ARTICLE IN PRESS

Journal of Retailing and Consumer Services 14 (2007) 35–44

www.elsevier.com/locate/jretconser

Does attitudinal loyalty influence behavioral loyalty?

A theoretical and empirical study
Subir Bandyopadhyay, Michael Martell
School of Business and Economics, Indiana University Northwest, 3400 Broadway, Gary, Indiana 46408, USA

Abstract

Despite a plethora of studies on brand loyalty spanning over last three decades, the research paradigm is unique in its inability to
produce generalizable results. The research suffers from a lack of agreement over the operationalization of the construct of the brand
loyalty. Many researchers [e.g., Kahn, B.E., Kalwani, M.U., Morrison, D.G., 1986. Measuring variety seeking and reinforcement
behaviors using panel data. Journal of Marketing Research 23, 89–100; Ehrenberg, A.S.C., Goodhardt, G.J., Barwise, P.B., 1990.
Double jeopardy revisited. Journal of Marketing 54, 82–91] have defined brand loyalty only from a behavioral perspective. They assumed
that repeat purchasing can capture the loyalty of a consumer towards the brand of interest. Other researchers [e.g., Day, G.S., 1969. A
two-dimensional concept of brand loyalty. Journal of Advertising Research 9, 29–35; Dick, A.S., Basu, K., 1994. Consumer loyalty:
towards an integrated conceptual approach. Journal of the Academy of Marketing Science 22 (2), 99–113; Baldinger, A.L., Rubinson, J.,
1996. Brand loyalty: the link between attitude and behavior. Journal of Advertising Research 36 (6), 22–34; Rundle-Thiele, S.R., Bennett,
R., 2001. A brand for all seasons: A discussion of loyalty approaches and their applicability for different markets. Journal of Product and
Brand Management 10 (1), 25–37; Rundle-Thiele, S.R., 2005. Elaborating customer loyalty: exploring loyalty to wine retailers. Journal
of Retailing and Consumer Services 12 (5), 333–344], however, have suggested that attitude should be included along with behavior to
define loyalty. In particular, Dick and Basu (1994) were precise in suggesting that a favorable attitude and repeat purchase were required
to define loyalty. By viewing loyalty as an attitude–behavior relationship in their framework, Dick and Basu (1994) were able to
investigate the concept from a causal perspective thereby permitting the identification of the factors that influence loyalty.
Our research extends the idea proposed by Dick and Basu (1994) in two ways. First, we provide a conceptually clear, and an
operationally meaningful way of segmenting the market on the basis of attitudes that govern this behavior. To be specific, we offer a
unique way to measure attitudinal loyalty. And we use a unique survey data with a large sample of 1800 respondents that includes both
behavioral (purchase) patterns and attitudes of the respondents for all major brands of toothpaste, to demonstrate that behavioral
loyalty is influenced by attitudinal loyalty across many brands of the toothpaste category. Finally, we propose that a third behavioral
pattern—non-user—should be also included along with two known behavioral patterns viz. single user and multiple user. This is because a
non-user has the potential to become a consumer in future.
r 2006 Elsevier Ltd. All rights reserved.

Keywords: Brand loyalty; Behavioral loyalty; Attitudinal loyalty; Market segmentation

1. Introduction understanding of the loyalty phenomenon. The research

In spite of the numerous studies devoted to under-
standing the phenomenon of brand loyalty over the past
three decades, the research paradigm is unique in its
inability to produce generalizable results. Previous research
in the field has been unable to contribute a clear conceptual
Corresponding author. Tel.: +1 219 980 6900; fax: +1219 9860 6916.
E-mail address: sbandyop@iun.edu (S. Bandyopadhyay).
also lacks a clear agreement over the operationalization of
the construct of the brand loyalty.
Broadly, there are two schools of thought when it comes
to define and operationalize brand loyalty. Many research-
ers (e.g. Ehrenberg et al., 1990; Blattberg and Sen, 1974;
Kahn et al., 1986) have defined brand loyalty strictly from
a behavioral perspective. A common theme across this
stream of work has been the attempt to look for a
surrogate behavioral measure to operationalize brand

