206 BtoB Brandextensions

Consumer Evaluations of
Brand Extensions:
Can B2B Brands be Extended
into the Consum er M arket?
Maastricht U niversity
Faculty of Economics and Business Administration
September 2004
Master Thesis
International Management Studies
Author: Leon Phang

ID number: 982261
Supervisor: Dr. H ans Ouwersloot
“Branding adds spirit and a soul to what would
otherwise be a robotic, automated, generic
price-value proposition. If branding is ultimately
about the creation of human meaning, it follows
logically that it is the humans who must ultimately
provide it.” – David Aaker
2
Can B2B Brands be Extended into the Consumer Market?
ABSTR ACT
Brand extensions allow companies to leverage the equity in established brands, and
thereby reducing risk associated with launching new products. A plethora of brand
extension studies have been done in recent years. H owever, there is a paucity of
research investigating business-to-business brand extensions. This thesis examines
whether business-to-business brands can leverage their brands in the consumer market
through brand extensions. A new model is developed by combining Aaker and Keller’s
brand extension model with theories from business-to-business branding as well as
other consumer branding concepts, and tested quantitatively to understand how
consumer evaluate brand extensions. The results of the present study show support for
this new model. More specifically, the results indicate that in the context of business-to-
business brand extensions, consumers use the transferability of skills and resources, and
brand concept consistency with the parent brand category as major cues to evaluate
extensions. Innovativeness and corporate social responsibility are also relevant cues. As
a consequence of these findings, branding strategies that stretch business-to-business
brands into the domain of consumer markets can be successful in cases where
consumers perceive a fit with respect to skills and resources, and brand concept, and
when the parent brand is perceived as being innovative and socially responsible.
3
CON TEN TS
Abstract 3
List of tables 8
List of figures 8
CH APTER 1. IN TRODU CTION 9
1.1 Background:Leveraging the Brand Across Sectors 9

1.2 Problem Statement 11
1.3 Contributions of the Study 12
1.3.1 Theoreticalcontribution 12
1.3.2 Practicalcontribution 13
1.4 Thesis Outline 13
PART I. LITERATU RE REVIEW
CH APTER 2. CON SU M ER BRAN D EX TEN SION S 15
2.1 Introduction to the Chapter 15

2.2 Consumer Brand Extensions 15
2.2.1 Brand Extension D efinition 15
2.2.2 Benefits 15
2.2.3 D raw backs 17
2.3 Concepts for Evaluating Brand Extensions 17
2.3.1 Extension reaction 17
2.3.2 Spillover and substitution effects 18
2.3.3 Categoricaland piecem ealevaluation processes 18
2.3.4 Brand-specific associations 19
2.4 Summary 20
CH APTER 3. BU SIN ESS-TO-BU SIN ESS BRAN DIN G 21

3.2 Corporate Branding Decisions 21
3.2.1 Corporate strategy 21
3.2.2 Corporate brand identity 22
4
3.2.3 Corporate brand strategy 23
3.2.4 Corporate brand architecture 24
3.3 Corporate Brand Extensions 25
3.3.1 Innovativeness 25
3.3.2 Corporate SocialResponsibility (CSR) 26
3.3.3 Environm entalconcern 26
3.4 Summary 27
PART II. CON CEPTU AL FRAMEW ORK
CH APTER 4. H YPOTH ESES 29

4.2 Knowledge About the Parent Brand 29
4.3 Attitude Towards the Parent Brand 29
4.4 Innovation 30
4.5 Corporate Social Responsibility 30
4.6 Environmental Concern 31
4.7 Transfer 31
4.8 Brand Associations 32
4.9 Perceived Difficulty 32
4.10 Moderating Factors 33
4.11 Summary 35
CH APTER 5. RESEARCH DESIG N 36

5.2 The H ypothesized Model 36
5.2.1 The originalm odelof Aaker and Keller 36
5.2.2 The current model 37
5.3 Methodology 39
5.4 Questionnaire Development 39
5.4.1 Brands and brand extensions 39
5.4.2 Scaling and m ulti-item scales 40
5.5 Data Collection 42
5.5.1 Pilot-testing 42
5.5.2 Sam pling and response 43
5
PART III. RESEARCH FIN DIN G S
CH APTER 6. AN ALYSIS,RESU LTS AN D IN TERPRETATION S 46

6.2 Reliability, Collinearity and Regression Analyses 46
6.2.1 Reliability assessm ent 46
6.2.2 M ulticollinearity 47
6.2.3 Regression analysis results 51
6.2 H ypothesis Testing 56
6.2.1 H 1:Parentbrand know ledge 56
6.2.2 H 2:Parentbrand quality 56
6.2.3 H 3:Innovative 56
6.2.4 H 4:Corporate SocialResponsibility 56
6.2.5 H 5:Environm entalconcern 57
6.2.6 H 6:Transfer 57
6.2.7 H 7:Brand conceptconsistency 57
6.2.8 H 8:D ifficult 57
6.2.9 H 10:Interaction effectbetw een Transfer and Brand conceptconsistency 58
6.3 Discussion and Interpretation 61
6.3.1 Parentbrand know ledge 61
6.3.2 Parentbrand quality 61
6.3.3 Innovative 61
6.3.4 Corporate SocialResponsibility 62
6.3.5 Environm entalconcern 62
6.3.6 Transfer 63
6.3.7 Brand conceptconsistency 63
6.3.8 D ifficult 64
6.3.9 Interaction effects 64
6.3.10 G eneraldiscussion of the currentm odel 64
6
PART IV. CLOSIN G PERSPECTIVES
7. CON CLU SION S 67
7.1 Summary of the Findings 67

7.2 Conclusions 67
7.3 Limitations and Directions for Future Research 67
7.4 Implications 69
7.4.1 Theoretical implications 69
7.4.2 Managerial implications 69
BIBLIOG RAPH Y 71
APPEN DIX A. QU ESTION N AIRES 78
APPEN DIX B. SPSS OU TPU TS 103
7
LIST OF TABLES
Table 1.1 Companies Formerly Operating in Business-to-Business Markets Only 11

Table 1.2 Expansion to Consumer Markets 11
Table 4.1 H ypotheses and Their References to Literature 35
Table 5.1 Overview of B2B Brands and B2C Extensions 40
Table 5.2 Multiple-scale Items 42
Table 5.3 Response Profile:Sampling Type 43
Table 5.4 Response Profile:G ender 43
Table 5.5 Response Profile:N ationality 43
Table 6.1 Cronbach’s Alpha Coefficients for Multiple-Item Constructs 47
Table 6.2 VIF Scores of Regression Variables Before and After Residual Centering 49
Table 6.3 VIF Scores of Regression Variables After Omitting Interaction Term 50
Table 6.4 Comparison Effects of Original and Replication Studies 53
Table 6.5 Regression Results:Main and Interaction Effects 54
Table 6.6 Full Model at Brand Level 55
Table 6.7 Summary of H ypothesis Testing 59
Table 6.8 Extension Level Means 60
Table 6.9 Perceived Quality of Parent Brands 60
LIST OF FIG U RES
Figure 3.1 An Integrated Approach to Communications 23

Figure 5.1 H ypothesized Model 38
8
CH APTER 1.
IN TRODU CTION
1.1 Background:Leveraging the Brand Across Sectors
The changing market dynamics and heightened competition of the global economy has
amplified the role of brands to an unsurpassed level. Brand marketers seek ways to
achieve growth while reducing both the cost of new product introductions as well as the
risk of new product failure. A popular way of launching new products has therefore been
to leverage the equity of an existing brand into a new sector, market, or product
category – a so-called brand extension.
A brand extension strategy can be beneficial because it reduces the new product
introduction costs and also increases the chance of success (Kapferer, 1994). The
rationale behind brand extensions is simple:when a strong brand has been established,
the brand has moved beyond the functional product into a realm of values. It makes
economical sense to try to deliver the same emotional benefits in a different market
(Mortimer, 2003). Since awareness of a certain brand already exists, costs of launching a
new product will, ceteris paribus, be lower than in the absence of a strong brand. The
main objective of brand extensions is hence to leverage the intangible qualities of a
brand since the functional benefits can generally be imitated (U rde, 1999).
Since brand extensions imply launching new products, a key issue is to what
extent these extensions are successful. A theoretical starting point to this discussion is
the concept of brand orientation, which can be defined as “an approach in which the
processes of the organization revolve around the creation, development, and protection
of brand identity in an ongoing interaction with target customers with the aim of
achieving lasting competitive advantages in the form of brands” (ibid., p. 117-118).
Brand orientation is an additional degree of sophistication to market orientation. In
other words, what is demanded by customers at any given moment in time is not
necessarily the same as what will reinforce the brand as a strategic resource. Market
orientation can therefore be a source of conflict to a brand’s long-term development
when achieving competitive advantage is the aim. Ideally, a firm is both market-oriented
9
and brand-oriented. Keller (2003) argues along the same line concerning brand
extensions –it is not a question of whether a brand should be extended, but rather
where, when, and how it should be extended. Simply put: extend the brand –if it is
possible. Reality shows that companies do this to a large extent:in the 1990s, 81 percent
of new products were brand extensions (Mortimer, 2003), for previously reasons
mentioned. This is not to say that brand extensions are risk-free –it is crucial to know
where the ‘boundaries’ of the brand are. U nderstanding these limits is not a simple
matter, however. As an example, the deodorant brand Lynx made an unsuccessful
attempt to stretch itself into hair care products. On the other hand, U nilever’s razor
brand G illette successfully stretched into after shaves and deodorants. Thus, even if the
product category of the extension is intuitively related to the product category of the
parent brand, there can still be a lack of fit. On the other hand, brand extensions do not
necessarily have to stick to their parent category. The department store chain Marks &
Spencer launched financial services, although it was a totally different area than
retailing. N evertheless, it worked well, because its customers associated the parent
brand and the financial services with trust.
Virtually all discussions of branding are structured in a consumer marketing
context (Aaker, 1996). That is not to say that industrial or business-to-business branding
is not as important and valuable as consumer branding. Some of the world’s most
powerful brands are B2B brands: ABB, Caterpillar, Cisco, Dupont, FedEx, G E, H ewlett
Packard, Intel, and Siemens1 (W ebster & Keller, 2004). An interesting question then is:
what if the business-to-business brand wants to make a stretch into the consumer
market? This may seem confusing, but many now-famous consumer brands have once
been business-to-business brands and now serving both sectors (see tables 1.1 and 1.2).
The global mobile phone brand N okia started out in forestry (B2B) in 1865, and then
began selling rubber boots in the 1960s, and it was not until the 1980s when it started
making mobile phones, for which it is now famous. Other examples include Philips,
Mitsubishi, Microsoft and IBM, three companies that through new innovations began
selling new products to consumers. This historical perspective highlights one fact:
brands are in constant flux. A stretch from B2B to the consumer market is perhaps not
that far-fetched or uncommon.
1
N ote that some of these brands are promoted to end-customers, giving them some characteristics of consumer
brands. Yet, they can be considered as industrial brands with respect to the development of purchase specifications,
the buying process and the actual purchase (which usually made by a B2B buyer) (W ebster & Keller, 2004).
10
TABLE 1.1 Companies Formerly Operating in Business-to-Business Markets Only
Com pany nam e Country Founded Initialbusiness at year 1
Microsoft US 1975 BASIC computer programming language

IBM US 1911* Tabulating machines
Morgan Stanley US 1935 Investment banking
Merrill Lynch US 1907 Stocks and bonds brokerage
Philips N etherlands 1891 Carbon-filament lamps
Vattenfall Sweden 1909 H ydro-electric powerplants
Mitsubishi Japan 1870 Shipping
N okia Finland 1865 Forestry
Source: Com pany w ebsites
TABLE 1.2 Expansion to Consumer Markets
Driver for strategic

Com pany nam e Year N ew offering B2B Offering at tim e expansion to B2C m arket
Microsoft 1981 MS-DOS BASIC computer programming Innovation

language
IBM 1981 IBM 5150 Personal Mainframes, industrial computing Innovation

computer systems
Morgan Stanley 1977 Private wealth Full-service banking Diversification and growth
management
Merrill Lynch 1977 Cash Management Retail brokerage, institutional- and Diversification and growth
Account® investment banking
Philips 1927 Radio Lamps, medical X-ray tubes Innovation

Vattenfall 1996 Electricity, heating, Public utility (domestic electricity) Deregulation
telephony
Mitsubishi 1917 Automobiles H eavy industries Diversification and

decentralization
N okia 1960s Rubber boots Industrial rubber and cables, radio Diversification
technology
Source:Com pany w ebsites
1.2 Problem Statem ent

In order to determine whether a brand extension is able to capitalize on its parent
brand whilst avoiding or minimizing potential disadvantages, it is crucial to understand
how the extensions are evaluated by consumers. The success of a brand extension is
largely determined by how customers evaluate the extension (Klink & Smith, 2001).
N umerous academic studies on brand extensions have been made. A landmark study in
11
this area was made by Aaker and Keller in 1990. In the current study, the issue of brand
extension evaluation will be investigated in a different context, namely that discussed in
the previous section. In other words, this paper tries to answer the problem statement:
Can business-to-business brands be extended into the consum er m arket?
The purpose of this research is also to: (1) determine whether a broad replication of
Aaker and Keller’s (1990) brand extension model is feasible with respect to the current
context, (2) link theory from consumer branding to corporate branding; (3) examine
whether concepts of evaluating brand extensions can be successfully combined to form
an effective model for predicting extension acceptance in the research context; and (4)
determine the relative importance of these concepts.
1.3 Contributions of the Study
1.3.1 Theoreticalcontribution
Aaker and Keller’s (1990) framework has been subject to substantial scrutiny as a
widely replicated study (Bottomley & H olden, 2001), and has also been confirmed in
marketing textbooks. W hile the series of so-called close replications confirm the
generalizability of Aaker and Keller’s (1990) predictions (Bottomley & H olden, 2001),
Barwise (1995) argues that a model which does not stretch beyond Aaker and Keller’s
(1990) choice of brand extensions is of limited scope and therefore of negligible value.
Replications are seen as crucial for empirical generalization (Leone & Schultz, 1980;
Barwise, 1995) and knowledge development (H ubbard et al. 1994; Bottomley & H olden,
2001). Furthermore, it can be argued that those academic research publications that are
peer-reviewed are worthy of replication (Bottomley & Doyle, 1996). Barwise (1995)
contends that a high-quality empirical generalization should be “characterized by its
scope, precision, usefulness and link with the theory”, and usefully defines a domain
which excludes most marketing practice, teaching and exercise.
The current study is a broad replication of Aaker and Keller’s (1990) model in the
sense that it does not generalize across brand extensions2, but does attempt to make a
generalization of the model by expanding the scope of which it has previously studied.
Scope can be defined as the domain (e.g. sectors, countries, situations) of which an
2
That is, it does not use the core brands of Aaker and Keller’s (1990) study.
12
empirical generalization holds. Concerning the scope, this study will examine whether
Aaker and Keller’s (1990) predictions are extra-sectoral, that is, whether parent brand
equity built up in one sector can be leveraged through brand extensions in another
sector.
Besides empirical replication, an additional contribution of the present study is
model development. Bottomley and Doyle (1996) point out “brand concept consistency”
as a better facilitator of brand extension acceptance than product-related similarity
(used in the A&K model). The present study will therefore integrate “brand concept
consistency” with Aaker and Keller’s (1990) to confirm Bottomley and Doyle’s (1996)
proposal.
A last contribution of the present study is to add to the relatively small but
growing supply of business-to-business/industrial branding theory.
1.3.2 Practicalcontribution
It may be of particular interest for managers to know whether their B2B parent
brands can be extended into the consumer market, especially when a strategic
opportunity arises, and consumer brand value can be created and delivered. In such an
event, the findings of the present study could tell managers whether brand extension
would be accepted by consumers. The findings of the study would also pinpoint which
of the brand extension factors they should focus on if an extension strategy is pursued.
1.4 Thesis Outline

The thesis is structured as follows. The first part is a literature review on brand
extension concepts and theories from both a consumer and business-to-business
perspective. This sets the stage for the second part, the conceptual framework. In this
part, hypotheses that are built on the theories of the preceding chapters are presented.
This is followed by the research design of the current study. The third part contains the
research findings, where the empirical data is analyzed, hypotheses tested and
interpreted. The final parts of the thesis are the closing perspectives, where a summary
and conclusion of the findings are given. Implications of the study, both theoretical and
managerial are mentioned, as well as limitations and directions for future research.
13
PART I.
LITERATU RE REV IEW
Chapter 2. Consumer Brand Extensions 15
Chapter 3. Business-to-Business Branding 21
14
CH APTER 2.
CON SU M ER BRAN D EX TEN SION S
2.1 Introduction to the Chapter
The discussion on brand extensions begins with a consideration of a firm’s sources of

growth. As a useful perspective, Keller (2003) proposes the so-called Ansoff’s
product/market expansion grid, which categorizes various growth strategies “according
to whether they involve existing or new products and whether they target existing or
new customers or markets (ibid., p. 577)”. An important point in the context of growth
strategies is that new-product introductions are crucial for a firm to sustain its long-term
competitive advantage (Keller, 2003). N ew-product introductions can, in turn, be
executed in three general ways:(1) as a new brand that is individually designed for the
new product; (2) as an application of an existing brand; and (3) as a combination of a
new brand with an existing brand. H ence, brand extension is a managerial topic.
N evertheless, for the purposes of the current study, the literature review in this chapter
will be of more theoretical nature, focusing on potential advantages and disadvantages,
as well as different theoretical concepts for how consumers evaluate brand extensions.
2.2 Consum er Brand Extensions

