 0.238 0.225 0.180 0.198 0.346 0.801 0.588 1.000

associate positive outcomes with co-branding strategy. Given all the
benefits recognized for co-branding strategy including the improve-
ment of brand image, awareness, transfer of positive associations, cost
reduction in manufacturing and advertising, enriching the attributes'
profile, and etc., one should know that co-branding strategy does not
improve the consumers' judgment about a product when it is compared
with a sole brand extension.
8. General discussion and conclusion
The intense competition in the dynamic marketplace and high cost
of investments involved in entering new markets have forced
companies to adopt innovative brand strategies such as brand extension
and co-branding (Desai & Keller, 2002). The extant literature has
discussed co-branding strategy as a key marketing technique for the
firms. It is proposed that this strategy can enable firms to strengthen the
overall quality, corporate reputation, and awareness of the participating
brands. This strategy is known as an efficient way of positioning new
products and a potential source of competitive advantage in the market.
Despite the prevalent assumptions in regard to advantages of co-
branding strategy, the current research extends the growing body of co-
branding literature by examining the effectiveness of co-branding as
well as brand extension strategies in terms of consumers' judgments
with the brand equity approach.
Study 1 investigates whether various co-branding strategies, pairing
partners with either equal or non-equal brand values, may differ in
terms of consumers' perceptions about a new product. Particularly,
study 1 explores whether synergy effects are observable for any brand
partnerships or whether they are merely limited to the specific type of
alliances. The findings suggest that high-equity brands do not benefit
from co-branding strategy for a new product in terms of consumers'
attitudes and quality perceptions. In other words, consumers do not hold
a different attitude and quality perception toward a high-equity brand
allied with a low-equity partner than they hold toward the same brand
paired with another high-equity one. This paper reveals that co-
branding strategy is not fruitful under every condition. This issue
promises appealing managerial implications. If a high-equity brand
seeks co-branding strategy to promptly respond to the market
turbulence, technological limitations, or market entry barriers, it
would be better off to make an alliance with a low-equity partner. This
decision not only does not dilute the consumers' quality perceptions and
attitude towards the high reputable brand, but also offers several
advantages. First, a high-equity brand because of its significance in the
market may demand more profit premiums from a low-equity partner.
Second, a well-known brand has more power in negotiation, meaning
that it can ask for financial incentives (e.g., manufacturing facilities,
advertising budget, sales personnel, and etc.) as well as non-financial
incentives (e.g., training, access to networks of suppliers, access to
customer databases, and etc.) from a low-equity brand.
Nevertheless, co-branding strategy, composed of at least one low-
equity brand, has its own drawbacks. First, the results from the study 1
show that people are more inclined to purchase a co-branded product
that has two high-equity brands than the same product having at least
one low-equity constituent brand. This issue is critical to the profitability
of the firms. Purchase intention is the ultimate outcome of the
consumers' judgment and is the significant predictor of an actual
purchase behavior. Therefore, making an alliance with a low-equity
brand is not a win–win scenario for a well-known brand. A high-equity
brand that intends to involve itself in co-branding strategy with a low-
equity partner should make a trade-off between the profit premiums
and financial incentives (if any) and the amount of monetary value it
may lose. Second, unlike the common wisdom about the overall synergy
effects of co-branding strategy, co-branding strategy between two
unknown brands will not enhance the consumers' perceptions and/or
purchase intentions to compete with any other co-branding strategies
having at least one high-equity partner. It means that managers who
have the propensity to form co-branding strategy should be extremely
cautious about the value of their own brands as well as those of any
potential partners they may consider. If brand values of both partners,
relative to those of other competitors in the same industry, are
considered low, this strategy will not be promising.
Study 2 addresses another research inquiry. To date, the existing
branding literature has treated co-branding and the brand extension
strategies as two independent entities. Although some scholars have
argued that co-branding is a special case of brand extension (Park et al.,
1996), they failed to test the advantages of each strategy over one
another in terms of the consumers' judgments. The second study in this
paper explores the challenges managers may encounter: should a brand
ally itself with a second brand to introduce a new product in the market
or should it launch the product independently? May the association of a
high-equity brand be diluted by pairing with a low-equity brand such
that the overall consumers' evaluations are negatively influenced? The
findings of this paper suggest that no significant statistical differences
exist between co-branding and the brand extension in terms of
consumers' attitudes, quality perceptions, and purchase intentions for
an identical product. This fact alerts brand managers to the con-
sequences of their decisions for co-branding. In particular, when a firm
decides to introduce a new product to the market, regardless of its brand
value, it should extend its own brand rather than searching for a partner
brand with which to co-brand. Because any strategic alliance entails
some behavioral uncertainties and requires additional monitoring, it is
not worth engaging in co-branding strategy unless it produces
significant market share or generates cash flow. Obviously, according
to the findings, companies will not benefit from co-branding strategy in
terms of consumers' evaluations of a product in the short-run.4 Building
positive brand attitudes and favorable quality perceptions requires time
and effort. If a company is seeking a quick boost in sales or improved
consumers' perceptions of a product by making an alliance with another
brand, this research does not suggest any significant outcome for this
particular decision.
4
Washburn et al. (2004) suggest that brand partners in co-branding strategy will
have higher brand equity ratings after than prior to the pairing.
1248 A. Besharat / Industrial Marketing Management 39 (2010) 1240–1249