0969-6989/$ - see front matter r 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.jretconser.2006.03.002
 ARTICLE IN PRESS
36 S. Bandyopadhyay, M. Martell / Journal of Retailing and Consumer Services 14 (2007) 35–44
loyalty. The major assumption here is that this (repeat)
purchasing could capture the loyalty of a consumer
towards the brand of interest. Thus, while some researchers
have observed purchasing patterns and made conclusions
based on the proportion of purchases devoted to a
particular brand (Cunningham, 1956; Blattberg and Sen,
1974), others have focused on the purchase sequence
(Kahn et al., 1986; McConell, 1968). In fact, many
researchers have struggled over the years to (1) to
distinguish between repeat purchase and brand loyalty,
and (2) to define brand loyalty—a complex multidimen-
sional phenomenon—on a single behavioral dimension
(Jacoby and Kyner, 1973).
Dick and Basu (1994) were precise in suggesting that a
favorable attitude and repeat purchase were required to
define loyalty. By viewing loyalty as an attitude–behavior
relationship in their framework, they were thus able to
investigate the concept from a causal perspective—thereby
permitting the identification of the antecedents of loyalty.
Their research, however, offers only a theoretical frame-
work, and did not provide any empirical validation of their
theory. In a recent paper, East et al. (2005) empirically
tested Dick and Basu’s (1994) theory that consumer loyalty
is characterized by a favorable attitude and repeat
purchase.1 Using data collected across countries (Britain
and New Zealand) and industries (cars and fashion stores),
they demonstrate that the combination concept of loyalty
(i.e., attitude plus behavior) is of limited value, particularly
in predicting phenomena such as recommendation, search
and retention. More empirical tests of the model under
different contexts (e.g., products, services, and usage
occasions) are needed, to thoroughly evaluate Dick and
Basu’s (1995) model.
Our research extends the work done by Dick and Basu
(1994) in two ways. First, we provide a conceptually clear,
and an operationally meaningful way of segmenting the
market on the basis of attitudes that influence the brand
purchase behavior. To be specific, we offer a unique way to
measure attitudinal loyalty. And we use a unique survey
data with a large sample of 1800 respondents that includes
both behavioral (purchase) patterns and attitudes of the
respondents for all major brands of tooth paste, to
demonstrate that behavioral loyalty is influenced by
attitudinal loyalty across many brands of the toothpaste
category.
Finally, we extend Dick and Basu’s (1994) model by
introducing a new segment of non-users. Dick and Basu
(1994) contend that the lowest form of loyalty is that of
someone with a low attitude and a low repeat purchase
pattern, but our position is that the non-users with a weak
attitude would be at the bottom of the ‘‘loyalty ladder’’. It
is important that non-users be included because they have
the potential to become consumers, especially if it is
unknown as to why they are non-users. Dick and Basu
1
We thank an anonymous reviewer for letting us know of this new
publication.
(1994) stated that, ‘‘customer loyalty is viewed as the
strength of the relationship between an individual’s relative
attitude and their repeat patronage’’. To a point this is
true; however we contend that a non-user with a strong
attitude towards a product could potentially be a loyal
customer, but due to financial reasons or lack of market
penetration or some other reasons, her purchase situation
was affected. We would concede that there is little that can
be done with a person who is a non-user and is
characterized with a weak attitude, and therefore few
resources should be exhausted in an attempt to convince
these individuals to switch brands. It is the non-user with a
strong attitude that should interest the brand manager.
This, in essence, is in consonance with one of the basic
functions of marketing: generate new customers who are
willing and able to purchase the brand. Our position is
supported by Uncles et al. (2003) who stated that ‘‘it is
possible that a loyalty program could be offered to people
who do not buy the target brand (but do buy from the
category)’’.
Our study thus is an attempt to improve earlier works in
the following ways:
(1) Behaviorally, we give a strict definition to a loyal
customer, i.e. a buyer of only one brand. Thus, as
defined by us, a loyal customer is one who buys only
one brand over the last 1 year. We call these customers
‘‘single users’’. We also include the ‘‘multiple users’’
(people who buy more than one brand) and ‘‘non-
users’’ in our analysis to investigate how attitudinal
influences may differ for these three types of con-
sumers.
(2) We define ‘‘favorable attitudes’’ much more explicitly
for the class of product category under consideration.
Our focus for this research is on low involvement
purchasing, and for such purchase behavior we look at
past consumer behavior research to understand how
the attitudes could be operationalized. Our definition of
strong and weak attitudes is based on the attribute
frequency levels. In other words, we argue that the
brand attitude will be stronger when consumers
associate more number of attributes with the
brand. We elaborate on this issue later on in the
‘‘Operationalizing Attitudinal Loyalty’’ section in
page 15.
(3) We propose a conceptually clear, and operationally
meaningful way of segmenting the market on the basis
of not only the behavior patterns (i.e. single-, multiple-,
and non-users), but also on the basis of the strength of
antecedent attitudes that govern this behavior.
The rest of the paper is organized as follows. In the next
section, we briefly review the literature on loyalty. Because
the loyalty literature has been thoroughly reviewed by