2.2.1 Brand Extension D efinition
According to Keller (2003, p. 577), a brand extension is defined as “when a firm

uses an established brand name to introduce a new product”. This is analogous to
approach (2) and (3). Brand extensions are made on an ad hoc basis or according to a
strategy to create a range brand (Aaker, 1996).
2.2.2 Benefits
The benefits of brand extensions is a much-discussed subject. Keller (2003)

distinguishes two kinds of benefits: (1) benefits that relate to the acceptance of the
brand extension, and (2) benefits that relate to the parent brand image. Kapferer (1997)
15
also makes a distinction between brand extensions and their benefits from an
operational point of view, and proposes that brand extensions that are intended to
boost sales should be distinguished from new products that carry brand image and exist
to fuel the brand.
2.2.2.1 Benefits related to brand extension acceptance
Brand extensions allow consumers to draw conclusions and form expectations

about the potential performance of a new product (i.e. the brand extension) based on
their existing knowledge about the brand (Keller, 2003). Provided that a strong brand
name is present, the perceived risk by consumers is substantially reduced when
familiarity and knowledge about the parent brand is present (Keller, 2003; Aaker & Keller,
1990). Benefits of introducing new products also include different ways of achieving
operational efficiencies. A favorable parent brand reduces costs associated with gaining
distribution since retailers are more positive to stock and promote a brand extension
(ibid.). Another benefit relates to marketing communications: since brand awareness
already exists, promotional activities (including introductory and follow-up advertising
and other marketing programs) of a brand extension can be less intensive and thus less
costly than those of a totally new brand and product (Keller, 2003; Kapferer, 1997). Other
efficiencies includes avoiding costly development of brand names, logos, symbols,
packages, characters, slogans, etc. (Keller, 2003).
2.2.2.2 Benefits relating to the parentbrand im age
Brand extensions also have positive spillover effects on the parent brand. Firstly,
extensions can clarify the brand m eaning to consumers and define the boundaries of the
domain in which it competes (Keller, 2003). Second, by improving the favorability of an
existing brand association, adding a new brand association, or a combination of these, a
brand extension can enhance the parent brand image (ibid.). Consistent with this view
are the findings of Morrin (1999), which propose that consumer exposure to brand
extensions will increase parent brand awareness in terms of recognition and recall.
Similarly, Balachander and G hose (2003) find evidence of beneficial spillover effects of
advertising of a child brand, for example a brand extension, on choice of a parent brand.
A third benefit involves brand revitalization—a new or rejuvenated product can be a
16
mean to renew interest and improve attitude towards the parent brand (Keller, 2003;
Kapferer, 1997).
2.2.3 D raw backs
Keller (2003) mentions several drawbacks of brand extensions. First, the image of
the parent brand can be hurt irrespective of the success or failure of the extension. This
happens when the attributes of the extension are seen as inconsistent or conflicting
with the corresponding attributes of the parent brand. Second, brand extensions may
obscure the identification of the brand with its original categories, reducing brand
awareness (Morrin, 1999) and/or diluting the brand meaning. Third, brand extensions
can lead to problems of practical nature, for example a large number of extensions
might confuse or frustrate customers, and there might be problems with retailers being
unwilling to shelf/store all the different extensions. Similarly, Loken and John (1993, p.
79) suggest that “unsuccessful brand extensions can dilute brand names by diminishing
the favorable attitudes that consumers have learned to associate with the family brand
name”.
2.3 Concepts for Evaluating Brand Extensions
2.3.1 Extension reaction
A principle study in the field of brand extensions is Aaker and Keller’s (1990) study
on how consumers evaluate brand extensions. The authors hypothesize that
“evaluations of brand extensions are based on the quality of the original brand, the fit
between the parent and extension categories and the interaction between the two”
(Bottomley & H olden, 2001, p. 494). Despite the fact that this study per se provides no
evidence that a direct relationship between the quality of the parent brand and the
consumer evaluation of the brand extension exist (Aaker & Keller, 1990; ibid.), the
empirical generalizability of Aaker and Keller’s (1990) model is well supported in
Bottomley and H olden’s (2001) secondary analysis, which examines seven replication
studies. Based on their findings, Bottomley and H olden (2001) draw three general
conclusions: (1) The quality of the parent brand and the fit between the parent brand
and the brand extension are key determinants of consumer evaluations of brand
extensions; (2) Consumer’s brand extension evaluations are also determined by (a) the
dimensions of fit (i.e. the complementarity and transferability of assets and skills)
17
between the parent brand and the brand extension, and (b) to what extent consumers
perceive the brand extension is difficult to produce; (3) Cultural differences influence
how brand extensions are evaluated with respect to relative measurement factors.
2.3.2 Spillover and substitution effects
W hile Aaker and Keller (1990) and consequent replication studies provide a
rationale for leveraging parent brand equity through brand extensions, from which
economic profits can be extruded, Balachander and G hose (2003) examine the reciprocal
effect of brand extensions on the parent brand. This effect or productivity is measured
by “brand-choice elasticities”, which measure the increase in choice probability that
results from increase in exposure (ibid., p. 11). The findings of Balachander and G hose
(2003) provide strong support to positive spillover effects from advertising of a brand
extension on choice of a parent brand. This reciprocal spillover effect does, however, not
seem to be symmetrical—that is, forward spillover effects from advertising of a parent
brand on choice of a brand extension are limited.
2.3.3 Categoricaland piecem ealevaluation processes
To understand how consumers evaluate new brand extensions, categorization

theory is a useful concept. It aims at identifying the processes by which consumers form
categories, and assigns certain objects to one category rather than another (Kapferer,
1997). Mervis and Rosch (1981, p. 89) propose that “a category exists whenever two or
more distinguishable objects are treated equivalently”. W hen a new brand extension is
launched, a set of attributes or beliefs in addition to the already existing family or parent
brand image is introduced (Loken & John, 1993). If these attributes or beliefs are
consistent with the parent brand image, an extension is considered to be acceptable
(Kapferer, 1997) or perceived to “fit” the category (Boush & Loken, 1991).
Another concept for attitude formation towards brand extensions is by so-called
“piecemeal”, “analytical” or “computational” processing (Fiske, 1982; Cohen, 1982;
Brooks, 1978), where attitude is “computed” from specific brand extension attributes.
This type of model does not aim to describe conscious evaluation processes (Boush &
Loken, 1991).
Categorical and computational evaluation processes are not mutually exclusive in
any given affective reaction. Fiske and Pavelchak (1986) propose a two-step process of
18
evaluation. In the first step, the consumer attempts to match a brand extension (or some
other new object) with the current category. If categorization is successful, in other
words, if there is a match, the affect that is associated with the category type is applied
to the brand extension and so the evaluation process is complete. If there on the other
hand is a poor match between the category and the brand extension, piecemeal
processes are initiated. Affect is then evaluated through a weighted combination of
attributes.
Even if inconsistency implies that the extension is not “integrated” in the parent
category, an inconsistent brand extension can have a negative impact on the parent
brand by “diluting” specific attribute beliefs that consumers have come to hold about an
established brand name, rather than “diluting” the global affect associated with the
established brand name (Loken & John, 1993). The negative impact of an inconsistent
extension depends on the typicality of the brand attribute at stake. H ence, brand
dilution is an important issue when launching new brand or category extensions.
2.3.4 Brand-specific associations
A brand-specific association is defined as an attribute or benefit that differentiates

a brand from competing brands (MacInnis & N akamoto, 1990). This means that a brand
can be associated with a salient attribute, but this association is per se not strongly
associated with competing brands or the product class as a whole (Broniarczyk & Alba,
1994).
Since the brand association varies depending on the benefits that are sought
within a particular product category, a consumer’s evaluation of a brand extension need
not correspond to evaluation of that brand in its original category (ibid.). Three
conclusions can be drawn from Broniarczyk and Alba’s (1994) research: (1) A perceived
lack of fit between the product category of the parent brand and the proposed
extension category can be overcome if key parent brand associations are salient and
relevant in the extension category; (2) brand-specific associations allow for brand
extensions to unrelated product categories. Brand-specific associations moderate the
role of product category similarity in brand extension judgments; a brand extension is
more preferred in an unrelated category that valued its association than in a similar
category that does not value its associations; and (3) the boundaries for the
appropriateness of a certain brand extension were determined by knowledge about the
incumbent brand.
19
2.4 Sum m ary
A brand can be leveraged by means of brand extensions. Two types of benefits

arise from an extension strategy:(1) an extension is easier accepted among consumers if
the parent brand is known; and (2) an extension can increase parent brand awareness
and positive attitudes. The success or failure of an extension strategy depends highly on
how consumers evaluate the extension. In general, extension acceptance is higher when
there is a perceived “fit” between the parent brand and the extension. This fit can be
based on associations of the extension product category of the, or on the consistency of
the extension with the parent “brand concept”. This, in turn, depends on how the
evaluation process of the consumer is structured.
20
CH APTER 3.
BU SIN ESS-TO-BU SIN ESS BRAN DIN G
W hile consumer branding tends to focus much on marketing areas such as consumer
behavior where emotions, perceptions and behaviors are central, B2B branding is less
concerned with the buyer-seller interface and more focused on functionality and
performance. Brand elements (such as brand names) are associated with offerings that
consistently deliver superior functionality and are hence valuable resources (Anderson &
N arus, 2004). B2B branding and brand building are concepts that are gaining in
popularity.
This chapter begins with a brief introduction to corporate strategy, which provides
the starting point to branding topics such as corporate brand strategy, brand identity
and brand architecture. Following this is a review of theory on corporate brand
extensions.
3.2 Corporate Branding Decisions
3.2.1 Corporate strategy
Ind (1997) proposes that successful corporate branding is derived from Porter’s
well-established analytical model of competitive strategy. In this model, corporate
strategy is defined as the search for a favorable competitive position in an industry in
order to establish a profitable and sustainable position against competitive forces
(Porter, 1998). The core of the Porter model is that strategy is defined according to (1)
competitive scope, and (2) competitive advantage, the main point being that firms must
choose between being different and being the lowest cost producer in a certain industry
(Ind, 1997).
An important concept of the Porter (1998) model is the value chain. Porter (ibid.)
argues that a firm’s competitive advantage cannot be understood by looking at a firm as
a whole. Instead, a firm should be disaggregated into activities –that are performed to
21
design, produce, market, deliver and support its offering (e.g. logistics, operations,
marketing and sales, and service). These activities represent the value chain, and if
deconstructed and analyzed properly can constitute a source of competitive advantage.
To link the Porter model of corporate business strategy to corporate branding
strategy, it is useful to first discuss the concept of corporate brand image.
3.2.2 Corporate brand identity
Brand identity is, in general, a set of brand associations that the brand strategist
aspires to create or maintain (Aaker & Joachimsthaler, 2000). Ind (1992, p. 19) states, with
respect to corporate brand identity, that:
“Identity is formed by an organization’s history, its beliefs and philosophy, the nature of its
technology, its ownership, its people, the personality of its leaders, its ethical and cultural values
and its strategies.”
In addition, Keller and Aaker (1997) argue that corporate brands may be more likely to
possess intangible attributes or organizational characteristics that span product classes,
compared to product brands, whose associations are more likely to be product specific.
Anderson and N arus (2004), on the contrary, state that corporate brand elements
may not enclose any intangible meaning –instead they are charged with meaning
through the performance of the supplier and its market offering over time. The value of
the brand, i.e. the brand equity, hence lies in the association that business customers or
suppliers have with offerings that consistently deliver superior functionality and
performance (Anderson & N arus, 2004).
The issue of tangibility versus intangibility is clarified by Mudambi (2002). H er
findings conclude that the importance of intangible or tangible industrial brand
attributes, respectively, is situation- and buyer-dependent. In addition, Mudambi (2002)
proposes that B2B branding in general is more important than commonly believed.
Michell et al. (2001) conclude that tenets of consumer brand equity theory such as
perceived quality, a recognizable image, market leadership, and a differentiated position
in the marketplace are also relevant among industrial buyers. Factors that are associated
with building brand loyalty in consumer markets (i.e. quality, reliability, and
performance) are also relevant in B2B markets.
22
3.2.3 Corporate brand strategy
The choice of business-to-business brand strategy is derived from the fact hat
industrial marketing and buying is increasingly focused on relationships rather than
individual transactions (W ebster & Keller, 2004). Industrial customers want ongoing
relationships with reliable suppliers of quality products and services. In a business-to-
business context, a brand is a relationship between buyer and seller.
Abratt (1989) presents a model (figure 3.1) which provides with a structure for
looking at a corporate brand. The underlying thought behind the model is to view
corporate brand management in a holistic point way: external and internal marketing
communications are interrelated. In other words, there must be a “fit” between the
corporate brand identity, the employees’view of identity (the internal environment), the
marketing communication, and stakeholders (the external environment).
FIG U RE 3.1 An Integrated Approach to Communications
M arketing
Identity com m unications Im age
IM AG E IN TERFACE
AU DIEN CES
Corporate mission Marketing Customers

communications
strategy G overnment
Corporate philosophy Local communities
Financial
Internal
Core values marketing Suppliers
Buyers
Corporate culture Employees’ Media

view of identity
Influential groups
Personal
communications
Products and services
Source:Abratt(1989)
23
An important aspect of this model are the so-called ‘feedback-loops’, which
suggest that the “marketing communications strategy, although based on the reality of
the identity, is a dynamic force” (Ind, 1997, p. 51). This implies that although the
corporate image and employees’ view of the corporate brand identity are transmitted
from the corporate identity via marketing communications, there are also backflows of
information (‘feedback’) from customers and other audiences to employees and the
marketing communications. The corporate identity and image are hence in a constant
flux and adjustment, in order to achieve a sustainable advantageous position for the
organization.
Consistent with Abratt’s (1989) idea of feedback loops is G riffin’s (2002) view that
all corporate branding strategy decisions stem both external and internal factors.
External factors are, theoretically, any factors that affect a corporate brand. Brand
strategy decisions try to reduce the uncertainty that is caused by external factors – brand
strategy is ‘outside-in’. Internal factors are the corporate values and culture, as well as
business exposure3. The more exposure a company has, the greater is the need for
consistent communications to stakeholders. H ence, brand strategy is also ‘inside-out’.
3.2.4 Corporate brand architecture

Brand architecture is defined by Rajagopal and Sanchez (2004, p. 236) as “the
organizing structure of the brand portfolio that specifies brand roles and the nature of
relationships between brands and different product-market brand contexts”. Aaker and
Joachimsthaler (2000) describe it as the medium by which the brand team functions as a
unit to create synergy, clarity, and leverage. A brand portfolio includes all the brands and
sub-brands attached to product-market offerings, including co-brands with other firms
(ibid.).
Very little academic research has been conducted with respect to brand
architecture, and there is to date no research in the context of business-to-business
branding. W ebster and Keller (2004) do however suggest that through observation,
industrial brands often use branding strategies that organize their brand portfolio under
a company “blanket brand” consistent with Aaker and Joachimsthaler’s (2000)
definitions of sub-brands and branded house strategies. Sub-brands are brands
connected to a master or parent brand and augment or modify the associations of that
master brand. H ence, the parent brand is the primary frame of reference. In a branded
3
The extent to which management and employees are ‘visible’.
24
house strategy, the role of the parent brand is to be a dominant driver across multiple
offerings. The difference between a sub-brand strategy and branded house strategy is
that in the later the role of the parent brand is more prominent. As Aaker and
Joachimsthaler (2000) describe it, it is like “putting a lot of eggs in one basket”.
3.3 Corporate Brand Extensions
Keller and Aaker (1997) examined how various types of corporate marketing
activities (communication activities portraying a firm as innovative, environmentally
concerned, and involved with the community) would influence corporate credibility (i.e.
perceived expertise, trustworthiness, and likeability) and thus have a positive effect on
brand extension evaluation. In their study, four hypothetical corporate brand extensions
outside the current brand offering were presented alongside corporate descriptions that
emphasized one of the following three types of attributes: (1) A firm’s reputation of
being innovative and philosophy of launching technologically advanced products; (2) a
firm’s policy to offer “environm entally friendly” products and to manufacture products in
an environmentally safe fashion; and (3) a firm's philosophy to im prove the quality of life
in localcom m unities through various activities and programs.
The findings of Keller and Aaker (1997) suggest that corporate marketing efforts
can be beneficial as it improves perceptions and evaluations of a corporate brand
extension. Creating a positive corporate image and executing a corporate brand
strategy can thus facilitate new product acceptance. The three types of brand attributes
mentioned above can be categorized into innovativeness, corporate socialresponsibility
and environm entalconcern.
3.3.1 Innovativeness
An innovative brand image involves being perceived as being modern and up-to-
date, investing in research and development, utilizing state-of-the-art manufacturing
technologies, and introducing the latest product features (Keller, 2003). Though this
suggests that positive brand attributes that signal innovativeness are important,
surprisingly little research has been done in this field. Indeed, most research in the topic
of innovativeness in marketing has been in the area of consumer innovativeness and the
innovation diffusion4 (Roerich, 2004). In this line of research it is argued that different
4
For example, the rate of adopting new technologies.
25
individuals have different predispositions to adopt or buy new products (ibid; Ostlund,
1974). An important point to make is that although innovative brand attributes is
favorable for building brand equity, the extent to which this is successful also depends
on how “innovative” the target audience is.
Marketing activities that emphasized innovation have a significant impact on
corporate brand extension evaluation as it leads to the favorable perceptions of
corporate expertise and to presumptions that the corporate brand extension will also be
innovative (Keller & Aaker, 1997). Emphasis on innovation is the only type of marketing
activity that enhances the customer’s perceived “fit” of the brand extension to the
parent brand, as well as evaluations of specific product attributes. H ence, marketing
efforts to emphasize innovation significantly increases both perceived quality and
purchase likelihood score for the brand extension.
3.3.2 Corporate SocialResponsibility (CSR)