9. Limitations and further research Appendix A. Customer-based brand equity scale

Despite the contribution of this research to both branding literature

and managerial practices, it contains several limitations that suggest
opportunities for further research. First, the main source of criticism of this
research is perhaps the methodology used. It involves a laboratory setting
where an experiment is carried out. The pretests and actual experiments
are all scenario-based. Although respondents, right after reading a
treatment's scenario, are directed to a website having a simulated 3D
image of the Bluetooth-enabled sunglasses to mimic an actual purchase
experience, this situation is not still realistic. External validity can be
enhanced by future research that tests hypotheses in a retail environment
setting. A more realistic shopping context would allow respondents to
evaluate a product by systematically searching the detail features and
options, yet at the same time, actual customers will be invited to a realistic
shopping environment. Contributing to the artificiality of this research is
also the use of a fictitious product. Although during a pretest respondents
reported that the product seemed realistic and found it to be practical, a
few sample of the respondents do not represent either the general public's
interest or a real world situation. Further research inquires are warranted
to use a real co-branded product in the experiment.5
Second, another limitation pertains to the sampling frame of this
study. Sample groups are limited to only undergraduate students
majoring in business. The demographic profiles of these subjects are
almost homogenous. This fact may create a bias in the responses that
students may provide. Although student samples can be a part of the
potential target market for the new product in this study, they do not
represent a heterogeneous sample of real customers who may be Appendix B. Scales used to measure attitudes, quality perception,
older, have higher income levels, or are less technology savvy. Future and purchase intentions
research should definitely use a probability sampling (e.g., cluster
sampling) for a group of customers to control for the bias associated
with a convenience sample.
Third, the generalizability of this study is also limited by the product
categories and brands chosen. The products categories, sunglasses and
cell phone, are selected according to the level of compatibility between
them and a new product. A different basis for product fit, besides the
physical attributes fit, could produce different results. Future research
should explore circumstances under which two brands with either
similar or dissimilar product categories form an atypical co-branded
product. In addition, future scholars should examine various combina-
tions of attribute profiles for a co-branded product under which one
brand may dilute the second partner. For instance, do customer evaluate
two utilitarian brands more favorably or two hedonic ones? Do
customers search for alignable attributes or non-alignable attributes
for product categories involved in co-branding strategy?
Finally, findings of this study are limited only to co-branding
context. Future research should examine consumers' evaluations in
different contexts such as product bundling or co-branding between
service providers. Another statistical tool, such as conjoint analysis,
might also help future researchers to predict the actual preferences
for co-branded products more effectively.

Acknowledgements

The author sincerely appreciates the insightful critiques from the

anonymous reviewers. The author also thanks Anand Kumar, Uday
Murthy, and Patricia Nickinson for their invaluable comments on
previous versions of this manuscript.