other researchers such as Rundle-Thiele (2005) and Uncles
et al. (2003), we will focus mostly on the earlier work done
on attitudinal loyalty, and skim the literature on behavioral
 ARTICLE IN PRESS
S. Bandyopadhyay, M. Martell / Journal of Retailing and Consumer Services 14 (2007) 35–44
loyalty.2 Next we discuss the conceptual framework used in
our study followed by the hypotheses to be tested. We then
describe the methodology used. In particular, we describe
the data obtained from a large multinational company. In
the next section, we discuss the results followed by a
discussion on managerial implications. We conclude with a
discussion on the scope of future research and limitations
of the study.
2. Review of the loyalty literature
Earlier research on brand loyalty has proposed two
perspectives in defining and operationalizing brand loyalty:
behavioral and attitudinal.
2.1. Behavioral loyalty
Many researchers (e.g., Ehrenberg et al., 1990; Kahn et
al., 1986; Ehrenberg, 2000) believe that repeat purchasing
can capture the loyalty of a consumer towards the brand of
interest. Similarly, Sharp et al. (2002) suggest that attitude
is not relevant to determining brand loyalty. Their research
is purely based on the notion that there is no ‘‘true’’
definition to brand loyalty, and that debating this topic is a
‘‘waste of time’’.
2.2. Limitation of the ‘‘loyalty is behavior’’ approach
It was Newman (1966) who first challenged the approach
of equating behavior patterns with preferences to infer
loyalty. Other researchers (e.g., Day, 1969; Coulson, 1966)
have highlighted the distinction between ‘‘spurious loyalty’’
as captured by the behavioral patterns, and ‘‘true/inten-
tional loyalty’’ that extends beyond the regular purchasing
of a brand. In a more recent study, Oliver (1999) suggested
that customer satisfaction developed by way of product
usage, is a necessary step in loyalty formation. But it
becomes less significant as loyalty begins to set through
other mechanisms such as individual fortitude (i.e., the
degree to which the consumer resists competitive pressure
to switch over to another brand) and social bonding (i.e.,
the degree to which the community or the society supports
the consumer to remain loyal).
Existence of situational factors (such as stock-out and
non-availability), intrinsic factors (such as individual
fortitude) or socio-cultural factors (such as social bonding),
that can differentiate between repeat purchase and brand
loyalty, have prompted recent works that have called for
understanding and operationalizing brand loyalty beyond
behavioral measures. For example, Dick and Basu (1994)
stressed upon the need to extend the concept of brand
loyalty to include attitudinal influences.
2
We thank an anonymous reviewer for making this suggestion.
37
2.3. Attitudinal loyalty
The need for the inclusion of ‘‘attitude’’ along with
behavior to define brand loyalty has been felt by many
researchers. Day (1969) was perhaps the first to recognize
and articulate this need. Thereafter, Jacoby and Kyner
(1973) defended Jacoby’s (1971) definition of brand loyalty.
This definition was expressed as a set of six necessary and
collectively sufficient conditions. According to them, brand
loyalty is (1) the biased (i.e., nonrandom) (2) behavioral
response (i.e., purchase), (3) expressed over time (4) by
some decision making unit, (5) with respect to one or more
alternative brands out of a set of such brands, and (6) is a
function of psychological (decision making, evaluative)
processes. The authors stated that it is the evaluation
process (the sixth condition) that makes an individual
develop a commitment towards a brand. It is this notion of
commitment, they argued, that provides an essential basis
of differentiating brand loyalty from other forms of repeat
purchasing behavior.
In a similar vein, Baldinger and Rubinson (1996)
expressed their belief that brand loyalty could be better
understood by extending the behavioral definitions of
loyalty so as to encompass attitudes (along with behavior)
for measuring loyalty. Their premise is that once con-
sumers are classified behaviorally (in terms of their loyalty
patterns), it is possible to link these behavioral segments to
the underlying attitudes towards the brands. Similarly,
Chaudhuri and Holbrook (2001) examine two aspects of
brand loyalty, purchase loyalty and attitudinal loyalty, as
linking variables in the sequence of effects from brand trust
and brand affect to brand performance (i.e., market share
and relative price). Their study results indicate that when
product- and brand-level variables are controlled for,
brand trust and brand affect combine to determine
purchase loyalty and attitudinal loyalty. Purchase loyalty,
in turn, leads to greater market share and attitudinal
loyalty leads to a higher relative price for the brand. In
another notable paper on brand loyalty, Dick and Basu
(1994) mentioned that the behavioral measures of loyalty
were inadequate in understanding the factors underlying
brand loyal purchase behavior. The behavioral definitions
were viewed as being insufficient towards explaining how
and why brand loyalty is developed and/or modified. Their
conceptual work was therefore developed on the notion
that brand loyalty encompasses both consumer attitude
and repeat purchase behavior. Amine (1998) also sup-
ported this view stating that while repeat purchases of a
brand over time is an expression of customer loyalty, it is
incomplete unless complemented with a positive attitude
towards the brand. The positive attitude towards the brand
ensures that previous purchase behavior will continue.
In a more recent study, Rundle-Thiele (2005) provides
insight into how and why consumers are loyal by
identifying the multiple dimensions of loyalty. The dimen-
sions that she explored are: situational loyalty, resistance to
competing offers, propensity to be loyal, attitudinal