Much has been written about corporate social responsibility (CSR) in recent years.
Kitchen (2003) argues that companies undeniably have responsibilities within their
surrounding community, and that these responsibilities must be clarified and aligned
with the companies’ core businesses. Since these responsibilities are relationships and
promises, CSR is ultimately a function of the brand (ibid.). H ence, Kitchen (2003) defines
CSR as:
“The brand-specific duties and resultant actions of commercial organizations in relation to their
communities of need – defined and delivered outside the core transactional context of the
business” (ibid, p. 10).
Similarly, Keller and Aaker (1997) define CSR as “a firm's philosophy to improve the
quality of life in local communities through various activities and programs”
3.3.3 Environm entalconcern

Keller and Aaker (1997) define environmental concern as “a firm’s policy to sell
"environmentally friendly" products and to manufacture products in an environmentally
safe fashion”. Corporate marketing efforts that emphasize environmental concern
enhance perceptions of corporate trustworthiness and likeability as well as inferences
that the corporate brand extension is environmentally aware.
26
3.4 Sum m ary
Business-to-business branding is buyer-focused and corporate brand image is
more likely to span product classes. H ence, much focus of current literature is on
corporate brand identity and communication of intangible brand attributes.
Observations indicate that product or service portfolios of B2B companies are often built
around one “blanket brand” consistent with branded-house and sub-brand strategies.
Virtually no academic research in the field of corporate brand extensions has been
published to date. Keller and Aaker (1997) did however examine in their working paper
how corporate marketing activities portraying a firm as innovative, environmentally
concern, and involved with the community might lead to positive corporate credibility
and thus have a positive effect on brand extension evaluation. Only perceived
innovativeness had a significant impact on corporate brand extension evaluation,
because emphasis on innovation was the only type of marketing activity that enhanced
the perceived fit between the corporate parent brand and the extension.
27
PART II.
CON CEPTU AL FRAM EW ORK
Chapter 4. H ypotheses 29
Chapter 5. Research Design 36
28
CH APTER 4.
H Y POTH ESES
Since this study is a broad replication of Aaker and Keller’s (1990), the original model will
be modified in order to be consistent with the scope of the study.
The dependent variable of the model is the overall attitude towards the business-
to-consumer brand extension, operationalized as the average of perceived quality of the
business-to-consumer brand extension and the likelihood of trial.
4.2 Know ledge About the Parent Brand
A prerequisite for drawing conclusions about the quality of a brand is to be familiar

with it. H ence, there has to exist a brand node in the consumer’s memory with a variety
of associations linked to it–this is conceptualized as brand knowledge (Keller, 2003).
Information stored in the memory network can be verbal, visual, abstract, or contextual
in nature. Brand knowledge can be characterized in terms of two components: brand
awareness and brand image (ibid.). If consumers are to appreciate the appropriateness
of the brand extension, knowledge of the brand-specific association is required
(Broniarczyk & Alba, 1994). Thus the first hypothesis is:
H ypothesis 1: H igher degrees of know ledge about the business-to-business parent

brand are associated w ith m ore favorable attitudes tow ard the consum er brand
extension.
4.3 Attitude Tow ards the Parent Brand
Attitude is conceptualized in this study as consumer’s perception of the brand’s

overall quality. Aaker and Keller (1990) propose a relation between perceived quality of
parent brand and consumers’ attitude toward the extensions in unrelated product
categories. As the perceived quality (termed QU ALITY) of the parent brand is higher, the
transfer of positive attitudes toward the extension is also higher.
29
Zeithaml (1988) defines perceived quality as a global assessment of a consumer’s
judgment about the superiority or excellence of a product, and also concludes that
perceived quality is a construct that is on a higher level of abstraction compared to a
specific product attribute. Following this last finding, it could then be argued that
despite the absence of any prior experience with a business-to-business brand,
consumer attitudes of that brand might still be present. H ence, it will be expected to find
similar results for business-to-business parent brands with business-to-consumer brand
extensions. Keller and Aaker (1997) mention that corporate brand equity lies in the
association of consistent delivery of superior functionality and performance that
customers or suppliers have with a firm’s offering. This leads to the second hypothesis:
H ypothesis 2:H igher quality perceptions tow ard the business-to-business parentbrand
are associated w ith m ore favorable attitudes tow ard the consum er brand extension.
4.4 Innovation
The following hypotheses (H 3, H 4 and H 5) relate to Keller and Aaker’s (1997) study
on corporate brand extensions. Since their study examines the impact of different types
of corporate descriptions on brand extension evaluation, the hypotheses will be
formulated in order to fit the original model of Aaker and Keller (1990).
Keller and Aaker (1997) propose that marketing efforts that emphasize innovation
leads to favorable perceptions of corporate expertise and thus has a positive impact on
corporate brand extension evaluation. For the purposes of the current study, it is then
reasonable to assume that a parent brand that is perceived to be innovative will lead to
a more favorable brand extension evaluation compared to a brand which is not
perceived as innovative. The third hypothesis then becomes:
H ypothesis 3: H igher perceptions of innovativeness tow ard the business-to-business

parent brand are associated w ith m ore favorable attitudes tow ard the consum er
brand extension.
4.5 Corporate SocialResponsibility
Keller and Aaker (1997) propose that marketing activities directed towards
environmental awareness and community involvement increase the perceived likeability
30
and trustworthiness, but has no significant effect on an extension’s perceived quality.
This leads to the fourth hypothesis:
H ypothesis 4:Perceptions of corporate socialresponsibility of the parent business-to-

business brand has no effect the on the attitude tow ards the consum er brand
extension.
4.6 Environm entalConcern
Following Keller and Aaker’s (1997) findings that marketing efforts emphasizing
environmental concern only has a modest impact on an extension’s perceived quality,
the fifth hypothesis is:
H ypothesis 5:Perceptions of environm entalconcern of the parentbusiness-to-business

brand has no effecton the attitude tow ards the consum er brand extension.
4.7 Transfer
As proposed by Aaker and Keller (1990), the transfer of positive attitudes is also
influenced by the similarity between the corporate brand and the extension. This follows
the categorization theory and category-based processing, where consumers evaluate a
new brand extension as to how well this “fits” with the parent brand. Similarly, Boush
and Loken (1991) propose that affect associated with the original brand is transferred to
the extension when similarity between the two products is high. Following this, if
consumers perceive a “fit” between a business-to-business brand and a consumer
product class, they will transfer perceptions of quality to the new brand extension.
The first dimension of fit is the variable TRAN SFER, which is defined as the degree
in which the overall skills of the corporate brand can be helpful for the development of
the brand extension. In this case, if the parent brand organization’s resources and
abilities are useful for making the consumer brand extension, the transfer of positive
attitudes is expected to be higher. Therefore, the sixth hypothesis is:
H ypothesis 6:The transfer of a business-to-business parent brand’s perceived quality is

enhanced w hen the product classes (of the parent brand and the consum er brand
extension) in som e w ay fittogether. W hen the fitis w eak, the transfer is inhibited.
31
4.8 Brand Associations
Aaker and Keller (1990) identify two additional dimensions of fit: the first one is
COMPLEMEN T, which is related to the product; the second one is SU BSTITU TE, which is
related to the producer. W hen the product of the parent brand and the brand extension
can be consumed jointly to satisfy some need, it is said that they complement each
other (H enderson & Quandt, 1980). On the other hand, when the brand extension can be
used instead of the parent brand product, they are substitutes.
Substitutability is in this model a generally weak predictor, since relatively few
brand extensions represent true substitutes (Bottomley & H olden, 2001). In the case of
an extra-sectoral movement, the brand extension can by definition not substitute for the
product of the parent brand. The Substitute dimension will hence be omitted from
further analyses. The original microeconomic definition of complementarity is limited for
current purposes, as according to above given definition it involves joint consumption
of two product classes. In the case of an extra-sectoral movement, it is not possible for a
brand extension to complement the original brand, since the customers for the original
brand and brand extension respectively are different (i.e. business-to-business and
consumers). H ence, the Complement dimension has to be replaced in order to fit the
model.
A useful alternative measure for complementarity is the concept of brand
associations (Broniarczyk & Alba, 1994; Section 2.3.4), which proposes that consumer do
not only evaluate the brand extension based on the perceived product category fit (c.f.
H ypothesis 6), but that their assessment are driven primarily by the associations of the
brand. In other words, if a person considers a brand extension to be relevant with the
original brand concept, the attitude towards the extension will be positive.
The seventh hypothesis is:
H ypothesis 7:If the brand associations of the consum er brand extension are consistent
w ith brand concept of the business-to-business parent brand, the attitude tow ard the
brand extension is positive.
4.9 Perceived Difficulty
Aaker and Keller (1990) introduce a last variable: DIFFICU LT. This denotes the
perceived difficulty in designing or producing the brand extension product. If the
32
extension is perceived to be too easy or difficult to make, consumers’attitude towards
the extension will be negative. This is because a quality brand combined with a trivial
product class is viewed as either incompatible with the parent brand or exploitative by
means of a premium pricing strategy (Aaker & Keller, 1990; Bottomley & H olden, 2001).
Intermediate levels of perceived difficulty lead to a favorable attitude, as suggested by
Kapferer (1997).
In the current study, it is reasonable too assume that consumers will evaluate the
difficulty of designing and making consumer brand extensions of business-to-business
parent brands as they would with “normal” brand extensions. Thus, the eighth
hypothesis is:
H ypothesis 8:The relationship betw een the difficulty of m aking the consum er product
class of the brand extension and the attitude tow ard the brand extension is positive.
4.10 M oderating Factors
In addition for testing for the effects described so far, some interaction effects will
be assessed.
First, it is expected that the effect of difficulty to make a brand extension will be
weakened in the case of higher levels of perceived innovativeness of the parent brand.
This is because an innovative parent brand might be perceived as being more capable of
producing brand extensions that are difficult to make. Reasoning in the same way, an
innovative parent brand might be perceived as being less capable of producing brand
extensions that are easy to make (easy is assumed to be the linear opposite of difficult).
H ence, the following will be tested:
H ypothesis 9a:A higher level of perceived parent brand innovativeness w eakens the
im pactof perceived difficulty(to produce an extension) on brand extension evaluation.
H ypothesis 9b:A higher levelof perceived parentbrand innovativeness strengthens the

im pactof perceived ease (to produce an extension) on brand extension evaluation.
Second, evaluation of a brand extension need not correspond to evaluation of that

brand in its original category because the value of a brand association differs depending
on the benefits sought within a particular product category (Broniarczyk & Alba, 1994).
33
H ence, there might be an interaction between brand-specific associations and the
similarity of the extension category. This leads to the following:
H ypothesis 10:A higher levelof perceived category fit betw een the parent brand and
the brand extension strengthens the im pactof perceived extension brand associations.
34
4.11 Sum m ary
Exhibit 4.1 below lists all hypotheses and also refers to its academic source from
which the concept was drawn to form each specific hypothesis.
TABLE 4.1 H ypotheses and Their References to Literature
H yp othesis Literature reference
H1 H igher degrees of know ledge aboutthe business-to-business parentbrand are Broniarczyk & Alba
associated w ith m ore favorable attitudes tow ard the consum er brand (1994)
extension.
H2 H igher quality perceptions tow ard the business-to-business parentbrand are Aaker and Keller (1990)
associated w ith m ore favorable attitudes tow ard the consum er brand
extension.
H3 H igher perceptions of innovativeness tow ard the business-to-business parent Adapted from Keller
brand are associated w ith m ore favorable attitudes tow ard the consum er and Aaker (1997)
brand extension.
H4 Perceptions of corporate socialresponsibility of the parentbusiness-to- Adapted from Keller

business brand has no effectthe on the attitude tow ards the consum er brand and Aaker (1997)
extension
.
H5 Perceptions of environm entalconcern of the parentbusiness-to-business Adapted from Keller
brand has no effecton the attitude tow ards the consum er brand extension. and Aaker (1997)
H6 The transfer of a business-to-business parentbrand’s perceived quality is Aaker and Keller (1990)
enhanced w hen the productclasses (of the parentbrand and the consum er
brand extension) in som e w ay fittogether. W hen the fitis w eak, the transfer is
inhibited.
H7 If the brand associations of the consum er brand extension are consistentw ith Adapted from
brand conceptof the business-to-business parentbrand, the attitude tow ard Broniarczyk & Alba
the brand extension is positive. (1994)
H8 The relationship betw een the difficulty of m aking the consum er productclass Aaker and Keller (1990)
of the brand extension and the attitude tow ard the brand extension is positive.
H 9a A higher levelof perceived parentbrand innovativeness w eakens the im pact Own construct
of perceived difficulty (to m ake an extension) on brand extension evaluation.
H 9b A higher levelof perceived parentbrand innovativeness strengthens the Own construct

im pactof perceived ease(to m ake an extension) on brand extension
evaluation.
H 10 A higher levelof perceived extension brand associations strengthens the Own construct
im pactof perceived category fitbetw een the parentbrand and the brand
extension.
35
CH APTER 5.
RESEARCH DESIG N
This chapter will provide with the research design of the present study. First, the
hypotheses brought forward in the previous chapter are operationalized through the
development of a new model. An argument for the choice of method for executing the
new model is then provided. A description of questionnaire development and sampling
procedure follows.
5.2 The H ypothesized M odel
5.2.1 The originalm odelofA aker and Keller
Aaker and Keller (1990) hypothesized that “the consumer’s attitude towards the brand
extension is a positive function of the quality of parent brand, the fit between the
parent’s brand category and the extension category (measured in terms of the
transferability of skills and expertise from one category to the other and the
complementarity and substitutability of one category and the other), the interactions of
quality with three fit variables, and the degree of difficulty in designing and making a
product in the extension category” (Bottomley & H olden, 2001, p. 495). Formally, the
following model was tested:
Y = α + β1Q + β2T + β3C + β4S + β5QT + β6QC + β7QS + β8D + ε,
where the dependent variable

Y = attitude towards the brand extension = [purchase + extension]/2,
and where the independent variables
Q = quality
T = transfer
C = complement
36
S = substitute
D = difficult
α = intercept
ε = error term.
5.2.2 The current m odel
Aaker and Keller (1990) and subsequent replications operationalize the

hypotheses by means of a linear regression model. The model in its original state cannot
be used in the current context, but has to be changed. First, following Section 4.2 and
hypothesis 1, the variable KN OW LEDG E will be added. Second, following the reasoning
from Section 4.3, the independent variables SU BSTITU TE and COMPLEMEN T will be
omitted. These will be replaced by the new variable BRAN D CON CEPT CON SISTEN CY
(hypothesis 4). H ence, there will only be two “fit” variables in the new model –
TRAN SFER and BRAN D CON CEPT CON SISTEN CY. Furthermore, all three original
interaction variables will be omitted from the original model. Third, three independent
variables corresponding to hypotheses 6, 7 and 8 will be added to the new model.
H ypothesis 6 relates to the IN N OVATIVEN ESS of the parent brand, hypothesis 7 relates to
CORPORATE SOCIAL RESPON SIBILITY, and hypothesis 8 corresponds to
EN VIRON MEN TAL CON CERN . This gives the following:
(2) Y = α + β1K + β2Q + β3T + β4B + β5D + β6I + β7C + β8E + ε,

K = PAREN T BRAN D KN OW LEDG E

B = BRAN D CON CEPT CON SISTEN CY
I = IN N OVATIVEN ESS
C = CORPORATE SOCIAL RESPON SIBILITY
E = EN VIRON MEN TAL CON CERN .
Lastly, there might be some interaction effects. The first interaction effect is
between IN N OVATIVEN ESS and DIFFICU LT, and the second interaction term is between
37
TRAN SFER and BRAN D CON CEPT CON SISTEN CY, corresponding to H ypothesis 9 and 10
respectively. This leads to:
(3) Y = α + β1K + β2Q + β3T + β4B + β5D + β6I + β7C + β8E + β9ID + β10BT + ε,
where the additional interaction terms
ID = IN N OVATIVEN ESS × DIFFICU LT

BT = BRAN D CON CEPT CON SISTEN CY × TRAN SFER.
The hypothesized model is summarized in figure 5.1.
FIG U RE 5.1 H ypothesized Model
Parent brand H1
knowledge
Parent brand
perceived quality H2
Parent brand corporate

H7
social responsibility
Parent brand
environmental H8 Attitude
concern towards
brand
extension
Brand concept
H4
consistency
H 10
Transfer H3
Difficult
H5
H 11
Parent brand
innovativeness H6
38
5.3 M ethodology
The current research follows the approach used in the quantitative analysis in
Aaker and Keller’s (1990) study and replication studies. In these studies, qualitative
analyses are also performed prior to the quantitative analysis. Since the current study
aims to validate the Aaker and Keller (1990) in a B2B context for the sake of
generalizability and reproducibility, the current study will exclusively be of quantitative
nature.
The quantitative part tries to formally assess the consumer’s evaluation of brand
extensions through measuring attitude for different variables. In other words,
hypotheses 1 to 10 will be operationalized through a new model adapted from Aaker
and Keller (1990) and other theories discussed in chapter two and three.
5.4 Questionnaire Developm ent
5.4.1 Brands and brand extensions

To test the hypothesized model, a questionnaire was developed. This
questionnaire consisted of various questions on five well-known global B2B brands and
11 hypothetical consumer brand extensions (see table 5.1). Out of these five B2B brands,
three were distinct service brands with little product features. The other two brands
were distinct product brands with minimal service features. All brands were chosen to
maximize the spread regarding industry in order to elicit specific associations per brand.
To avoid any confounding reactions, only the logos of each brand were presented
without any additional information.
The brand extensions used were chosen had to be reasonable and not illogical,
but aimed to provide with heterogeneity on the three fit measures of the hypothesized
model. To achieve this, some extensions were deliberately chosen to be of “lower
quality”, thus allowing for some variance with respect to the perceived quality of the
brand extension.
39
Table 5.1 Overview of B2B Brands and B2C Extensions
Service (S)/ H ypotheticalExtensions

Originalbrand Product (P) OriginalProduct or Service U sed in the Survey
Reuters S Information provider N ews radio station,