5
The data for this article are collected on March, 2007. Before the data collection
phase, a careful market analysis was carried out such that a real co-branded product
that satisfies the criteria of both studies is selected. No real product that may meet the
expectations of co-branding and brand extensions was found. Ironically, this product is
not a fictitious product anymore. Recently, Oakley, sunglasses manufacturer, has allied
with Motorola to launch their Bluetooth-enabled sunglasses. Therefore, future
research efforts with a new pool of respondents can enhance the external validity of
the current study by using the real product.
 A. Besharat / Industrial Marketing Management 39 (2010) 1240–1249 1249

References Lim, J. S., Darley, W. K., & Summers, J. O. (1994). An assessment of country of origin
effects under alternative presentation formats. Journal of the Academy of Marketing
Science, 22(3), 274−282.
Aaker, D. A., & Keller, K. L. (1990). Consumer evaluations of brand extensions. Journal of
Maheswaran, D., & Sternthal, B. (1990). The effects of knowledge, motivation, and type
Marketing, 54(1), 27−41.
of message on ad processing and product judgments. Journal of Consumer Research,
Abratt, R., & Motlana, P. (2002). Managing co-branding strategies: Global brands into
17(1), 66−73.
local markets. Business Horizons, 45(5), 43−50.
McCarthy, M. S., & Norris, D. G. (1999). Improving competitive position using branded
ACNielsen (1999). New product introduction-successful innovation/failure: Fragile
ingredients. Journal of Product and Brand Management, 8(4), 267−285.
boundary. New York, AC Nielsen BASES and Ernst & Young Global Client Consulting.
Montgomery, C. A., & Wernerfelt, B. (1992). Risk reduction and umbrella branding.
Baker, M., & Churchill, G. A. (1977). The impact of physically attractive models on
Journal of Business, 65(1), 31−50.
advertising evaluation. Journal of Marketing Research, 14(4), 538−555.
Park, C. W., Jun, S. Y., & Shocker, A. D. (1996). Composite branding alliances: An investigation
Baumgarth, C. (2004). Evaluations of co-brands and spill-over effects: Further empirical
of extension and feedback effects. Journal of Marketing Research, 33(4), 453−466.
results. Journal of Marketing Communications, 10(2), 115−131.
Raaijmakers, J., & Shiffrin, R. (1981). Search of associative memory. Psychological
Bengtsson, A., & Servais, P. (2005). Co-branding on industrial markets. Industrial
Review, 88(2), 93−134.
Marketing Management, 34(7), 706−713.
Rao, A. R., & Ruekert, R. W. (1994). Brand alliances as signals of product quality. Sloan
Biswas, A., & Blair, E. (1991). Contextual effects of reference prices in retail adver-
Management Review, 36(1), 87−97.
tisements. Journal of Marketing, 55(3), 1−12.
Rao, A. R., Qu, L., & Ruekert, R. W. (1999). Signaling unobservable product quality
Bottomley, P. A., & Doyle, J. R. (1996). The formation of attitudes toward brand
through a brand ally. Journal of Marketing Research, 36(2), 258−268.
extensions: Testing and generalizing Aaker and Keller's model. International Journal
Samu, S., Krishnan, H. S., & Smith, R. E. (1999). Using advertising alliances for new
of Research in Marketing, 13(4), 365−377.
product introduction: Interactions between product complementarity and promo-
Cobb-Walgren, C. J., Ruble, C. A., & Donthu, N. (1995). Brand equity, brand preference,
tional strategies. Journal of Marketing, 63(1), 57−74.
and purchase intent. Journal of Advertising, 24(3), 25−40.
Shimp, T. A. (1991). Neo-Pavlovian conditioning and its implications for consumer
d'Astous, A., Colbert, F., & Fournier, M. (2007). An experimental investigation of the use
theory and research. Handbook of consumer behavior. New York: Prentice Hall.
of brand extension and co-branding strategies in the arts. Journal of Services
Simonin, B., & Ruth, J. (1995). Bundling as a strategy for new product introduction:
Marketing, 21(4), 231−240.
Effects on consumers' reservation prices for the bundle, the new product, and its
Desai, K. K., & Keller, K. L. (2002). The effects of ingredient branding strategies on host
tie-in. Journal of Business Research, 33(3), 219−230.