 ARTICLE IN PRESS
38 S. Bandyopadhyay, M. Martell / Journal of Retailing and Consumer Services 14 (2007) 35–44
loyalty, and complaining behavior. Her research based on
a survey of wine retailers in Australia, indicates that
consumers can be loyal in different ways thereby demon-
strating the superiority of a multidimensional model of
consumer loyalty over a unidimensional model. Other
authors (e.g. East et al., 2005; Rundle-Thiele and Bennett,
2001) make a similar argument against the use of a single
brand loyalty measure. They suggest that the measures
should be classified according to market types: consumable
goods, durable goods and services. They argue that
consumable goods markets, which are typically stable,
have a high rate of brand switching, and low involvement
and risk, behavioral measures are appropriate for predict-
ing future brand loyalty levels. Conversely, for durable
goods markets, buyers do not frequently switch back and
forth between brands. Hence, in any given time period,
buyers are sole-loyal or dual-loyal. For example, a given
product (such as a Sanyo fax machine) may be used by all
members of a family. Hence behavioral measures of loyalty
may not be sufficient to capture all these characteristics.
Lastly, they discuss loyalty in the service context. As the
perceived risk increases, the likelihood of loyalty to one
brand increases. Thus a behavioral loyalty measure is not
sufficient in the service context. We find support of their
observation in several other studies on service quality and
customer satisfaction that use multiple dimensions of
loyalty. Notable studies in this genre include Bloemer et
al. (1999), Ganesh et al. (2000) and Zeithaml et al. (1996).3
2.4. Limitations of attitudinal loyalty measurement
Despite all these calls for better understanding of brand
loyalty through the consideration of attitudes as well as
behavior, we feel that the meaning and operationalization
of brand loyalty are not thoroughly understood. Most of
the shortfalls have been in the understanding and inclusion
of attitudes in defining loyalty. Many attitudinal loyalty
measures are considered to be incapable of distinguishing
between brands (Dall’Olmo Riley et al., 1997). Day (1969)
stated the importance of attitudes for defining loyalty, but
he also believed that imposition of attitudinal criteria
would make loyalty a brand specific concept instead of a
general concept describing product-class behavior. This
assumption can be viewed as somewhat inconsistent in that
similar attitudinal patterns might be observable across all
(or most of) the brands of a product category, thereby
increasing the generalizability of the results. Wansink et al.
(2001) supported this assumption by stating that product
category loyalty is related to brand level switching within
the same product category, and not switching within the
same brand. Furthermore, the construction of the attitude
scale that Day (1969) used was not described in detail, and
the trial and error fitting of the constants also cast doubt
on the accuracy of the research. These limitations were
3
We thank an anonymous reviewer for making this important
observation.
recognized by Day (1969) himself who stated that he was
unsure of what weights should be given to the relative
influence of attitude and behavior components in the
loyalty score.
Similarly, though Jacoby and Kyner (1973) did attempt
to defend their six-condition definition of loyalty, the
definitions and understanding of attitudes seemed unclear.
They stated that their research design guaranteed a
comparative evaluative judgment, but unfortunately no
details of attitude conceptualization and measurement were
given. Baldinger and Rubinson (1996) did emphasize
the importance of attitudes for measuring loyalty. They
categorized people into two groups—‘‘prospects’’ and
‘‘vulnerables’’—by comparing the strengths of their atti-
tudes and behavior. Comparing two different constructs on
a common scale, we feel, can lead to the possibility of
inaccurate data. Again, Baldinger and Rubinson (1996) did
not present a sufficient explanation of their attitude scale,
and of how they trichotomized the scale to define low,
moderate, and high attitudes.
Other research has attempted to further understand an
individual’s personality traits, in contrast to brand-specific
traits, and the effects that those traits have on attitudinal
loyalty (see, for example, Raju, 1980; Mellens et al., 1996).
We feel, as do Sheth et al. (1999), that personality traits are
often too general and do not offer reliable empirical
measures that would help to explain product- and brand-
specific purchasing behavior. Bennett and Rundle-Thiele
(2002) also concluded that only brand-specific measures
will be relevant when determining brand loyalty. It is
ultimately more important to understand which product
attributes are important to the individuals in a household.
Determining which observable product attributes are
important to consumers allows product managers to
effectively position products, and aids in determining
whether or not to extend a product line at the risk of
cannibalization.
3. Conceptual framework
We use a modified typology of loyalty in Table 1 that is
based on the work of Dick and Basu (1994) and Uncles
et al. (1994). In line with Uncles et al. (1994) we believe
that a consumer may use only one brand (single user),
which has also been termed monogamous (e.g. Uncles
et al., 1994), multiple brands (multiple user) also referred to
Table 1
Consumer brand loyalty classification according to their behavioral and
attitudinal characteristics
Attitudinal Behavioral loyalty
loyalty
Single users Multiple users Non-users
Strong Brand loyal Variety seeker Potential buyer
Weak Constrained buyer Deal prone Indifferent
 ARTICLE IN PRESS
S. Bandyopadhyay, M. Martell / Journal of Retailing and Consumer Services 14 (2007) 35–44