24 hour news TV channel
S Assurance, tax and legal, and financial Private tax planning service,
advisory services Institute of Accountancy
(Executive MBA education),
off-shore banking service
Saatchi & Saatchi S Advertising Marketing/advertising books,

Institute of Advertising (executive
training courses)
Intel P Chips, boards, systems, and software Portable MP3 music player,
components notebook PC
Boeing P Commercial airplanes, military aircraft, Chronograph wristwatch,

satellites, missile defense, human space flight simulator computer game,
flight, and launch systems travel luggage
5.4.2 Scaling and m ulti-item scales

The Likert scale is a widely used rating scale that requires the respondents to
indicate a degree of agreement or disagreement with each of a series of statements
about the stimulus objects (Albaum, 1997; Brody & Dietz, 1997; Likert, 1932). The
advantages of Likert-type scales are that they are easy to construct and administer, and
respondents are familiar about how to use them. This makes Likert-type scales suitable
for Internet surveys, mail, telephone or personal interviews. A major disadvantage is that
it takes longer to complete Likert-type scales than other itemized rating scales because
the respondents have to read and fully reflect upon each statement (Malhotra and Birks,
2003).
To assess the overall quality (QU ALITY) of each parent brand as well as each
brand extension, a 7-point Likert scale was used (1 = poor, 7 = outstanding). A second 7-
point scale measure assessed the attributes of the parent brand (IN N OVATIVE, CSR and
40
EN VIRON MEN TAL CON CERN ) where (1 = not at all, 7 = very). A third 7-point scale
measured the three fit variables (TRAN SFER, BRAN D CON CEPT CON SISTEN CY, and
DIFFICU LT) where (1= totally disagree, and 7 = totally agree). This scale was also used to
assess the likelihood of brand extension trial (TRIAL). A fourth 7-point scale measured
the familiarity about the parent brand (KN OW LEDG E) where (1 = not at all, and 7 = very).
In Likert-type scales, the neutral points (4) can be considered to be natural
origins5. As in the original Aaker and Keller (1990) model and consequent replication
studies, the measured ordinal scales can be interpreted as interval scales. This allows for
using multiple regression as an analytical method.
Multiple-item rating scales are often use in social sciences research to measure
abstract constructs (Aaker et al., 1998). By having several items that measure the same
so-called construct, the problem of having single unrepresentative questions is solved
(ibid.). The greater the number of initial items generated, the better will be the final
scale. The larger the scale, the greater is the reliability, but shorter scales are easier for
respondents to answer. H ence, a balance between brevity and reliability has to be struck
to determine the optimal scale.
Multiple-item scales were developed for the variables IN N OVATIVE, CSR,
TRAN SFER, BRAN D CON CEPT CON SISTEN CY, and DIFFICU LT by means of 7-point Likert-
scales. Several items relevant for each variable were generated to form constructs. The
multiple-item scales were evaluated in the pilot-testing of the questionnaire, after which
some items were rephrased for clarity and some were omitted for brevity with respect to
the final length of the questionnaire. The final set of multiple-item constructs can be
seen in table 5.2.
5
This solves the problem of the absence of a real zero in the scale (van Riel et al., 2001).
41
TABLE 5.2 Multiple-scale items
Innovative
This brand is modern and up-to-date
This brand invests in R&D
This brand introduces the latest product/service features
Corporate SocialResponsibility
This brand is involved in helping its community
This brand is socially responsible
Transfer
The competences of Reuters are useful to make this extension
The resources of Reuters are useful to make this extension
Brand concept consistency

This brand extension is consistent with the Reuter brand
This brand extension fits with my associations of the Reuter brand
D ifficult
This brand extension will be difficult to make
Reuters is out to make short-term profits with this brand extension
5.5 Data Collection
5.5.1 Pilot-testing
Pilot-testing can be used to test the questionnaire on a small sample of
respondents to identify and eliminate potential problems (Martin & Polivka, 1995). A
paper version of the questionnaire along with additional pages with open-ended
questions concerning question content, wording, sequence, form, layout, question
difficult and instructions were handed out to ten graduate students enrolled at the
Faculty of Economics and Business Administration of Maastricht U niversity in the
N etherlands. The respondents of the pilot-test were similar to those who were included
in the subsequent survey, i.e. they were drawn from the same population.
Some changes were made to the first version of the questionnaire, including
removal of some brand extensions to make the questionnaire length shorter. Some
42
questions were also rephrased for sake of clarity. The pilot-testing as well as the final
version of the questionnaire can be found in appendix A.
TABLE 5.3 Response Profile:Sampling Type
Frequency Percentage
Paper 35 34.0 %
Online 68 66.0 %
Total 103 100.0 %
TABLE 5.4 Response Profile:G ender
M ale 59 57.3 %
Fem ale 44 42.7 %
Total 103 100.0 %
TABLE 5.5 Response Profile:N ationality
Dutch 49 47.6 %
G erm an 20 19.4 %
Other 34 33.0 %
Total 103 100.0 %
5.5.2 Sam pling and response

In the original Aaker and Keller (1990) study as well as in most of the replication
studies, the samples were drawn from student populations. The survey of the current
study was conducted both online and via regular paper questionnaires. Both survey
types sampled graduate and undergraduate students enrolled at the Faculty of
Economics and Business Administration of Maastricht U niversity (U M) in the
N etherlands. The online (Internet) questionnaire was arranged by the Marketing
Department of the Faculty of Economics and Business Administration of Maastricht
U niversity by which e-mail invitations were sent out to graduate and undergraduate
faculty students who were registered as willing to participate in online surveys
conducted by the faculty. The regular (paper) questionnaires were handed out among
faculty students at U M.
43
A total of 103 students were sampled out of which 66% were sampled via the
Internet and 34% sampled via regular paper questionnaires. Respondents varied in age
between 19 and 29 years old, 57% were male and 43% were female. In terms of
nationality, two thirds of the sample consisted of two large groups:47% were Dutch and
19% were G erman.
44
PART III.
RESEARCH FIN DIN G S
Chapter 6. Analysis, Results and Interpretation 46
45
CH APTER 6.
AN ALY SIS,RESU LTS AN D IN TERPRETATION S

In this chapter, the findings of the empirical research are presented. First are the results
of the three types of analyses in the order in which they were performed. The first two
analyses aim to increase internal consistency and reduce collinearity, and thus enable a
more solid regression analysis. The regression results pave the way for the hypothesis
testing, the next section of this chapter. Following this is an interpretation and
discussion of each hypothesis.
6.2 Reliability,Collinearity and Regression Analyses

The data provided by the questionnaires was analyzed using the statistical
software application SPSS (Statistical Package for the Social Sciences). The reliability of
multiple-item constructs were assessed by measuring internal consistency. The problem
of multicollinearity was dealt with using the method of residual centering. To test the
hypothesized model, multiple regression analyses were used.
6.2.1 Reliability assessm ent

Prior to performing the regression, reliability analyses were performed on the
following constructs:(1) the dependent variable, operationalized by the average of the
perceived quality of the extension and the likelihood of trial, and (2) variables
operationalized through multiple-item scales.
Because the two items of the DIFFICU LT construct were the same measure but in
two different directions6, the item that measured “ease” was reversed as necessary step
to perform the reliability analysis as well regression analyses.
Reliability was assessed by measuring the reliability coefficient, or Cronbach’s
alpha, which is defined as the average of all possible split-half coefficients resulting from
6
One item measured the difficulty to make an extension, while the other item measured the ease to make an
extension.
46
different ways of splitting the scale items (Cronbach, 1951). This coefficient varies
between 0 and 1, where a value of 0.6 or less generally indicates unsatisfactory internal
consistency reliability (Malhotra & Birks, 2003). An important property of Cronbach’s
alpha is that its value tends to increase as the number of scale items increases. Therefore,
the coefficient alpha may be artificially inflated by including several redundant scale
items (Peterson, 1994).
The reliability analysis for the dependent variable and the five independent
variable constructs reveled that all measures had an alpha coefficient exceeding 0.7
except for the construct DIFFICU LT ( = 0.21) (see table 6.1). The results from the
reliability analysis indicated that the scale variance would be reduced the most if the
item corresponding to the ease to make an extension (labeled DIFF2; see appendix B.1)
was removed. Consequently, this item of the DIFFICU LT construct was eliminated from
further analyses.
Because the multiple-item scales only included two or three items per construct,
the problem of artificially inflated coefficients is considered not to be an issue in the
current study.
TABLE 6.1 Cronbach’s Alpha Coefficients for Multiple-Item Constructs
D ependent variable construct

Attitude towards brand extension 0.74
Independent variable constructs
Innovative 0.87
Corporate social responsibility 0.85
Transfer 0.93
Brand concept consistency 0.94
Difficult 0.21
6.2.2 M ulticollinearity
Multicollinearity is when two or more independent variables are correlated with
each other. W hen highly correlated independent variables are present in a regression
model, the results are confusing and interpretation of estimated variables is difficult
(McClave et al., 1998). The original Aaker and Keller study (1990) and replication studies
(Sunde & Brodie, 1993; Bottomley & Doyle, 1996; van Riel et al., 2001) have all noted a
presence of multicollinearity between main effects and interaction terms.
47
In the current model, multicollinearity caused by a high correlation between
main effects and the interaction term was also present. This was indicated by the
Variance Inflation Factors (VIFs) of the independent variables, as shown in table 6.2. The
VIF scores indicate to what extent each independent variable is explained by the other
independent variables (H air et al., 1998). A VIF score of four or higher is considered to be
indicative of severe collinearity (Olsen, n.d.).
A method was developed by Lance (1988) to deal with the effects of
multicollinearity. This method, called “residual centering” is a two-step regression that
“serves to substantially reduce multicollinearity among first-order terms in polynomial
regression equations for any given independent variable” (N eter et al., 1990, p. 411).
Lance (1988) argues that the benefits of the “residual centering” approach over
conventional OLS regression analysis in the presence of high degrees of multicollinearity
between the main and interaction terms are fourfold:(1) multicollinearity between the
exogenous variables is substantially reduced; (2) standard errors are reduced; (3) main
and interaction effects are separated; and (4) the relative importance of the main and
interaction variables can easier be identified.
48
TABLE 6.2 VIF Scores of Regression Variables Before and
After Residual Centering
Before After
Indep endent variables Residual Residual
Centering Centering
Parent brand-specific effects

Parent brand knowledge 1.91 1.91
Parent brand quality 2.23 2.23
Innovative 5.80 2.32
CSR 1.97 1.97
Environmental concern 1.95 1.95
Brand extension-specific effects

Transfer 7.73 4.35
Brand concept consistency 13.54 2.89
Difficult 15.38 15.38
Interaction effects
Innovativeness × Difficult* 18.91 13.93*
Brand concept consistency ×
Transfer * 29.71 1.13*
* Formally, the two interaction variables are labeled as “IN N OVATIVE ×

DIFFICU LT[Residual]” and “TRAN SFER × BRAN D CON CEPT
CON SISTEN CY[Residual]after the residual centering has been applied.
Replications of Aaker and Keller’s (1990) model by N ijssen and H artman (1994) and by
Bottomley and Doyle (1994) apply Lance’s (1988) residual centering regression
approach, arguing that their researches offer an improved view on the interaction
effects.
Following N ijssen and H artman (1994) and like Bottomley and Doyle (1994) as well
as Riel et al. (2001), the residual centering technique was applied in the current study. In
this two-stage regression procedure each interaction variable (e.g. IN N OVATIVEN ESS ×
DIFFICU LT) was regressed on its two component parts (i.e. IN N OVATIVN ESS and
DIFFICU LT in the example) using OLS regression. The resulting residuals (e.g.
IN N OVATIVEN ESS × DIFFICU LT[Residual]) were then used in place of the respective
interaction term when estimating the full effects model using OLS (cfr. Bottomley &
Doyle, 1994).
Multicollinearity was successfully eliminated for independent variables
IN N OVATIVE and BRAN D CON CEPT CON SISTEN CY, and the interaction term BRAN D
49
CON CEPT CON SISCEN TY × TRAN SFER, whose VIF scores fell below 2.89 or lower after
residual centering. For the independent variable TRAN SFER, and for the interaction term
IN N OVATIVEN ESS × DIFFICU LT, multicollinearity was only partially eliminated (VIF scores
remain above four). N o multicollinearity could be eliminated for the variable DIFFICU LT.
Exhibit 6.4 summarizes the VIF scores of regression variables before and after residual
centering.
As a consequence of the high VIF scores, the interaction term IN N OVATIVEN ESS ×
DIFFICU LT was omitted from further analysis, after which multicollinearity was totally
eliminated (VIF < 2.89). The variable DIFFICU LT was kept in the model. N ote that its VIF
score after the removal of the interaction term is well below four. Exhibit 6.5 shows the
results of this action.
TABLE 6.3 VIF Scores of Regression Variables After

Omitting Interaction Term
Indep endent variables

Parent brand knowledge 1.96
Parent brand quality 2.27
Innovative 2.77
CSR 1.97
Environmental concern 1.95

Transfer 2.96
Brand concept consistency 2.88
Difficult 1.27
Interaction effects
Innovativeness ×
Difficult[Residual] removed
Brand concept consistency ×
Transfer [Residual] 1.14
50
6.2.3 Regression analysis results
After residual centering and omitting one interaction term, the hypothesized
model can be summarized as follows:
(4) Y = α + β1K + β2Q + β3I + β4C + β5E + β6T + β7B + β8D + β9BT[Residual]+ ε,

K = PAREN T BRAN D KN OW LEDG E

Q = PAREN T BRAN D QU ALITY
I = IN N OVATIVEN ESS
C = CORPORATE SOCIAL RESPON SIBILITY
E = EN VIRON MEN TAL CON CERN .
T = TRAN SFER
B = BRAN D CON CEPT CON SISTEN CY
D = DIFFICU LT
BT[Residual]= BRAN D CON CEPT CON SISTEN CY × TRAN SFER [Residual].
The hypothesized model (4) was formally tested by means of linear regression. The
analysis included the data from the 103 respondents, giving a total sample size of 1,133.
6.1.3.1 O verallm odelevaluation

The significance of the regression model as a whole was tested automatically by
the SPSS software. The computed F statistic equals 180.77, which is significant at
p= 0.000. Theoretically, this indicates that one or more regression coefficients have a
value different from zero (Thomas, 1997), i.e. at least one variable is relevant. Section 6.2
will assess the significance of specific coefficients.
The “closeness of fit” of the regression model is determined by its R2. It captures the
percentage deviation from the mean in the dependent variables that could be explained
by the model. It is always possible to increase R2 by adding extra explanatory variables,
regardless of the true significance of these. A fairer measure of closeness of fit is adjusted
R2, i.e. adjusted for degrees of freedom (ibid.). The adjusted R2 for the current model is
51
0.62. This compares favorably with the original Aaker and Keller (1990) model and
replications studies. The regression model was also tested for each brand as well as for
services and products. The adjusted R 2 values of these tests are shown in table 6.6b.
52
TABLE 6.4 Comparison Effects of Original and Replication Studies (t values are in parentheses)
Aaker Sunde and N ijssen Bottomley Bottomley Van Riel et Van Riel et Current study1
and Brodie and and Doyle and Doyle al. (2001) al. (2001)
Keller (1993) H artman (1996) U K (1996) N Z Services G oods
(1990) (1994)
Quality N /A 0.38 (5.3) 0.25 (9.1) 0.22 (11.2) .25 (12.91) 0.15 (9.38) 0.17 (6.78) 0.07* (2.39)
Transfer 0.24 0.21 (3.1) 0.58 (18.9) 0.31 (13.6) .26 (12.48) 0.34 (13.5) 0.41 (13.1) 0.27* (8.41)
Complement 0.17 0.29 (4.2) 0.01 (0.3) 0.31 (13.2) .30 (14.48) 0.38 (18.7) 0.20 (7.59) NA
Substitute 0.08 0.13 (1.9) .08 (2.7) 0.18 (7.67) .19 (8.86) 0.06 (3.40) 0.19 (7.13) NA
Difficult N /A 0.00 (0.1) Omitted 0.02 (1.22) .03 (1.59) -0.09 (-4.1) –0.16 (–5.3) -0.06* (-3.03)
Adjusted R2 N /A 0.48 0.48 0.47 0.43 0.57 0.53 .61

Sample size 2,140 1,413 693 1,358 1,558 1,616 808 1,133
*Significant at 0.000 level.