brand extendibility. Journal of Marketing, 66(1), 73−93.
Simonin, B. L., & Ruth, J. A. (1998). Is a company known by the company it keeps?
Dickinson, S., & Barker, A. (2007). Evaluations of branding alliances between non-profit
Assessing the spillover effects of brand alliances on consumer brand attitudes.
and commercial brand partners: The transfer of affect. International Journal of
Journal of Marketing Research, 35(1), 30−42.
Nonprofit and Voluntary Sector Marketing, 12(1), 75−89.
Spence, A. M. (1974). Market signaling. Boston: Harvard University Press.
DiPietro, R. B. (2005). The case against multibranding strategy. Cornell Hotel and
Tauber, E. M. (1988). Brand leverage: Strategy for growth in a cost-control world.
Restaurant Administration Quarterly, 46(1), 96−99.
Journal of Advertising Research, 28(4), 26−30.
Farquhar, P. H. (1994). Strategic challenges for branding. Marketing Management, 3(2),
Van Osselaer, S. M. J., & Alba, J. W. (2000). Consumer learning and brand equity. Journal
8−15.
of Consumer Research, 27(1), 1−16.
Grewal, D., Krishnan, R., Baker, J., & Borin, N. (1998). The effect of store name, brand
Van Osselaer, S. M. J., & Janiszewski, C. (2001). Two ways of learning brand associations.
name and price discounts on consumers' evaluations and purchase intentions.
Journal of Consumer Research, 28(2), 202−223.
Journal of Retailing, 74(3), 331−352.
Volckner, F., & Sattler, H. (2006). Drivers of brand extension success. Journal of
Grossman, R. P., & Till, B. D. (1998). The persistence of classically conditioned brand
Marketing, 70(2), 18−34.
attitudes. Journal of Advertising, 27(1), 23−31.
Voss, K. E., & Gammoh, B. S. (2004). Building brands through brand alliances: Does a
Hillyer, C., & Tikoo, S. (1995). Effect of co-branding on consumer product evaluations.
second ally help? Marketing Letters, 15(2), 147−159.
Advances in consumer research, 22, 123−127.
Washburn, J. H., Till, B. D., & Priluck, R. (2000). Co-branding: Brand equity and trial
James, D. (2005). Guilty through association: Brand association transfer to brand
effects. Journal of Consumer Marketing, 17(7), 591−604.
alliances. Journal of Consumer Marketing, 22(1), 14−24.
Washburn, J. H., Till, B. D., & Priluck, R. (2004). Brand alliance and customer-based
Janiszewski, C., & Van Osselaer, S. M. J. (2000). A connectionist model of brand-quality
brand-equity effects. Psychology & Marketing, 21(7), 487−508.
associations. Journal of Marketing Research, 37(3), 331−350.
Yoo, B., & Donthu, N. (2001). Developing and validating a multidimensional consumer-
Keller, K. L. (1993). Conceptualizing, measuring, and managing customer-based brand
based brand equity scale. Journal of Business Research, 52(1), 1−14.
equity. Journal of Marketing, 57(1), 1−22.
Keller, K. L. (2003). Strategic brand management: Building, measuring, and managing
brand equity. Englewood Cliffs, NJ: Prentice Hall. Ali Besharat is a PhD candidate of marketing in the University of South Florida, USA. His
Kilbourne, W. E. (1986). An exploratory study of the effect of sex role stereotyping on research interests include the role of affect/cognition in consumer information processing,
attitudes toward magazine advertisements. Journal of the Academy of Marketing investors' psychology, and brand management. Besides numerous proceedings in leading
Science, 14(4), 43−46. conferences such as American Marketing Association, Academy of Marketing Science,
Leuthesser, L., Kohli, C., & Suri, R. (2003). Academic papers 2 + 2 = 5? A framework for Society of Consumer Psychology, and Society of Marketing Advances, he has two papers
using co-branding to leverage a brand. Journal of Brand Management, 11(1), 35−47. under the review of European Journal of Marketing and Journal of Consumer Psychology.
Levin, I. P., & Levin, A. M. (2000). Modeling the role of brand alliances in the assimilation
of product evaluations. Journal of Consumer Psychology, 9(1), 43−52.
Levin, A. M., Davis, J. C., & Levin, I. (1996). Theoretical and empirical linkages between
consumers' responses to different branding strategies. Advances in consumer research,
23, 296−300.