as polygamous (See Uncles et al. 1994) in a category. Our
typology differs from that of Dick and Basu (1994) in that
ours includes an additional segment for ‘‘non-users’’.
According to Dick and Basu (1994), the lowest form of
loyalty is demonstrated by a person with a low attitude and
a low repeat purchase pattern while Uncles et al. (1994)
acknowledge that customers may be promiscuous or not
loyal to any brand. But our position is that the non-users
with a weak attitude would be at the bottom of the loyalty
hierarchy. We contend that non-users be included in the
loyalty hierarchy because they have the potential to
become consumers, especially if reasons for non-use are
know and within a marketers control.
Multiple users, are the largest group and the most
common among consumers. They are the brand switchers
who have favorable attitudes towards several brands, and
often feel it unnecessary to devote total loyalty to one
brand as long as there are others that can ‘‘do the job’’.
This may be a reflection of the functional similarity of the
brands in a product category.
Finally, single users are those who display the highest
form of loyalty. We must, however, differentiate between
those with a strong attitude and considered truly ‘‘brand
loyal’’, and those with a weak attitude who may be forced
to purchase a brand (we call them ‘‘constrained buyers’’).
These constrained buyers may be purchasing a brand
strictly because they are without choice of competing
brands or the brand is continually being discounted.
Allenby and Rossi (1991) note that a person could be
perceived as being loyal to a brand through repeat
purchasing, but only when it is being heavily discounted.
If other choices became readily available, these customers
may slip all the way down the loyalty ladder and become
non-users. On the other hand, single users with strong
attitude are hardly influenced by the marketing mix tools.
Oliver (1997) supports this argument in his definition of
customer loyalty: ‘‘A deeply held commitment to rebuy or
repatronize a preferred product/service consistently in the
future, thereby causing repetitive same-brand or same-
brand set purchasing despite situational influences and
marketing efforts having the potential to cause switching
behavior’’. These customers provide a steady revenue
stream that becomes very predictable. Their past purchase
behavior has already been calculated into future sales, and
so the only way this group can provide increased revenue is
by convincing them to buy more of the brand. Unfortu-
nately, these customers comprise only a small portion
(approximately 10%) of all customers, and hence cannot be
the focal point of any marketing strategy.4 These customers
are also typically light category users. Accordingly,
marketers should focus on increasing the penetration rate
among these customers. It is important to note that these
customers are not necessarily drawn by every attribute of
the brand, but have over time built a relationship with
the brand itself. Fournier (1998) sees this loyalty as a
4
We thank an anonymous reviewer for highlighting this fact.
39
committed and affect-laden partnership between consu-
mers and brands.
4. Conceptual development
Based on the conceptual framework outlined above, we
suggest that true loyalty should be seen as comprising: (1) a
favorable repeat purchase (behavioral) pattern, and (2) a
favorable disposition (attitude) towards the brand.
4.1. Operationalizing behavioral loyalty
To operationalize behavioral pattern, the first of these
necessary and sufficient conditions, we group our survey
respondents into three behavioral categories—single users,
multiple users, and non-users. As these labels suggest,
single users exhibit highest level of (behavioral) loyalty as
they purchase only a single brand on every purchase
occasion; an intermediate (behavioral) loyalty is demon-
strated by users on multiple brands (multiple users), and
the least amount of behavioral loyalty is shown by those
respondents who do not use the brand even once over the
study period (non-users).
4.2. Operationalizing attitudinal loyalty
The behavioral classification, in support of our basic
hypothesis, should also exhibit parallel attitudinal mea-
sures to suggest that loyalty is true, and is not spurious
loyalty driven by marketing mix elements. Several
researchers have used attitudinal measures of consumer
loyalty. Notable among them are Bandyopadhyay et al.
(2005) Rundle-Thiele (2005), Chaudhuri and Holbrook
(2001), Mittal and Lasser (1998), and Berne et al. (1997). In
fact, Rundle-Thiele (2005) and Rundle-Thiele and Bennett
(2001) suggest that in consumable markets characterized by
high switching and low involvement and low risk,
behavioral measures can predict brand loyalty levels
appropriately. We propose here a simplified attitudinal
loyalty measure that is capable of accurately predicting
behavioral loyalty levels in low involvement purchasing
situations.
Typically, attitude of a consumer towards a brand is
operationalized by measuring consumer perceptions of the
‘‘overall rating’’ of the brand (Fishbein and Ajzen, 1975).
For the low involvement purchasing, however, a favorable
attitude towards a brand can be linked to the concept of
attribute frequency (Bettman and Park, 1980; Russo and
Dosher, 1983). Thus, brand liking can be equated with the
number of positive attributes a respondent believes the
brand possesses.
Our approach is also supported by Petty and Cacioppo
(1984) who posited that the number of arguments
presented in support of a position (for example, the
number of attributes a brand is believed to possess) could
dominate the argument importance (for example, brand
attitude strength) if subject involvement is low. Thus
 ARTICLE IN PRESS
40 S. Bandyopadhyay, M. Martell / Journal of Retailing and Consumer Services 14 (2007) 35–44