1
Beta coefficients are taken from the full model. Since the variables of the current model are different from the variables of the
original and replication studies, a formal comparison is not suitable. The comparison is nevertheless interesting on an intuitive level.
53
TABLE 6.5 Regression Results:Main and Interaction Effects (residually
centered)
Indep endent variables Fullm odel
Predicted
direction of Std.
hypotheses b Error Beta T-value
Constant -0.20 0.18 -1.13

H 1:Parent brand knowledge + -0.03 0.02 -0.03 -1.32
H 2:Parent brand quality + 0.08** 0.03 0.07 2.39
H 3:Innovative + 0.15**** 0.03 0.13 4.31
H 4:CSR 0 0.18**** 0.03 0.16 5.98
H 5:Environmental concern 0 0.06** 0.03 0.05 2.11

H 6:Transfer + 0.26**** 0.03 0.27 8.41
H 7:Brand concept consistency + 0.37**** 0.03 0.42 13.33
H 8:Difficult + -0.06*** 0.02 -0.06 -3.03
Interaction effects
H 9:Innovativeness × Difficult -/+ Omitted
H 10:Transfer × Brand concept
consistency[Residual] + -0.02 0.01 -0.03 -1.50
Controlvariables
Service brand -0.05 0.07 -0.02 -0.75
R2 0.62
Adjusted R2 0.61
Sample size n 1133
****p<.000; ***p<.01; **p<.05; *p<.10
54
TABLE 6.6a Full Model at Brand Level (residually centered, t-values in parentheses)
Parent brand-sp ecific effects
Product/ Parent Parent Environ-

service brand brand m ental
Brand brand know ledge quality Innovative CSR concern
Reuters S 0.09 -0.07 0.07 0.21*** 0.06

(1.33) (-0.92) (1.04) (2.71) (0.74)
Ernst & Young S -0.11** 0.10* 0.18*** 0.15*** 0.09

(-2.07) (1.78) (2.81) (3.01) (1.62)
Saatchi & Saatchi S 0.06 0.07 0.08 0.38**** -0.11

(0.97) (0.77) (0.96) (4.96) (-1.33)
Intel P -0.01 0.15*** 0.11 0.16*** 0.02

(-0.23) (2.27) (1.63) (2.51) (0.28)
Boeing P 0.05 0.03 0.02 0.15**** 0.11***

(1.05) (0.47) (0.37) (2.75) (1.99)
Services S -0.06 0.05 0.06* 0.14**** 0.18****

(-1.60) (1.17) (1.78) (4.03) (5.00)
Products P 0.00 0.08** 0.06 0.08* 0.13***

(-0.04) (1.98) (1.56) (1.75) (3.21)
****p<.000; ***p<.01; **p<.05; *p<.10
TABLE 6.6b Full Model at Brand Level (residually centered, t-values in parentheses)
Brand extension-specific effects Interaction effects

Transfer ×
Product/ Brand brand
service concept Innovativeness concept Sam ple Adjusted
Brand brand Transfer consistency Difficult × difficulty consistency Size n R2
Reuters S 0.20** 0.35**** -0.08 Omitted 0.03 206 0.51

(2.42) (4.37) (-1.47) (0.50)
Ernst & Young S 0.24**** 0.34**** -0.15*** Omitted 0.00 206 0.63
(3.57) (4.95) (-3.15) (-0.05)
Saatchi & Saatchi S 0.28*** 0.29**** -0.07 Omitted -0.02 206 0.60
(3.52) (4.02) (-1.55) (-0.34)
Intel P 0.36**** 0.25*** -0.05 Omitted -0.01 206 0.57

(4.00) (2.85) (-0.89) (-0.16)
Boeing P 0.17**** 0.48**** -0.01 Omitted -0.13**** 309 0.62

(3.28) (8.59) (-0.29) (-3.07)
Services S 0.26**** 0.34**** -0.10**** Omitted 0.01 618 0.58

(5.71) (8.32) (-3.58) (0.20)
Products P 0.26**** 0.45**** -0.03 Omitted -0.06** 515 0.64

(5.69) (9.83) (-0.85) (-2.04)
****p<.000; ***p<.01; **p<.05; *p<.10
55
6.2 H ypothesis Testing
6.2.1 H 1:Parent brand know ledge

The beta coefficient for the variable KN OW LEDG E is not significant at any level (P >
0.1), thus the first hypothesis has to be rejected. The KN OW LEDG E variable also is
insignificant on a brand level, except for cases where only the Ernst & Young brand
analyzed. For these cases, the KN OW LEDG E coefficient is both significant (P < 0.05) and
substantial (-0.11). It is noteworthy, however, that the coefficient has a negative sign as
opposed to the predicted direction of the first hypothesis.
To sum m arize, w e can reject the first hypothesis at a 0.05 significance level. At a brand
levelw e rejectH 1 for allbrands exceptErnst& Young for w hich w e acceptitatP<0.05.
6.2.2 H 2:Parent brand quality

The beta coefficient for the variable QU ALITY is significant (P < 0.05). The effect of
QU ALITY is relatively weak (0.07) compared to the original and replication studies. The
second hypothesis is henceforth accepted. On a brand level, QU ALITY is significant at
P>0.10 for the Ernst & Young brand, and at P<0.01 for Intel. The coefficients for these
brands are 0.10 and 0.15, respectively. For all other brands we reject the second
hypothesis.
In sum m ary, the second hypothesis is accepted at P=0.05. At a brand levelw e m ust
rejectH 2 exceptfor the tw o brands Ernst& Young (P<0.10) and Intel(P<0.01.
6.2.3 H 3:Innovative
The beta coefficient for the variable IN N OVATIVE is significant at P<0.000 with a
moderate effect (0.13). The third hypothesis is therefore accepted. On a brand level, we
reject H 3 except for the Ernst & Young brand for which the beta coefficient is 0.18
(P<0.01).
In sum m ary, the third hypothesis is accepted atP=0.00. Ata brand level, w e m ustreject
H 3 w ith exception of the brand Ernst& Young.
6.2.4 H 4:Corporate SocialResponsibility

The beta coefficient for the variable CORPORATE SOCIAL RESPON SIBILITY (CSR) is
significant at P<0.00, with a value of 0.18. Since the CSR effect was hypothesized to be
zero, the fourth hypothesis is rejected. H 4 was also rejected for all brands at P<0.01. The
56
CSR beta coefficient for brands are inline with the aggregate beta with exception for the
beta for the Saatchi & Saatchi brand, which is 0.38 (P<0.000).
To sum up, the fourth hypothesis is rejected atboth an aggregate level(P<0.000) and a
brand level(allbrands;P<0.01).
6.2.5 H 5:Environm entalconcern

The beta coefficient for EN VIRON MEN TAL CON CERN is significant at P<0.05,
although its effect is rather unsubstantial (0.05). On a brand level, only the beta
coefficient for the brand Boeing is significant (P<0.01) with a slightly larger effect
compared with the aggregate level beta (0.11).
In sum m ary, the fifth hypothesis is rejected at P<0.05 on an aggregate level, and
rejected for the brand Boeing.
6.2.6 H 6:Transfer
The beta coefficient for the variable TRAN SFER is significant (P<0.00) and highly
substantial (0.27). Thus, the sixth hypothesis is accepted. The effect of TRAN SFER is inline
with findings from previous replication studies. On a brand level, TRAN SFER is
significant for all of the five brands tested at P<0.05, although their effect vary between
0.20 for Reuters and 0.36 for Intel.
In sum m ary, the third hypothesis is accepted at P=0.00. At a brand level, w e accept H 3
for all brands.
6.2.7 H 7:Brand concept consistency

The beta coefficient for the variable BRAN D CON CEPT CON SISTEN CY is significant
(P<0.00) and highly substantial. The beta equals 0.42. Thus, the seventh hypothesis is
accepted. BRAN D CON CEPT CON SISTEN CY is also relevant for all brands at a 0.01
significance level. W hat is noteworthy is that the beta coefficient for Boeing is 0.48,
which is much higher than for the other brands.
To sum m arize, w e accept the seventh hypothesis at P=0.00. At a brand level, H 4 is
accepted for allbrands (P<0.01).
6.2.8 H 8:D ifficult

The beta coefficient for the variable DIFFICU LT is significant at P<0.05. H owever,
the coefficient indicates a negative effect (-0.06), which is contradictory to the direction
57
of the fifth hypothesis. W e can therefore not accept the eighth hypothesis. On a brand
level, only the beta coefficient for the brand Ernst & Young is significant (P<0.01). It too is
negative, with a beta of -0.15.
To sum m arize, w e cannot accept the eighth hypothesis because of sign reversal. The
beta coefficient is insignificant on a brand level w ith exception for Ernst & Young.
N evertheless, because of sign reversalw e cannotacceptthe eighth hypothesis for this brand.
6.2.9 H 10:Interaction effect betw een Transfer and Brand concept consistency
The moderator variable TRAN SFER x BRAN D CON CEPT CON SISTEN CY[Residual]
shows no significance (P>0.10). W e must therefore reject the tenth hypothesis. On a
brand level, the moderator variable shows no significance except for the brand Boeing,
for which the beta coefficient is negative (-0.13) at P<0.000.
In sum m ary, w e reject the tenth hypothesis at an aggregate level. W e also reject the
hypothesis at a brand levelfor allbrands except for Boeing for w hich w e accept it at a 0.000
significance level.
58
TABLE 6.7 Summary of H ypothesis Testing
Aggregate
level
H yp othesis results Brand levelresults

H 1:Parent brand knowledge Rejected Accepted for Ernst & Young brand only
H 2:Parent brand quality Accepted Accepted for Ernst & Young and Intel only
H 3:Innovative Accepted Accepted for Ernst & Young only
H 4:CSR Rejected Rejected for all brands
H 5:Environmental concern Rejected Rejected for Boeing only

H 6:Transfer Accepted Accepted for all brands
H 7:Brand concept consistency Accepted Accepted for all brands
H 8:Difficult Rejected Rejected for all brands
Interaction effects
H 9:Innovativeness × Difficult N /A N /A
H 10:Transfer × Brand concept
consistency[Residual] Rejected Accepted for Boeing brand only
59
TABLE 6.8 Extension Level Means
Attitude
tow ard Parent Parent Environ- Brand
brand brand brand m ental concept
Brand extension extension know ledge quality Innovative concern CSR Transfer Difficult consistency
Intel N otebook PC 5.08 5.58 5.67 6.22 3.62 3.90 5.57 2.85 5.31
Reuters 24-H our N ews TV
Channel 5.02 4.21 5.31 4.55 3.45 3.89 5.49 3.26 5.28
Intel Portable MP3 player 4.82 5.58 5.67 6.22 3.62 3.90 5.00 3.33 4.41
Reuters N ews Radio
Station 4.68 4.21 5.31 4.55 3.45 3.89 5.40 3.33 5.00
Ernst & Young Private tax
planning service 4.61 4.99 5.17 4.61 3.31 3.81 5.40 3.12 5.15
Boeing Flight simulator
video game 4.30 5.27 5.55 5.46 3.62 4.01 4.83 3.62 4.42
Saatchi & Saatchi Institute
of Advertising 4.24 2.69 4.55 4.38 3.52 3.73 4.67 3.77 4.40
Saatchi & Saatchi
Marketing & Advertising
books 4.22 2.69 4.55 4.38 3.52 3.73 4.50 3.84 4.36
Ernst & Young Institute of
Accountancy 4.06 4.99 5.17 4.61 3.31 3.81 5.16 3.89 4.44
Boeing "U ltra-tough"
travel luggage 4.06 5.27 5.55 5.46 3.62 4.01 3.53 3.50 3.56
Boeing Chronograph
wristwatch 3.16 5.27 5.55 5.46 3.62 4.01 2.95 4.06 2.40
Average 4.39 4.61 5.28 5.08 3.52 3.88 4.77 3.51 4.43
Standard deviation 0.55 1.06 0.40 071 0.12 0.11 0.84 0.37 0.85
TABLE 6.9 Perceived Quality of Parent Brands
Standard
Brand M ean Deviation
Intel 5.67 1.00

Boeing 5.55 1.18
Reuters 5.31 1.27
Ernst & Young 5.17 1.17
Saatchi & Saatchi 4.55 1.08
Average 5.25 1.14
60
6.3 Discussion and Interpretation
6.3.1 Parent brand know ledge

Surprisingly, knowledge about the parent brand had no significant effect on the
dependent variable. Possible reasons for this could be that brand familiarity is an
insufficient condition for evaluating brand extensions. This may suggest that pure
affective responses do not take place, but that cognitive processing is necessary in order
to evaluate the extension. This is inline with Fiske and Pavelchak’s (1986) proposed
concept of piecemeal processing.
Interestingly, the KN OW LEDG E beta coefficient was negative (-0.11) on a brand
level for the brand Ernst & Young. A possible explanation for this may be that the sample
population (i.e. business students) in part possess so much knowledge about Ernst &
Young that they feel that the brand extensions would be unrealistic.
6.3.2 Parent brand quality

As expected, parent brand quality plays a role in evaluating the brand extension.
H owever, the beta coefficient for the QU ALITY variable of the current study is
considerably lower than in previous replication studies (0.07). In other words, the
perceived quality of a B2B brand has less spillover effects than a B2C brand, ceteris
paribus. Brand quality may not be the optimal way to measure brand equity in the
context of B2B brands judged by consumers. N evertheless, the beta coefficient for
QU ALITY is positive. Thus, even though consumers lack experience of B2B brands they
still are able to make some inferences on their overall quality albeit in a cautious way.
6.3.3 Innovative
Parent brand innovativeness has a positive effect (0.13) on brand extension
evaluation. Thus, Keller and Aaker’s (1997) prediction also holds in the case of consumer
brand extensions of B2B brands. As mentioned in Section 3.3.1, not much research has
been done with respect to brand innovativeness (most research has been done in the
are of consumer innovativeness). The results show that innovativeness is a stronger
predictor of the dependent variable than parent brand quality. There might be two
reasons for this phenomenon: (1) Innovativeness may be a salient attribute of brand
quality, at least partially (i.e. there is an interaction effect between QU ALITY and
IN N OVATIVEN ESS); and (2) the quality of B2B may be too abstract for consumers to
evaluate, while innovativeness can be more easily judged.
61
On a brand level, only the Ernst & Young brand showed significance for the
innovativeness coefficient (0.18), indicating that innovativeness is brand-specific.
6.3.4 Corporate SocialResponsibility

Contrary to expectations, parent brand corporate social responsibility has a
positive effect (0.16) on extension evaluation. In addition, this is larger than both parent
brand quality and innovativeness.
On a brand level, the beta for CSR was consistently high on corporate social
responsibility, with exception of the Saatchi & Saatchi brand, which scored exceptionally
high.
There is a noteworthy difference in the importance of CSR when comparing
service and product brands. The beta coefficient for CSR is higher for service brands
(0.14) than for product brands (0.08). Although this may be brand specific, another
reason for this can be the drawn from the recent corporate scandals involving e.g.
accounting firms and other service companies (e.g. Enron).
Although the importance of this coefficient may have been inflated by the fact
that all brands included in the survey were global brands with high degree of perceived
CSR and codes of conduct, the effect of corporate social responsibility on consumer
extension evaluations cannot be discounted for. A likely explanation for the positive and
relatively large effect of the CSR coefficient compared to Keller and Aaker ‘s (1997)
findings is that consumers values corporate social responsibility significantly more than
industrial buyers do. This highlights the difference of perceived social responsibility
among end-consumers vis-à-vis B2B customers.
6.3.5 Environm entalconcern

Environmental concern has a small but significant effect on extension evaluation
(0.05). On a brand level, only the beta coefficient for the brand Boeing was significant.
This might be seen as logical since Boeing might be seen as a company that produces
heavy equipment and airplanes with environmental impact. This is in contrast to the
other product brand Intel, which manufactures microprocessors, and can be considered
as being “clean” and having less impact on the environment.
There is a notable difference between services and products –service brands have
a higher beta coefficient than product brands. This is remarkable since services have no
“production processes” that impact the environment.
62
The effect of environmental concern was hypothesized to be zero, but turned out
to be positive. This fact alongside the positive effect of corporate social responsibility
indicates that consumers to a certain extent take ethical considerations when evaluating
parent brands.
6.3.6 Transfer
The first fit variable, TRAN SFER is significant and its effect is substantial (0.27). This
is an important finding that indicates that current skills and resources of a company are
perceived as being transferable from business markets to consumer markets.
This finding was also relevant across all brands with a notably higher beta
coefficient for the Intel brand (0.36) compared to the other brands. A likely explanation
for the higher beta of the Intel brand may be the high extension attitude scores for the
two Intel brand extensions (see table 6.8).
The beta coefficients are the same (0.26) for product and service brands, implying
that consumers do not distinguish between skills and resources in making services or
goods respectively.
6.3.7 Brand concept consistency

The second fit variable, brand concept consistency, is the coefficient with the
most explanatory power (0.42) in the model. Although this variable was introduced in
this study and has not been used in previous studies, it has proven to be a highly
relevant fit variable compared to the variables of fit used in the original and replication
studies (see table 6.4) –adding the fact that there were no interaction effects between
the two fit variables.
Brand concept consistency is also highly relevant on a brand level, with an
exceptionally high beta for the Boeing brand. A likely explanation might be the
interaction effect between transfer and brand concept consistency. This phenomenon
will be further discussed in Section 6.3.7.
There is also an observable difference in beta coefficient weight between service
and product brands. Brand concept consistency has a higher beta for products (0.45)
than services (0.34). A reason for this may be the intangible nature of services, which
makes it more difficult to evaluate their respective brand concepts when no consumer
service offering exists.
63
6.3.8 D ifficult
A positive relationship between extension attitude and perceived difficulty of
producing the extension was hypothesized. Although the variable Difficult was
significant, the hypothesis was rejected because of sign reversal, i.e. the beta coefficient
turned out to be negative (-0.06). This might indicate that some brand extensions were
perceived to be too difficult to make and that this leads to negative extension
evaluation. This might be an important distinction in comparison to “traditional” brand
extensions that operate in the same market etc. For example, in the original Aaker and
Keller (1990) study and replication studies with exception from van Riel, Lemmink and
Ouwersloot’s (2001) study, the Difficult beta coefficients were positive. In the latter study
(van Riel et al., 2001), which tested service brand extensions, the Difficult beta was
negative. In other words, the more difficult an extension is to produce, the higher will be
the evaluation, except in cases when the extension is a service and the parent brand is a
product brand, or –as concluded in the current study –when the extension is a consumer
product/service and the parent brand is a business-to-business brand. This suggests that
there is an invisible “barrier” that brands face when extending brands to unrelated
markets.
6.3.9 Interaction effects

The first hypothesized interaction effect between Innovativeness and Difficult
was omitted because of multicollinearity, hence this effect cannot be measured. The
second hypothesized interaction effect between Transfer and Brand concept
consistency was insignificant on an aggregate level, and on a brand level only significant
for one brand, Boeing. The beta coefficient for this interaction effect was negative in this
case (-0.13), implying that no synergy effect is achieved (i.e. extension evaluation does
not increase even though brand concept consistency is perceived to be high when there
is no category fit). A reasonable explanation for this is that the brand extensions under
the Boeing brand were perceived to be of low quality (see table 6.8).
The beta coefficient of the second interaction effect was also significant for
product brands, although its direction was in opposite of the hypothesis.
6.3.10 G eneraldiscussion ofthe current m odel

The current model shows a remarkably high adjusted R2 (0.61). This is an
improvement from previous studies, although a direct comparison is unfair since
64
different variables were used in the regression analysis of the current model. Van Riel et
al. (2001) propose that the possible explanation for the high adjusted R2 of their model
could be explained by the fact that brand extensions have become more common over
the years and the explanatory power of the used constructs have been reinforced by
successful extensions (and thus positive extension evaluations). This seems like a
plausible explanation in the case of consumer brand extensions of business brands:
consumers are able to “make sense” out of these type of extensions.
The model of the current study also seems to have a better fit for products than
services, with respective adjusted R2 values of 0.64 and 0.58.
65
PART IV .
CLOSIN G PERSPECTIV ES
Chapter 7. Conclusions 67
66
CH APTER 7.
CON CLU SION S
7.1 Sum m ary of the Findings

In general, the variables that by far have the largest effects in explaining extension
attitude are the two fit variables Transfer and Brand concept consistency. This is inline
with previous studies where the effect of fit variables to a great extent superseded those
of other variables (see table 6.4). Parent brand quality has, compared to previous studies,
surprisingly little effect on the dependent variable. Variables corresponding to Keller and
Aaker’s (1997) study on corporate brand extension play, contrary to expectations, a
significant part in explaining extension evaluation.
7.2 Conclusions
The current study has shown that it indeed is possible to extend business-to-
business brands into the consumer market. It presented the brand extension model of
Aaker and Keller (1990), which was modified to fit the current context by drawing
theories from Broniarczyk and Alba (1994) and Keller and Aaker (1997). More specifically,
this study provided evidence that in the context of business-to-business brand
extensions, consumers use the transferability of skills and resources and brand concept
consistency with the parent brand category as major cues to evaluate extensions. H ence,
abstract factors facilitate extension acceptance beyond the limitations of product-
related similarity. Corporate branding attributes such as innovativeness and corporate
social responsibility also play a large role. The goodness of fit of the current model also
seems to be higher for product brands than services.
7.3 Lim itations and Directions for Future Research

There are some limitations to the present study. A first concern relates to the way
parent brand equity was measured –by measuring parent brand quality with a single-
item measure. Single-item measures are sometimes subject to criticism with respect to
67
their unreliability and low validity (Churchill, 1979). Multiple-item scales is one solution,
but the emphasis should be on constructing better measures of brand equity, in order to
fully assess the extent to which a brand can be leveraged.
A second concern relates to the one-sidedness of the current study. It measured
only consumer acceptance of the brand extensions. A relevant question in the context of
the current research would also address the attitudes of existing B2B customers when
launching consumer brand extensions. In other words, the reciprocal impacts of
consumer brand extensions on brand equity can be measured with respect to both
consumers and business-to-business customers.
A third concern relates to the fit variables. Because of the non-applicability of the
fit variables Substitute and Complement used in previous replication studies, only two fit
variables were used (Transfer and Brand concept consistency). Brand concept
consistency proved to be a useful factor, probably because of its abstractness. Future
studies on brand extensions could include Brand concept consistency as well as the
original fit variables Substitute and Complement to further examine whether the
abstractness of the former or the concreteness of the latter two are superior in attitude
formation.
A fourth limitation relates to the number of brands used in the study. Only five
brands were used, all of which were well-known global brands. Differences in adjusted
R2 on a brand level suggest that there are attributes unique to each brand, which are not
measured by Brand concept consistency. A more detailed study on brand extensions
could take into account numerous factors such as previous extensions (cfr. Keller &
Aaker, 1992), effects of extensions on a company’s brand portfolio (cfr. Dacin & Smith,
1994), or brand architecture (e.g. how brand architecture facilitates brand extension
acceptance).
A fifth limitation concerns the way the brand extension were presented. Since
each brand extension was presented only as a non-branded generic product and
without any accompanying text or visual cues, the extent to which a true assessment of
the quality and likelihood of trial by the consumer might have been limited. Related to
the brand extension presentation is the absence of pricing. Van Riel et al. (2001) suggest
that consumers may use “price clues” to assess [service]quality (Zeithaml, 1988).
68
7.4 Im plications
7.4.1 Theoreticalim plications

Bottomley and Doyle (1996) called for further research on the role of Brand
concept consistency as an important factor in determining how consumers form
attitudes towards brand extensions. The current study has proven, at least in the current
context, that brand-specific associations are more important than category similarity, in
consumer attitude formation of B2B-brand extensions. H ence, Broniarczyk and Alba’s
(1994) claim that the “brand” in brand extension is superior to category-based similarity
is supported by the current study.
The present study has also proven that variables of Keller and Aaker’s (1997)
corporate brand extension model can be used in the original A&K model, showing
further evidence that the original A&K model can be contextually adapted.
By replacing irrelevant variables of the A&K model with contextually relevant
concepts and theories as discussed above, the present study shows that a broad
empirical replication of Aaker and Keller’s (1990) is both possible and valuable for
additional explanatory power and insights in more complex or specific cases.
7.4.2 M anagerialim plications

W hether to extend a business-to-business brand to the consumer market or not
remains a predominantly managerial topic. W hat the current study has shown is that it is
possible to do so.
Brand extension strategies in the current context will be most successful when
there is a “fit” between the parent brand and the extension. This fit is determined by the
extent to which consumers perceive that the skills and resources of a company are
useful in making the extension, and more importantly whether the extension is
consistent with the brand concept of the parent brand. Any extension must therefore
begin with examining the parent brand itself, and what it stands for.
The quality of the parent brand plays a lesser role in brand extension acceptance;
although quality should not be discounted since consumers can assess brand equity in
different ways than measured by the current study.
Corporate brand attributes such as corporate social responsibility (CSR) is highly
important and can be achieved by e.g. sponsoring local communities through various
activities and programs, especially for service companies. It is, however, unclear whether
69
CSR is a trend or whether it will remain as an important attribute. N evertheless, a B2B
company should be aware of the difference in ethical values among consumers and
industrial buyers.
Another corporate brand attribute that facilitates brand extension acceptance is
innovativeness. A company should therefore strive to build an innovative reputation and
form a philosophy of constantly launching advanced products or services.
An important practical constraint with respect to the current model is that it is
only tested on B2B brands. The validity or importance of the above is hence not
confirmed in cases after a brand makes the transition from B2B to both B2B and
consumer brand. Some time after a transition is made, i.e. when the former B2B brand is
both a B2B and B2C brand, consumers may evaluate the brand extensions according to
the original model by Aaker and Keller (1990). This pinpoints the context-specificity of
the current model.
70
BIBLIOG R APH Y
Aaker, D. A. (1996). Building Strong Brands. N ew York:The Free Press.
Aaker, D. A. and Joachimsthaler, E. (2000). Brand Leadership. London:Free Press Business.
Aaker, D. A. and Keller, K. L. (1990). ”Consumer evaluations of brand extensions”. Journal

of M arketing Research, 54(1), 27-41.
Aaker, D. A., Kumar, V. and Day, G . S. (1998) Marketing Research (6th Ed.). N ew York:John
W iley & Sons.
Aaker, J. L. (1997). "Dimensions of brand personality". Journal of M arketing Research,

34(3), 347-356.
Abratt, R. (1989). “A new approach to the corporate image management process”.

Journalof M arketing, 5(1), 63-76.
Albaum, G . (1997). “The Likert scale revisited – and alternative version”. Journal of the
M arketResearch Society, 39(2), 331-48.
Anderson, J. C. and N arus, J. A. (2004). Business M arket M anagem ent - U nderstanding,

Creating and D elivering Value (2nd Ed.). U pper Saddle River:Prentice-H all.
Balachander, S. and G hose, S. (2003). “Reciprocal spillover effects:a strategic benefit of

brand extensions”. Journalof M arketing, 67(1), 4-13.
Barwise, P. (1995). “G ood empirical generalizations”. M arketing Science, 14(3), 29-36.
Berthon, P., H olbrook, M.B. and H ulbert. J.M. (2003). “U nderstanding and managing the
brand space”. M IT Sloan M anagem entReview , 44(2), 49-54.
71
Bhattacharya, C. B. and Sen, S. (2003). “Consumer-company identification:a framework
for understanding consumers'relationships with companies”. Journalof M arketing,
67(2), 76-88.
Bottomley, P. A. and Doyle, J. R. (1996). “The formation of attitudes towards extensions:

testing and generalising Aaker and Keller's model”. International Journal of
Research in M arketing, 13(4), 365-377.
Bottomley, P. A. and H olden, S. J. S. (2001). “Do we really know how consumers evaluate
brand extensions? Empirical generalizations based on secondary analysis of eight
studies”. Journalof M arketing Research, 38(4), 494-500.
Boush, D. M. and Loken, B. (1991). “A process-tracing study of brand extension

evaluation”. Journalof M arketing Research, 28(1), 16-28.
Brody, C. J. and Dietz, J. (1997). “On the dimensionality of 2-question format Likert scale
attitudes”. SocialScience Research, 26(2), 197-204.
Broniarczyk, S. M. and Alba, J. W . (1994). “The importance of the brand in brand

extension”. Journalof M arketing Research, 31(2), 214-239.
Brooks, L. (1978). “N onanalytic concept formation and memory for instances”, in

Cognition and Categorization, Rosch, E. and Lloyd B. B., eds. H illsdale, N J:Lawrence
H illbaum Associates, 169-215.
Brown, T. J. and Dacin, P. A. (1997). “The company and the product: corporate
associations and consumer product responses”. Journalof M arketing, 61(1), 68-84.
Churchill, G . A. Jr. (1979). “A paradigm for developing better measures of marketing

constructs”. Journalof M arketing Research, 16, 64-73.
Cohen, J. B. (1982). “The role of affect in categorization:towards a reconsideration of the

concept of attitude”, in Advances in Consum er research, 9, Mitchel, A., ed. Ann
Arbor, MI:Association for Consumer Research, 94-100.
72
Cronbach, L. J. (1951) “Coefficient alpha and the internal structure of tests”,
Psychom etrica, 16, 297-334.
Dacin, P. A. and Smith, D. C. (1994). “The effect of brand portfolio characteristics on

consumer evaluations of brand extensions”. Journalof M arketing Research, 31, 229-
242.
de Chernatony, L. and Segal-H orn, S. (2001). “Building on services' characteristics to

develop successful services brands”. Journal of M arketing M anagem ent, 17(7/8),
645-669.
de Ruyter, K. and W etzels, M. (2000). “The role of corporate image and extension
similarity in service brand extensions”. Journalof Econom ic Psychology, 21(6), 639-
660.
Douglas, S. P., Craig, C. S. and N ijssen, E. J. (2001). “Executive insights: integrating

branding strategy across markets: building international brand architecture”.
Journalof InternationalM arketing, 9(2), 97-114.
Fiske, S. T. (1982). “Schema-triggered affect: applications to social perception” in Affect

and Cognition:The 17th Annual Carnegie Sym posium on Cognition, Clark, M. S. and
Fiske, S. T., eds. H illsdale, N J:Lawrence Erlbaum Associates, 171-190.
Fiske, S. T. and Pavelchak, M. A. (1986). “Category-based versus piecemeal-based

affective responses:developments in schema-triggered affect”, in The H andbook of
M otivation and Cognition: Foundation for Social Behaviour, Sorrentino, R. W . and
H iggins, E. T., eds. N ew York:G uilford Press, 167-203.
G riffin, J. J. (2002). “Branding as strategy: new forms of decision-processes”. Corporate

Reputation Review , 5(2/3), 227-240.
H air, J. F., Anderson, R. E., Tatham, R. L. and Black, W . C. (1998). M ultivariate D ata Analysis
(5th Ed.). U pper Saddle River, N J:Prentice H all.
73
H enderson, J. M. and Quandt, R. E. (1980). M icroeconom ic theory: a m athem atical
approach. N ew York:McG raw-H ill.
H ubbard, R. and Armstrong, J. S. (1994). “Replications and extensions in marketing:rarely

published but quite contrary”. InternationalJournalof Research in M arketing, 11(3),
233-249.
Ind, N . (1997). The Corporate Brand. London:McMillan Press.
Ind, N . (1992). The Corporate Im age (revised ed.). London:Kogan Page.
Kapferer, J-N . (2000). Strategic Brand M anagem ent. London:Kogan Page.
Keller, K. L. (2003). “Brand equity dilution”. M IT Sloan M anagem entReview , 45(1), 12-14.
Keller, K. L. (2003). Strategic Brand M anagem ent:Building, M easuring, and M anaging Brand
Equity. U pper Saddle River:Prentice H all.
Keller, K. L. and Aaker, D. A. (1992). “The effects of sequential introduction of brand

extensions”. Journalof M arketing Research, 29, 35-50.
Keller, K. L. and Aaker, D. A. (1997). “Managing the corporate brand: the effect of
corporate marketing activity on consumer evaluations of brand extensions”.
Cambridge, MA:Marketing Science Institute.
Kitchin, T. (2003). “Corporate social responsibility: a brand explanation”. Brand

M anagem ent, 10(4-5), 312-326.
Krishnan, B. C. and H artline, M. D. (2001). “Brand equity:is it more important in services?”.

Journalof Services M arketing, 15(5), 328-342.
Lambert, Z. V. (1972). “Perceptual patterns, information handling and innovativeness”.

Journalof M arketing Research, 9(4), 427.
74
Lance, Charles E. (1988), “Residual centering, exploratory and confirmatory moderator
analysis, and decomposition of effects in path models containing interactions,”
Applied PsychologicalM easurem ent, 12(2), 163-175.
Leone, R. P. and Schultz, R. L. (1980). “A study of Marketing G eneralizations”. Journalof

M arketing, 44(1), 11-18.
Likert, R. (1932). “A technique for the measurements of attitudes”. Archives of Psychology,

140.
Loken, B. and John, D. R. (1993). “Diluting brand beliefs:when do brand extensions have
a negative impact?”. Journalof M arketing, 57(3), 71-85.
MacInnis, D. J. and N akamoto, K. (1990). “Examining factors that influence the perceived
goodness of brand extensions”. W orking Paper #54, Karl Eller G raduate School of
Management, U niversity of Arizona.
Mackay, M. M. (2001). “Application of brand equity measures in service markets”. Journal

of Services M arketing, 15(3), 210-221.
Malhotra, N . K. and Birks, D. F. (2003). M arketing Research:An Applied Approach (3rd Ed.).
H arlow:Prentice-H all.
Martin, E. and Polivka, A. E. (1995). “Diagnostics for redesigning survey questionnaires –

measuring work in the current population survey”. Public O pinion Q uarterly, 59(4),
547-564.
McAlexander, J. H ., Schouten, J. W . and Koenig, H . F. (2002). “Building brand community”.

Journalof M arketing, 66(1), 38-54.
McClave, J. T., Benson, P. G . and Sinich, T. (1998). Statistics for Business and Econom ics (7th
Ed.). U pper Saddle River:Prentice-H all.
75
Mervis, C. B. and Rosch, E. (1981). “Categorization of N atural Objects”. AnnualReview of
Psychology, 32, 89-115.
Michell, P., King, J. and Reast, J. (2001). “Brand values related to industrial products”.
IndustrialM arketing M anagem ent, 30(5), 415-425.
Moorthi, Y. L. R. (2002). “An approach to branding services”. Journalof Services M arketing,

16(3), 259-274.
Morrin, M. (1999). “The impact of brand extensions on parent brand memory structures
and retrieval processes”. Journalof M arketing Research, 36(4), 517-526.
Mudambi, S. (2002). “Branding importance in business-to-business markets -Three buyer

clusters”. IndustrialM arketing M anagem ent, 31(6), 525-533.
N eter, J., W asserman, W . and Kutner, M. H . (1990), Applied Linear Statistical M odels:
Regression, Analysis of Variance, and Experim entalD esigns. Burr Ridge, IL:Irwin.
O'Cass, A. and G race, D. (2003). “An exploratory perspective of service brand

associations”. Journalof Services M arketing, 17(5), 452-475.
Olsen, B. L. (n.d.) “Regression Analysis:Issues of Multicollinearity too often Overlooked”

[PDF document], retrieved 7th July 2004 from
http://www.iaca.net/resources/Articles/multicollinearity_article.pdf
Ostlund, L. E. (1974). “Perceived innovation attributes as predictors of innovativeness”.