attribute frequency can be considered as a peripheral route Table 2

to persuasion. This hypothesis was further supported by
Alba and Marmorstein (1987). According to them, the net
number of positive attributes towards a brand could
influence judgment and choice if no other information is
available. They suggested that, according to the theory of
automation (see Hasher and Zacks, 1979, 1984), the
frequency information (i) acts continually (ii) cannot be
improved by practice (iii) cannot be inhibited (iv) does not
require conscious awareness and (v) drains minimal
cognitive resources.
These characteristics make the frequency heuristic very
different from other consumer decision heuristics as the
frequency information may be acquired with minimum
effort, or may even be acquired unconsciously. Thus it is
very likely that for low involvement product categories, the
frequency information may be used as a decision criterion
employed by the customer to reduce cognitive load. This is
to say that an average respondent can be said to have a
more favorable disposition towards Brand A vis-à-vis
towards Brand B if she believes Brand A has more positive
attributes than brand B. In fact, Bandyopadhyay and
Gupta (2004) have made a similar argument in their study
of attitudinal measures of double jeopardy. In line with the
approach suggested by them, and by Bettman and Park
(1980), Russo and Dosher (1983), Petty and Cacioppo
(1984), and Alba and Marmorstein (1987), we, therefore,
operationalize attitude strength of a brand (or brand
liking) by the number of positive attributes (or attribute
score) associated with the brand. A person has a stronger
(or weaker) attitude toward a brand when she believes that
the brand possesses more (or less) positive attributes (we
call it the attribute scores).
Put together, these two conditions (behavior and
attitude) that have been suggested to operationalize true
loyalty therefore suggest that the single users should
demonstrate the highest attribute scores, followed by the
attribute scores of multiple users and the non-users in that
order. We, therefore, test the following hypothesis:
H1a. The mean attribute score is greater for single users
than for multiple users.
H1b. The mean attribute score is greater for multiple users
than for non-users.
5. The data
The data for testing the research hypotheses was
provided by Procter & Gamble, a leading consumer goods
manufacturer. The data set contains the results from a
survey conducted in June 2000 of 1800 panelists on ten
major brands of toothpaste. The list of brands along with
their market shares is given in Table 2. Those who did not
respond to one or more questions in the survey were not
included in our analysis. Out of 1800, 1096 panelists
List of toothpaste brands and their market shares
Brand Market share (%) Attribute scoresa
1. Crest Tartar 38 8.36
2. Crest Original 18 9.14
3. Colgate Original 10 5.86
4. Aquafresh Original 9 4.21
5. Colgate Tartar 7 5.20
6. Close Up 7 1.40
7. Macleans 4 3.40
8. Sensodyne 3 1.00
9. Aquafresh Tartar 3 3.87
10. Arm & Hammer 1 3.00
a
An attribute score of 8.36 for Crest Tartar indicates that each of the
respondents who have purchased the brand at least once during the last
year believes that Crest Tartar has on average 8.36 of the attributes among
30 attributes.
Table 3
List of behavioral questions

Brands of toothpaste seen advertised on TV in past 4 weeks

Brands of toothpaste seen advertised in magazines in past 4 weeks

How often brush teeth (self)

How often brush teeth (spouse)

How often brush teeth (children less than 12 years)

How often brush teeth (children older than 12 years)

Last time used (each of the brand names)

How often bought (each of the brand names)
responded to all questions included in the survey. Thus,
our analysis is based on the feedback from 1096 panelists.
The survey includes eight behavioral questions to
measure the respondents’ usage patterns and their satisfac-
tion towards various brands over a period of six months,
and 30 attribute-related questions on respondents’ beliefs
about these brands (e.g., ‘‘leaves mouths fresh,’’ ‘‘reverses
tooth decay,’’ ‘‘suitable for children,’’ ‘‘prevents root
cavities,’’ ‘‘good value for money,’’ ‘‘has a low price,’’
etc.). The survey used a two-point (yes/no) scale to measure
if respondents felt that a given brand had each of the 30
attributes. Table 3 includes the list of behavioral questions
and Table 4 contains the list of attributes.
Although the behavioral measures are ‘‘recalled mea-
sures’’ instead of actual measures, the potential lack of
reliability in such measures is minimized because of the use
of a free-choice as opposed to a forced-choice question-
naire format (Bird and Ehrenberg, 1970; Barnard and
Ehrenberg, 1990). As well, all the attributes are positively
correlated to circumvent the problem of low response rates
for negative attributes (Barnard and Ehrenberg, 1990).
6. Results
Table 5 provides the distribution of the types of users
(i.e., single users, multiple users, and non-users) for the top
 ARTICLE IN PRESS
S. Bandyopadhyay, M. Martell / Journal of Retailing and Consumer Services 14 (2007) 35–44 41

Table 4 Table 6
List of brand attributes Attribute scoresa of single users and multiple users of major toothpaste
brands