Journalof Consum er Research, 1(2), 23-29.
Peterson, R. A. (1994). “A meta-analysis of Cronbach’s coefficient alpha”, Journal of

Consum er research, 21, 381-91.
Porter, M. (1998). Com petitive Advantage -Creating and Sustaining Superior Perform ance.
N ew York:Free Press.
76
Rajagopal and Sanchez, R. (2004). “Conceptual analysis of brand architecture and
relationships within product categories”. Brand M anagem ent, 11(3), 233-247.
Riel, A.C.R. van, Lemmink, J. and Ouwersloot, H . (2001). “Consumer evaluations of service
brand extensions”. Journalof Service Research, 3(3), 220-231.
Roerich, G . (2004). “Consumer innovativeness:concepts and measurements”. Journalof

Business Research, 57(6), 671-677.
Sunde, L. and Brodie, R.J., (1993). “Consumer evaluations of brand extensions: Further
empirical evidence”. InternationalJournalof Research in M arketing, 10, 47-53.
The 100 Top Brands (2004, August). Business W eek, 3864, 68-71.
Thomas, R. L. (1997). M odern Econom etrics –an Introduction. Essex:Addison-W esley.
U rde, M. (1999). “Brand orientation: a mindset for building brands into strategic
resources”. Journalof M arketing M anagem ent, 15(1-3), 117-133.
U rde, M. (2003). “Core value-based corporate brand building”. European Journal of

M arketing, 37(7/8), 1017-1040.
W ebster, F. E. Jr. and Keller, K. L. (2004). “A roadmap for branding in industrial markets”.
Brand M anagem ent, 11(5), 388-402.
Yoo, B., Donthu, N . and Lee, S. (2000). “An examination of selected marketing mix
elements and brand equity”. Journal of the Academ y of M arketing Science, 28(2),
195-211.
Zeithaml, V. A. (1988). “Consumer perceptions of price, quality, and value:a means-end

model and synthesis of evidence”. Journalof M arketing, 52, 2-22.
77
APPEN DIX A. QU ESTION N AIR ES
A.1 Pilot-testing Questionnaire
THESIS SURVEY: PILOT-TESTING

Hello,
I am conducting a survey for my final thesis on brands. To make my survey as robust as possible, I
am performing a pilot-test (of the questionnaire I will be using later in my research) for which you
have been invited. Instructions are as follows:
• Your task is to fill out the questionnaire like you usually do, but in addition to the
questionnaire questions I would like you to also think about the quality of the questionnaire
itself.
• You can write down your opinions on the questionnaire on the FEEDBACK FORM which is
found on the last two pages.
• You are also allowed to make notes/markings on the questionnaire itself, but if you do,
please make sure that these are readable.
• Since the pilot-testing only samples a small number of people, it is important that you do
your best to answer all questions of the feedback form.
• Please be as thorough and clear as possible!
The “real” questionnaire starts on the next page.
Please return this questionnaire to me as soon you can. Thanks for your help!
Leon
Please write down your personal details (you can skip this part in the actual questionnaire).
Name Age Gender Male Female
78
Nationality Dutch German Other, namely
Study programme IB IM IES FE INF Exchange student
If you are an exchange student: do you study business, economics, etc. in your home country?
Yes No
Year of study 1 2 3 4 5 or more
79
SURVEY ON BRAND EXTENSIONS
Dear participant,
My name is Leon Phang, and I am conducting a survey for my final thesis on

brands.
I am curious about what you think about different business-to-business brands and
their hypothetical brand extensions.
Please help me do this survey by taking the time to fill out this questionnaire. All
information will be treated in the strictest confidence and results will be produced in
the form of aggregated data only.
Thank you for your time and help.
INSTRUCTIONS
As mentioned, this questionnaire is about business-to-business (B2B) brands and hypothetical

brand extensions. A brand extension is a new product that uses an already existing brand name.
You will be asked about your opinion about 5 different brands. The questionnaire will take some
time and effort to fill out, so please try to do your best and make both your and my time worthwhile!
80
BRAND 1 OUT OF 5
Consider the following brand:
Now, please write down your immediate associations with this brand in the empty box:
POOR OUTSTANDING
How do you perceive the overall quality of this brand? 1 2 3 4 5 6 7
NOT AT ALL VERY

This brand is modern and up-to-date 1 2 3 4 5 6 7
This brand invests in R&D 1 2 3 4 5 6 7
This brand introduces the latest product/service features 1 2 3 4 5 6 7
This brand is concerned with the environment 1 2 3 4 5 6 7
This brand is involved in helping its community 1 2 3 4 5 6 7
This brand is socially responsible 1 2 3 4 5 6 7
Also consider the following brand extensions, and please write down your immediate associations to
these in the empty box:
a) Reuters 24 hour news TV channel
b) Reuters news radio station
81
Now, please indicate your opinion on the previously mentioned hypothetical brand extensions by
marking the appropriate number (1-7).
a) Consider the brand extension Reuters 24 hour news TV channel again.

POOR OUTSTANDING
How do you perceive the overall quality of this brand extension? 1 2 3 4 5 6 7
TOTALLY DISAGREE TOTALLY AGREE

If this brand extension existed, I would be likely to try it 1 2 3 4 5 6 7
The competences of Reuters are useful to make this extension 1 2 3 4 5 6 7
The resources of Reuters are useful to make this extension 1 2 3 4 5 6 7
This brand extension is consistent with the Reuter brand 1 2 3 4 5 6 7
This brand extension fits with my associations of the Reuter brand 1 2 3 4 5 6 7
This brand extension is too easy to make 1 2 3 4 5 6 7
b) Consider the brand extension Reuters news radio station again.

POOR OUTSTANDING

82
BRAND 2 OUT OF 5
POOR OUTSTANDING
NOT AT ALL VERY

these in the empty boxes:
a) KPMG Private tax planning service
b) KPMG Institute of Accountancy (Executive MBA education)
c) KPMG Offshore banking service
83
a) Consider the brand extension KPMG Private tax planning service again.
POOR OUTSTANDING

The competences of KPMG are useful to make this extension 1 2 3 4 5 6 7
The resources of KPMG are useful to make this extension 1 2 3 4 5 6 7
This brand extension is consistent with the KPMG brand 1 2 3 4 5 6 7
This brand extension fits with my associations of the KPMG brand 1 2 3 4 5 6 7
b) Consider the brand extension KPMG Institute of Accountancy (Executive MBA education)
again.
POOR OUTSTANDING

c) Consider the brand extension KPMG Offshore banking service again.

POOR OUTSTANDING

84
BRAND 3 OUT OF 5
1) Consider the following brand:
2) Now, please write down your immediate associations with this brand in the empty box:
POOR OUTSTANDING
3) How do you perceive the overall quality of this brand? 1 2 3 4 5 6 7
NOT AT ALL VERY

4) This brand is modern and up-to-date 1 2 3 4 5 6 7
5) This brand invests in R&D 1 2 3 4 5 6 7
6) This brand introduces the latest product/service features 1 2 3 4 5 6 7
7 This brand is concerned with the environment 1 2 3 4 5 6 7
8) This brand is involved in helping its community 1 2 3 4 5 6 7
9) This brand is socially responsible 1 2 3 4 5 6 7
a) Saatchi & Saatchi marketing & advertising books
b) Saatchi & Saatchi Institute of Advertising (executive training courses)
c) Saatchi & Saatchi Espresso Bar & Café
85
a) Consider the brand extension Saatchi & Saatchi marketing & advertising books again.
POOR OUTSTANDING

The competences of Saatchi & Saatchi are useful to make this extension 1 2 3 4 5 6 7
The resources of Saatchi & Saatchi are useful to make this extension 1 2 3 4 5 6 7
This brand extension is consistent with the Saatchi & Saatchi brand 1 2 3 4 5 6 7
This brand extension fits with my associations of the Saatchi & Saatchi brand 1 2 3 4 5 6 7
b) Consider the brand extension Saatchi & Saatchi Institute of Advertising (executive
training courses) again.
POOR OUTSTANDING

c) Consider the brand extension Saatchi & Saatchi Espresso Bar & Café again.
POOR OUTSTANDING

86
BRAND 4 OUT OF 5
POOR OUTSTANDING
NOT AT ALL VERY

a) Intel® Portable MP3 Music Player
b) Intel® Home Wireless (wi-fi) Network Hardware
c) Intel® Notebook PC
87
a) Consider the brand extension Intel® Portable MP3 Player again.

POOR OUTSTANDING

The competences of Intel are useful to make this extension 1 2 3 4 5 6 7
The resources of Intel are useful to make this extension 1 2 3 4 5 6 7
This brand extension is consistent with the Intel brand 1 2 3 4 5 6 7
This brand extension fits with my associations of the Intel brand 1 2 3 4 5 6 7
b) Consider the brand extension Intel® Home Wireless (wi-fi) Network Hardware again.
POOR OUTSTANDING

c) Consider the brand extension Intel® Notebook PC again.

POOR OUTSTANDING

88
BRAND 5 OUT OF 5
POOR OUTSTANDING
3) How do you perceive the overall quality of this brand? 1 2 3 4 5 6 7
NOT AT ALL VERY

4) This brand is modern and up-to-date 1 2 3 4 5 6 7
5) This brand invests in R&D 1 2 3 4 5 6 7
6) This brand introduces the latest product/service features 1 2 3 4 5 6 7
7 This brand is concerned with the environment 1 2 3 4 5 6 7
8) This brand is involved in helping its community 1 2 3 4 5 6 7
9) This brand is socially responsible 1 2 3 4 5 6 7
3) Also consider the following brand extensions, and please write down your immediate association
to these in the empty box:
a) Boeing chronograph wristwatch
b) Boeing flight simulator computer game
c) Boeing travel luggage
89
4) Now, please indicate your opinion on the previously mentioned hypothetical brand
extensions by placing a tick in the appropriate box.
a) Consider the brand extension Boeing chronograph wristwatch again.

POOR OUTSTANDING

The competences of Boeing are useful to make this extension 1 2 3 4 5 6 7
The resources of Boeing are useful to make this extension 1 2 3 4 5 6 7
This brand extension is consistent with the Boeing brand 1 2 3 4 5 6 7
This brand extension fits with my associations of the Boeing brand 1 2 3 4 5 6 7
b) Consider the brand extension Boeing flight simulator computer game again.
POOR OUTSTANDING

c) Consider the brand extension Boeing travel luggage again.

POOR OUTSTANDING

90
PERSONAL INFORMATION
Name (this will be used for administrational purposes only)
Age Gender: Male Female
Do you study business or economics? Yes No
Nationality
Dutch
German
Other, namely:
91
PILOT-TESTING FEEDBACK FORM
Please give your opinions, suggestions, and possible corrections concerning the following aspects:
Instructions
Content of questions
Wording (spelling, grammar, choice of words etc.)
Sequence of questions
Form and layout
Please turn to the next page >>
92
Question difficulty
Any other comments, remarks, suggestions, etc?
93
A.2 Questionnaire (Finalversion)
SURVEY ON BRAND EXTENSIONS
Dear participant,
My name is Leon Phang, and I am conducting a survey for my final thesis in brand
management.
I am curious about what you think about various business-to-business brands and
their hypothetical brand extensions for the consumer market.
Please help me do this survey by taking the time to fill out this questionnaire, and I
shall be eternally grateful.
All information will be treated in the strictest confidence and results will be
produced in the form of aggregated data only.
Thank you for your time and help.
INSTRUCTIONS
As mentioned, this questionnaire is about strictly B2B brands and hypothetical B2C brand
extensions. A brand extension is a new product that uses an already existing brand name.
You will be asked about your opinion about 5 different brands. The questionnaire will take some
time and effort to fill out, so please try to do your best and make both your and my time worthwhile!
94
BRAND 1 OUT OF 5
• Consider the following brand
Please indicate your opinion on this brand by marking the appropriate number (1-7).
NOT AT ALL VERY
How knowledgeable are you about this brand? 1 2 3 4 5 6 7
POOR OUTSTANDING
NOT AT ALL VERY


Two hypothetical B2C brand extensions of the above brand will be mentioned.
Please indicate your opinion on these brand extensions by marking the appropriate number (1-7).
• Consider the brand extension
“24-Hour News TV Channel”

POOR OUTSTANDING



This brand extension will be difficult to make 1 2 3 4 5 6 7

Reuters is out to make short-term profits with this brand extension 1 2 3 4 5 6 7
”News Radio Station”

POOR OUTSTANDING




Reuters is out to make short-term profits with this brand extension 1 2 3 4 5 6 7
95
BRAND 2 OUT OF 5
• Consider the brand
NOT AT ALL VERY
POOR OUTSTANDING
NOT AT ALL VERY


Three hypothetical B2C brand extensions of the above brand will be mentioned.

“Private tax planning service”.
POOR OUTSTANDING

The competences of Ernst & Young are useful to make this extension 1 2 3 4 5 6 7
The resources of Ernst & Young are useful to make this extension 1 2 3 4 5 6 7
This brand extension is consistent with the Ernst & Young brand 1 2 3 4 5 6 7
This brand extension fits with my associations of the Ernst & Young brand 1 2 3 4 5 6 7

Ernst & Young is out to make short-term profits with this brand extension 1 2 3 4 5 6 7
96
“Institute of Accountancy” (Executive MBA education).
POOR OUTSTANDING



“Offshore banking service”.
POOR OUTSTANDING


97
BRAND 3 OUT OF 5
NOT AT ALL VERY
POOR OUTSTANDING
NOT AT ALL VERY


“Marketing & Advertising Books”

POOR OUTSTANDING


Saatchi & Saatchi is out to make short-term profits with this brand extension 1 2 3 4 5 6 7
“Institute of Advertising” (executive training courses)

POOR OUTSTANDING


Saatchi & Saatchi is out to make short-term profits with this brand extension 1 2 3 4 5 6 7
98
BRAND 4 OUT OF 5
NOT AT ALL VERY
POOR OUTSTANDING
NOT AT ALL VERY


”Portable MP3 Player”

POOR OUTSTANDING




Intel is out to make short-term profits with this brand extension 1 2 3 4 5 6 7
99
”Home Wireless (wi-fi) Network Hardware”

POOR OUTSTANDING




”Notebook PC”
POOR OUTSTANDING




100
BRAND 5 OUT OF 5
NOT AT ALL VERY
POOR OUTSTANDING
NOT AT ALL VERY


“Chronograph wristwatch”
POOR OUTSTANDING




Boeing is out to make short-term profits with this brand extension 1 2 3 4 5 6 7
“Flight simulator computer game”

POOR OUTSTANDING




101
“Travel luggage”
POOR OUTSTANDING




PERSONAL INFORMATION
Name (this will be used for administrational purposes only)
Age Gender: Male Female
Do you study business or economics? Yes No
Nationality
Dutch
German
Other, namely:
102
APPEN DIX B. SPSS OU TPU TS
B.1 Reliability Analyses
B.1.1 A ttitude tow ards brand extension
R E L I A B I L I T Y A N A L Y S I S - S C A L E (A L P H A)
Mean Std Dev Cases
1. EXTQUAL 4.7767 1.4177 1124.0

2. EXTTRIAL 3.9929 1.7649 1124.0
N of Cases = 1124.0
N of
Statistics for Mean Variance Std Dev Variables
Scale 8.7696 8.1597 2.8565 2
Item Means Mean Minimum Maximum Range Max/Min Variance

4.3848 3.9929 4.7767 .7838 1.1963 .3072
Reliability Coefficients 2 items
Alpha = .7439 Standardized item alpha = .7551
B.1.2 Innovativeness
Mean Std Dev Cases
1. INNOV1 5.0872 1.3162 1113.0

2. INNOV2 5.0916 1.5555 1113.0
3. INNOV3 5.1024 1.4200 1113.0
N of
SCALE 15.2812 14.6106 3.8224 3
Item-total Statistics
Scale Scale Corrected

Mean Variance Item- Alpha
if Item if Item Total if Item
Deleted Deleted Correlation Deleted
INNOV1 10.1941 7.8148 .6880 .8647

INNOV2 10.1896 6.4074 .7344 .8298
INNOV3 10.1788 6.5570 .8301 .7335
Reliability Coefficients
N of Cases = 1113.0 N of Items = 3
Alpha = .8667
103
B.1.3 Corporate SocialResponsibility
Mean Std Dev Cases
1. SOCIAL 3.8978 1.3330 1125.0

2. COMMUN 3.8667 1.3036 1125.0
N of Cases = 1125.0
N of
Scale 7.7644 6.0290 2.4554 2
Item Means Mean Minimum Maximum Range Max/Min Variance

3.8822 3.8667 3.8978 .0311 1.0080 .0005
Reliability Coefficients 2 items
Alpha = .8468 Standardized item alpha = .8470
B.1.4 Transfer
Mean Std Dev Cases
1. TRANSFE1 4.7384 1.5765 1124.0

2. TRANSFE2 4.8167 1.5170 1124.0
N of
SCALE 9.5552 8.9507 2.9918 2

TRANSFE1 4.8167 2.3012 .8706 .

TRANSFE2 4.7384 2.4854 .8706 .
Alpha = .9304
104
B.1.5 Brand concept consistency
Mean Std Dev Cases
1. BCC1 4.5009 1.6452 1114.0

2. BCC2 4.3348 1.7094 1114.0
N of
SCALE 8.8357 10.5723 3.2515 2

BCC1 4.3348 2.9219 .8790 .

BCC2 4.5009 2.7066 .8790 .
Alpha = .9352
B.1.6 D ifficult
Mean Std Dev Cases
1. DIFFI1 3.5067 1.5741 1125.0

2. DIFFI2 3.9636 1.4635 1125.0
N of
SCALE 7.4702 5.1497 2.2693 2

DIFFI1 3.9636 2.1419 .1150 .

DIFFI2 3.5067 2.4779 .1150 .
Alpha = .2058
105
B.2 Regression Analyses
B.2.1 Fullm odelbefore residualcentering
Variables Entered/Removed
Model Variables Entered Variables Method

Removed
1 Difficult, Environmental . Enter

concern, Parent brand
knowledge, Transfer,
Parent brand quality,
Corporate social
responsibility,
Innovative, Brand
concept consistency,
Innovativeness*difficulty,
Transfer*brand concept
consistency
a All requested variables entered.

b Dependent Variable: Attitude toward brand extension
Model Summary
Model R R Adjusted R Std. Error of

Square Square the Estimate
1 .785 .617 .613 .8894
a Predictors: (Constant), Difficult, Environmental concern, Parent brand knowledge, Transfer,

Parent brand quality, Corporate social responsibility, Innovative, Brand concept consistency,
Innovativeness*difficulty, Transfer*brand concept consistency
ANOVA
Model Sum of df Mean F Sig.
Squares Square
1 Regression 1429.08 10 142.908 180.642 .000

4
Residual 887.630 1122 .791
Total 2316.71 1132

4
a Predictors: (Constant), Difficult, Environmental concern, Parent brand knowledge, Transfer,

Parent brand quality, Corporate social responsibility, Innovative, Brand concept consistency,
Innovativeness*difficulty, Transfer*brand concept consistency
106
Coefficients
Unstandardized Standardized t Sig. Collinearity

Coefficients Coefficients Statistics
Model B Std. Beta Tolerance VIF
Error
1 (Constant) -.521 .340 -1.532 .126
Parent brand -2.254E-02 .019 -.031 -1.200 .230 .524 1.908

knowledge
Parent brand quality 7.587E-02 .033 .064 2.314 .021 .448 2.234
Environmental concern 5.802E-02 .028 .054 2.092 .037 .513 1.951
Corporate social .181 .030 .155 5.960 .000 .507 1.971
responsibility
Innovative .162 .050 .144 3.242 .001 .173 5.795
Transfer .310 .049 .324 6.301 .000 .129 7.733
Brand concept .444 .060 .505 7.425 .000 .074 13.540
consistency
Innovativeness*difficulty -6.840E-04 .012 -.005 -.058 .954 .053 18.912
Transfer*brand concept -1.552E-02 .011 -.145 -1.444 .149 .034 29.714
consistency
Difficult -5.596E-02 .066 -.061 -.847 .397 .065 15.384
a Dependent Variable: Attitude toward brand extension
107
B.2.2 Fullm odelafter residualcentering

Removed
1 Innovativeness*difficulty . Enter
(resid), Transfer,
Environmental concern,
Transfer*brand concept
consistency (resid),
Parent brand
knowledge, Parent
brand quality,
Corporate social
responsibility,
Innovative, Brand
concept consistency,
Difficult

Model Summary
Model R R Adjusted Std.