Is suitable for children

Prevents root cavities Brand Single users Multiple users t-value p-value

Prevents cavities/tooth decay

Prevents receding gums Aquafresh Original 0.301 0.309 0.13 0.55

Reverses tooth decay at early stages Colgate Original 0.380 0.351 +0.59 0.28

Prevents red/puffy/tender gums Colgate Tartar 0.433 0.428 +0.09 0.46

Prevents tartar build-up Crest Original 0.430 0.260 +3.92 0.0001

Prevents gum disease/gingivitis Crest Tartar 0.535 0.395 +2.89 0.0028

Cleans teeth well a

Leaves teeth clean An attribute score of 0.301 for single users for Aquafresh Original

Whitens and brightens teeth means that each respondent who have purchased only Aquafresh Original

Freshens breath and no other toothpaste brand in the last year, believes that Aquafresh

Removes film/plaque from teeth Original possesses on average 30.1% of the listed attributes (or 9.03 of the

Removes stains from teeth 30 listed attributes). The same formula has been used to derive the

Keeps teeth looking attractive attribute scores for all brands and for all types of users.

Leaves mouth refreshed

Cleans gums well

Strengthens tooth enamel
Table 7

Leaves mouth feeling clean
Attribute scores of multiple users and non-users of major toothpaste

Has a good taste
brands

Recommended by dentists

Brand I can trust Brand Multiple users Non-users t-value p-value

Is gentle on teeth

Is good for teeth that are sensitive to heat/cold Aquafresh Original 0.309 0.109 +4.60 0.0000

Has an effective fluoride Colgate Original 0.351 0.166 +4.93 0.0000

Has research to prove it’s effective Colgate Tartar 0.428 0.151 +7.29 0.0000

Product/package reduces environmental impact Crest Original 0.260 0.224 +1.18 0.12

Is a good value for the money Crest Tartar 0.395 0.194 +5.80 0.0000

Has a low price

the attribute score for a given brand (e.g., Aquafresh

Table 5 Original) and for a given user group (e.g., single user) is
Distribution of type of users among top 5 toothpaste brands measured as the average number of attributes possessed by
Brand Single users Multiple users Non-users
the brand, in the opinion all respondents from that user
group. Refer footnote in Table 5 for an example of this
Aquafresh Original 36 117 943 measurement procedure.
Colgate Original 43 141 912 Table 7 presents the average attribute ratings for
Colgate Tartar 25 85 986
multiple users and non-users. Interestingly, for all brands
Crest Original 49 217 830
Crest Tartar 105 415 576 other than Crest Original, the average attribute ratings for
multiple users are significantly higher than those for non-
users.
There are several noteworthy features in the aforesaid
5 brands of toothpastes. It is evident that for each brand, results:
the number of multiple users is more than the number of
single users but less than the number of non-users. (1) Single users of the Crest brand (both original and
Tables 6 and 7 provide the average attitude ratings for tartar) demonstrate a much stronger attitude than
five (as against ten) brands; as for the remaining brands, other brands. This means that the brand loyal
the cell sizes did not permit a meaningful statistical consumers of Crest are more strongly opinionated in
analysis. Results are shown as the significance of the favor of their brand than their counterparts for other
difference between the means of attribute ratings. brands. This indicates the existence of double jeopardy
Results in Table 6 show that for Crest Original and Crest in the product category wherein the small brands not
Tartar, average attitude ratings for single users are only suffer from a narrower customer base but also
significantly higher than those of multiple users from less purchase frequency compared to the large
(M ¼ 0:430 vs. 0.260, t ¼ 3:92 for Crest Original, and M ¼ brands.
0:535 vs. 0.395, t ¼ 2:89 for Crest Tartar). For the (2) Differences in attitude strengths of multiple users vs.
remaining three brands—Aquafresh Original, Colgate non-users are stronger than those of single users vs.
Original, and Colgate Tartar—no significant difference multiple users (see Table 7). This is, in fact, expected
between single and multiple users were found. Note that because both single users and multiple users of a brand
 ARTICLE IN PRESS
42 S. Bandyopadhyay, M. Martell / Journal of Retailing and Consumer Services 14 (2007) 35–44

Table 8

Factor Single users Multiple users t-value p-value

Factor scores of single users and multiple users of Crest Original

Disease 0.299 0.191 1.60 0.069
Functional 0.466 0.271 3.04 0.0048
Quality 0.570 0.338 2.53 0.014
Value 0.325 0.215 1.24 0.14
Factor scores of multiple users and non-users of Crest Original
Disease 0.191 0.177 0.27 0.40
Functional 0.271 0.226 1.09 0.15
Quality 0.338 0.302 0.51 0.31
Value 0.215 0.158 0.96 0.19
Factor scores of single users and multiple users of Crest Tartar
Disease 0.496 0.349 1.35 0.10
Functional 0.573 0.435 2.05 0.029
Qaulity 0.587 0.434 1.50 0.079
Value 0.417 0.308 0.78 0.24
Factor scores of multiple users and non-users of Crest Tartar
Disease 0.349 0.179 2.22 0.023
Functional 0.435 0.245 3.75 0.001
Quality 0.434 0.193 3.32 0.003
Value 0.308 0.101 3.37 0.022