Square R Error of
Square the
Estimate
1 .785 .617 .613 .8894
a Predictors: (Constant), Innovativeness*difficulty (resid), Transfer, Environmental concern,

Transfer*brand concept consistency (resid), Parent brand knowledge, Parent brand quality,
Corporate social responsibility, Innovative, Brand concept consistency, Difficult
ANOVA

Squares Square
1 Regression 1429.084 10 142.908 180.642 .000
Residual 887.630 1122 .791
Total 2316.714 1132
a Predictors: (Constant), Innovativeness*difficulty (resid), Transfer, Environmental concern,

Transfer*brand concept consistency (resid), Parent brand knowledge, Parent brand quality,
Corporate social responsibility, Innovative, Brand concept consistency, Difficult
108
Coefficients

Error
1 (Constant) -.254 .348 -.730 .466
Parent brand -2.254E-02 .019 -.031 -1.200 .230 .524 1.908
knowledge
Environmental concern 5.802E-02 .028 .054 2.092 .037 .513 1.951
responsibility
Innovative .159 .032 .142 5.046 .000 .432 2.316
Transfer .252 .037 .263 6.814 .000 .230 4.354
Brand concept .366 .028 .416 13.283 .000 .347 2.879
consistency
Difficult -5.596E-02 .066 -.061 -.847 .397 .065 15.384
Transfer*brand concept -1.552E-02 .011 -.028 -1.444 .149 .884 1.131
consistency (resid)
Innovativeness*difficulty -6.840E-04 .012 -.004 -.058 .954 .072 13.925
(resid)
109
B.2.3 Fullm odelafter residualcentering and om itted m oderator variable

Removed
1 Service brand, . Enter

Difficult,
Environmental
concern,
Transfer*brand
concept
consistency[Residual],
Parent brand quality,
Transfer, Parent
brand knowledge,
Corporate social
responsibility,
Innovative, Brand
concept consistency

Model Summary
Model R R Adjusted R Std. Error
Square Square of the
Estimate
1 .786 .617 .614 .8892
a Predictors: (Constant), Service brand, Difficult, Environmental concern, Transfer*brand concept

consistency[Residual], Parent brand quality, Transfer, Parent brand knowledge, Corporate social
responsibility, Innovative, Brand concept consistency
ANOVA
Squares Square
1 Regression 1429.521 10 142.952 180.786 .000
Residual 887.193 1122 .791
Total 2316.714 1132
a Predictors: (Constant), Service brand, Difficult, Environmental concern, Transfer*brand concept

consistency[Residual], Parent brand quality, Transfer, Parent brand knowledge, Corporate social
responsibility, Innovative, Brand concept consistency
110
Coefficients

Error
1 (Constant) -.200 .177 -1.131 .258
Parent brand -2.506E-02 .019 -.034 -1.318 .188 .512 1.955
knowledge
Innovative .148 .034 .132 4.306 .000 .362 2.766
responsibility
Environmental 5.843E-02 .028 .054 2.107 .035 .513 1.951
concern
Transfer .256 .030 .268 8.412 .000 .337 2.964
Brand concept .368 .028 .418 13.335 .000 .348 2.877
consistency
Difficult -5.757E-02 .019 -.063 -3.029 .003 .785 1.274
Transfer*brand -1.622E-02 .011 -.030 -1.505 .133 .878 1.139
concept
consistency[Residual]
Service brand -5.174E-02 .069 -.018 -.746 .456 .585 1.710
111
B.2.4 Regression at brand level:brand 1
Model Summary
R R Square Adjusted R Std.

Square Error of
the
Estimate
Model +Brand =
Reuters
(Selected)
1 .727 .528 .507 .9580
a Predictors: (Constant), Transfer*brand concept consistency[Residual], Environmental concern,

Difficult, Parent brand knowledge, Innovative, Parent brand quality, Brand concept consistency,
Corporate social responsibility, Transfer
ANOVA

Squares Square
1 Regression 201.360 9 22.373 24.379 .000
Residual 179.874 196 .918
Total 381.234 205

Difficult, Parent brand knowledge, Innovative, Parent brand quality, Brand concept consistency,
Corporate social responsibility, Transfer
c Selecting only cases for which +Brand = Reuters
112
Coefficients

Error
1 (Constant) .684 .550 1.243 .215
Parent brand 6.914E-02 .052 .091 1.329 .186 .508 1.967
knowledge
Parent brand quality -7.240E-02 .078 -.067 -.923 .357 .454 2.200
Innovative 8.317E-02 .080 .067 1.041 .299 .584 1.712
responsibility
Environmental 5.707E-02 .077 .055 .744 .458 .439 2.278
concern
Transfer .233 .096 .202 2.416 .017 .343 2.913
Brand concept .352 .081 .349 4.369 .000 .378 2.643
consistency
Difficult -7.587E-02 .052 -.082 -1.471 .143 .778 1.285
Transfer*brand 1.844E-02 .037 .033 .496 .621 .556 1.799
concept

b Selecting only cases for which +Brand = Reuters
113
Model Summary
R R Square Adjusted Std. Error

R Square of the
Estimate
Model +Brand =
Ernst &
Young
(Selected)
1 .805 .649 .632 .8285
a Predictors: (Constant), Transfer*brand concept consistency[Residual], Innovative, Difficult,

Corporate social responsibility, Parent brand knowledge, Environmental concern, Brand concept
consistency, Parent brand quality, Transfer
ANOVA

Squares Square
1 Regression 248.338 9 27.593 40.195 .000
Residual 134.551 196 .686
Total 382.888 205
a Predictors: (Constant), Transfer*brand concept consistency[Residual], Innovative, Difficult,

Corporate social responsibility, Parent brand knowledge, Environmental concern, Brand concept
consistency, Parent brand quality, Transfer
c Selecting only cases for which +Brand = Ernst & Young
114
Coefficients

Error
1 (Constant) -.293 .426 -.687 .493
Parent brand -9.209E-02 .044 -.110 -2.074 .039 .633 1.579
knowledge
Parent brand quality .121 .068 .104 1.777 .077 .527 1.898
Innovative .204 .073 .176 2.813 .005 .460 2.174
responsibility
Environmental 9.772E-02 .060 .086 1.618 .107 .635 1.574
concern
Transfer .287 .081 .243 3.567 .000 .385 2.598
Brand concept .301 .061 .337 4.947 .000 .387 2.586
consistency
Difficult -.132 .042 -.154 -3.149 .002 .749 1.335
Transfer*brand -1.581E-03 .033 -.002 -.047 .962 .696 1.437
concept

b Selecting only cases for which +Brand = Ernst & Young
115
Model Summary
R R Square Adjusted R Std. Error

Square of the
Estimate
Model +Brand =
Saatchi &
Saatchi
(Selected)
1 .785 .617 .599 .7307
a Predictors: (Constant), Transfer*brand concept consistency[Residual], Corporate social

responsibility, Difficult, Parent brand knowledge, Brand concept consistency, Parent brand quality,
Environmental concern, Transfer, Innovative
ANOVA

Squares Square
1 Regression 168.389 9 18.710 35.040 .000
Residual 104.656 196 .534
Total 273.046 205

responsibility, Difficult, Parent brand knowledge, Brand concept consistency, Parent brand quality,
Environmental concern, Transfer, Innovative
c Selecting only cases for which +Brand = Saatchi & Saatchi
116
Coefficients


Error
1 (Constant) -.724 .463 -1.564 .119
Parent brand 3.739E-02 .038 .063 .973 .332 .473 2.115
knowledge
Parent brand quality 7.194E-02 .093 .067 .773 .441 .258 3.877
Innovative .101 .105 .085 .963 .337 .253 3.951
responsibility
Environmental -.110 .082 -.108 -1.334 .184 .297 3.367
concern
Transfer .316 .090 .284 3.523 .001 .300 3.334
Brand concept .312 .078 .294 4.024 .000 .367 2.722
consistency
Difficult -6.868E-02 .044 -.074 -1.553 .122 .866 1.155
Transfer*brand -1.385E-02 .040 -.020 -.342 .732 .573 1.746
concept

b Selecting only cases for which +Brand = Saatchi & Saatchi
117
Model Summary
R R Square Adjusted Std. Error

R Square of the
Estimate
Model +Brand =
Intel
(Selected)
1 .770 .594 .575 .9211

Parent brand knowledge, Parent brand quality, Difficult, Brand concept consistency, Corporate
social responsibility, Innovative, Transfer
ANOVA

Squares Square
1 Regression 242.933 9 26.993 31.817 .000
Residual 166.282 196 .848
Total 409.215 205

Parent brand knowledge, Parent brand quality, Difficult, Brand concept consistency, Corporate
social responsibility, Innovative, Transfer
c Selecting only cases for which +Brand = Intel
118
Coefficients

Error
1 (Constant) -1.092 .565 -1.934 .055
Parent brand -1.269E-02 .056 -.012 -.226 .822 .718 1.393
knowledge
Innovative .188 .115 .113 1.629 .105 .433 2.308
responsibility
Environmental 2.237E-02 .080 .018 .281 .779 .534 1.873
concern
Transfer .375 .094 .361 4.005 .000 .254 3.930
Brand concept .226 .079 .251 2.847 .005 .268 3.734
consistency
Difficult -4.409E-02 .049 -.053 -.893 .373 .597 1.676
Transfer*brand -5.030E-03 .032 -.008 -.158 .874 .736 1.359
concept

b Selecting only cases for which +Brand = Intel
119
Model Summary

Square of the
Estimate
Model +Brand =
Boeing
(Selected)
1 .795 .632 .621 .9032
a Predictors: (Constant), Transfer*brand concept consistency[Residual], Parent brand quality,

Difficult, Environmental concern, Transfer, Parent brand knowledge, Brand concept consistency,
Corporate social responsibility, Innovative
ANOVA

Squares Square
1 Regression 419.341 9 46.593 57.122 .000
Residual 243.889 299 .816
Total 663.230 308
a Predictors: (Constant), Transfer*brand concept consistency[Residual], Parent brand quality,

Difficult, Environmental concern, Transfer, Parent brand knowledge, Brand concept consistency,
Corporate social responsibility, Innovative
c Selecting only cases for which +Brand = Boeing
120
Coefficients

Error
1 (Constant) .346 .315 1.099 .273
Parent brand 4.881E-02 .046 .053 1.054 .293 .491 2.038
knowledge
Parent brand quality 3.492E-02 .075 .028 .466 .642 .337 2.969
Innovative 2.857E-02 .078 .024 .366 .714 .295 3.386
responsibility
Environmental 9.719E-02 .049 .108 1.993 .047 .417 2.398
concern
Transfer .155 .047 .174 3.280 .001 .435 2.299
Brand concept .404 .047 .481 8.590 .000 .392 2.551
consistency
Difficult -9.681E-03 .033 -.011 -.291 .771 .868 1.152
Transfer*brand -5.631E-02 .018 -.129 -3.066 .002 .692 1.446
concept

b Selecting only cases for which +Brand = Boeing
121
B.2.9 Regression: Service brands
Model Summary

Square of the
Estimate
Model Service
brand =
Service
brand
(Selected)
1 .766 .587 .580 .8580

Difficult, Parent brand knowledge, Innovative, Brand concept consistency, Corporate social
responsibility, Parent brand quality, Transfer
ANOVA

Squares Square
1 Regression 635.010 9 70.557 95.847 .000
Residual 447.573 608 .736
Total 1082.583 617

Difficult, Parent brand knowledge, Innovative, Brand concept consistency, Corporate social
responsibility, Parent brand quality, Transfer
c Selecting only cases for which Service brand = Service brand
122
Coefficients

Error
1 (Constant) -.173 .268 -.646 .518
Parent brand -3.744E-02 .023 -.057 - .111 .526 1.901
knowledge 1.596
Environmental 6.725E-02 .038 .062 1.777 .076 .557 1.794
concern
Innovative .176 .044 .144 4.033 .000 .530 1.886
responsibility
Transfer .287 .050 .256 5.705 .000 .336 2.972
Brand concept .334 .040 .345 8.315 .000 .396 2.526
consistency
Difficult -9.479E-02 .026 -.104 - .000 .806 1.241
3.582
Transfer*brand 4.065E-03 .020 .007 .200 .841 .643 1.556
concept

b Selecting only cases for which Service brand = Service brand
123
B.2.10 Regression: Product brands
Model Summary

Square of the
Estimate
Model Service
brand =
Product
brand
(Selected)
1 .805 .648 .642 .9233

responsibility, Difficult, Parent brand knowledge, Transfer, Parent brand quality, Environmental
concern, Innovative, Brand concept consistency
ANOVA

Squares Square
1 Regression 793.594 9 88.177 103.441 .000
Residual 430.482 505 .852
Total 1224.076 514

responsibility, Difficult, Parent brand knowledge, Transfer, Parent brand quality, Environmental
concern, Innovative, Brand concept consistency
c Selecting only cases for which Service brand = Product brand
124
Coefficients

Error
1 (Constant) -.219 .262 -.833 .405
Parent brand -1.426E-03 .035 -.001 -.041 .967 .613 1.632
knowledge
Environmental 6.479E-02 .041 .061 1.563 .119 .462 2.165
concern
Innovative .104 .059 .077 1.747 .081 .358 2.790
responsibility
Transfer .236 .041 .262 5.690 .000 .330 3.033
Brand concept .388 .039 .455 9.829 .000 .326 3.071
consistency
Difficult -2.327E-02 .028 -.026 -.846 .398 .761 1.315
Transfer*brand -2.920E-02 .014 -.057 - .042 .892 1.121
concept 2.042

b Selecting only cases for which Service brand = Product brand
125

206 BtoB Brandextensions

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

206 BtoB Brandextensions

Uploaded by

Copyright:

Available Formats

Consumer Evaluations of

Author: Leon Phang

CH APTER 1. IN TRODU CTION 9

1.1 Background:Leveraging the Brand Across Sectors 9

PART I. LITERATU RE REVIEW

CH APTER 2. CON SU M ER BRAN D EX TEN SION S 15

2.1 Introduction to the Chapter 15

CH APTER 3. BU SIN ESS-TO-BU SIN ESS BRAN DIN G 21

3.1 Introduction to the Chapter 21

PART II. CON CEPTU AL FRAMEW ORK

CH APTER 4. H YPOTH ESES 29

4.1 Introduction to the Chapter 29

CH APTER 5. RESEARCH DESIG N 36

5.1 Introduction to the Chapter 36

CH APTER 6. AN ALYSIS,RESU LTS AN D IN TERPRETATION S 46

6.1 Introduction to the Chapter 46

7. CON CLU SION S 67

7.1 Summary of the Findings 67

Table 1.1 Companies Formerly Operating in Business-to-Business Markets Only 11

LIST OF FIG U RES

Figure 3.1 An Integrated Approach to Communications 23

1.1 Background:Leveraging the Brand Across Sectors

Com pany nam e Country Founded Initialbusiness at year 1

Microsoft US 1975 BASIC computer programming language

Source: Com pany w ebsites

TABLE 1.2 Expansion to Consumer Markets

Driver for strategic

Microsoft 1981 MS-DOS BASIC computer programming Innovation

IBM 1981 IBM 5150 Personal Mainframes, industrial computing Innovation

Philips 1927 Radio Lamps, medical X-ray tubes Innovation

Mitsubishi 1917 Automobiles H eavy industries Diversification and

Source:Com pany w ebsites

1.2 Problem Statem ent

Can business-to-business brands be extended into the consum er m arket?

1.3 Contributions of the Study

1.4 Thesis Outline

Chapter 2. Consumer Brand Extensions 15

Chapter 3. Business-to-Business Branding 21

CON SU M ER BRAN D EX TEN SION S

2.1 Introduction to the Chapter

The discussion on brand extensions begins with a consideration of a firm’s sources of

2.2 Consum er Brand Extensions

According to Keller (2003, p. 577), a brand extension is defined as “when a firm

The benefits of brand extensions is a much-discussed subject. Keller (2003)

2.2.2.1 Benefits related to brand extension acceptance

Brand extensions allow consumers to draw conclusions and form expectations

2.2.2.2 Benefits relating to the parentbrand im age

2.2.3 D raw backs

2.3 Concepts for Evaluating Brand Extensions

2.3.1 Extension reaction

2.3.2 Spillover and substitution effects

2.3.3 Categoricaland piecem ealevaluation processes

To understand how consumers evaluate new brand extensions, categorization

2.3.4 Brand-specific associations

A brand-specific association is defined as an attribute or benefit that differentiates

A brand can be leveraged by means of brand extensions. Two types of benefits

BU SIN ESS-TO-BU SIN ESS BRAN DIN G

3.1 Introduction to the Chapter

3.2 Corporate Branding Decisions

3.2.1 Corporate strategy

3.2.2 Corporate brand identity

FIG U RE 3.1 An Integrated Approach to Communications

Corporate mission Marketing Customers