should find a lot of common attributes attractive
perhaps due to their prior usage experience. Con-
versely, non-users lack the brand usage experience and
hence may not hold strong beliefs about the brand that
are normally formed due to repeated positive usage
experience. Interestingly, even the non-users of Crest
Original hold such a strong attitude that its difference
with multiple users is statistically insignificant (0.224 vs.
0.260, ns).
Thus, we find moderate support for H1a (single user vs.
multiple user) and strong support for H1b (multiple user
vs. non-user). In particular, H1a is supported for the two
largest brands viz. Crest Original and Crest Tartar. For the
smaller brands, results are not statistically significant but
are generally in the right direction (except Aqua Fresh
Original). Results also validate our conceptual approach to
consider ‘‘non-users’’ with strong attitude as a potentially
attractive consumer segment.
To demonstrate the usefulness of the proposed attitu-
dinal loyalty measurement approach for fast moving
consumer goods marketers we performed a confirmatory
factor analysis to reduce the list of attributes. All 30
attributes loaded onto four factors—functional, disease/
health, quality and value.5 Table 8 shows the mean scores
of these four factors.
Results are given for the two market-leading brands:
Crest Original and Crest Tartar. Interestingly, the results
suggest that functional and quality attributes may be the
most important attributes for customers purchasing tooth-
5
These four factors are identical to the factors used by the data supplier
in their data analysis.
paste with approximately one-half of sole loyal customers
believing their brand (Crest Original or Crest Tartar)
possessed these attributes. This dropped to less than one in
five non-users. These results may suggest that improving
perceptions regarding functional and quality attributes will
assist marketers to improve market share. This is highly
actionable for marketers and demonstrates the usefulness
of this measure for marketers.
7. Managerial implications
Results of our study have substantial managerial
implications. First, our approach to measure both attitu-
dinal loyalty and behavioral loyalty offers more managerial
insights than measuring either type of loyalty indepen-
dently. For example, this approach can demonstrate to the
brand manager that non-users of a brand, particularly
those with a strong positive attitude towards the brand,
may be an important target market. Results given in Table
7 show that non-users of Crest Original and Crest Tartar
score very high in the attribute rating scale (0.224 and
0.194, respectively). In fact, for Crest Original, there
is no statistically significant difference between the
attribute ratings for non-users and multiple users (0.224
vs. 0.260, ns).
Non-users may not buy a given brand for several reasons
such as non-availability of the brand in their market, and/
or lack of awareness of the brand. Brand mangers should
take cognizance of this potential demand and try to devise
strategies to circumvent the problems that keep non-users
from buying the desired brand. For example, they may
offer free samples to the non-users to induce them to switch
brands, and hopefully, retain most of them as single or
 ARTICLE IN PRESS
S. Bandyopadhyay, M. Martell / Journal of Retailing and Consumer Services 14 (2007) 35–44
multiple users. Also, if the product distribution in a given
market is inadequate, the brand manager must re-evaluate
the distribution strategy including relationship with the
retailers, and co-operative merchandising strategies.
Also, attitudinal loyalty measurements provide more
insight about the possible reasons a customer demonstrates
behavioral loyalty. For example, a customer may be driven
by functional attributes, quality attributes or simply by
price. Attitudinal loyalty measures will help brand mangers
understand (1) why and for what reasons, customers
purchase their brands as well as those of their competitors,
and (2) what are the strengths and vulnerabilities of their
brands. These insights will help them plan their marketing
programs more effectively. For example, if a manager finds
that a majority of their customers have a low quality
perception about their brand, they may consider several
remedial strategies such as (1) wrapping the product in
plastic, (2) printing a certificate from an authentic testing
agency, and (3) offering extended product warranty. If the
customers are price-driven, they may offer price discounts
(such as cents-off coupons, shelf price mark-downs etc.)
and volume discounts (such as ‘‘buy one get the second at
half price’’ deals and large-pack discounts).
Our approach to segment customers into single user,
multiple user and non-user categories also helps brand
managers to devise segment-specific marketing strategies.
For example, single users or sole-loyal customers should be
targeted with market penetration strategies (i.e., strategies
that will induce greater use of the product). Multiple users
should be targeted with strategies that highlight the
strengths of your brand vis-à-vis competing brands. For
example, comparative advertisements that portray the
advantages of a brand over its competitors on price and/
or non-price (e.g., functional) attributes will be suitable for
this purpose. As mentioned earlier, non-users should be
targeted with ‘‘free trial’’ offers and product demonstra-
tions to induce first-time trials.
8. Scope of future research and limitations of the study
We have shown how to segment the market on the basis
of not only the behavioral patterns (i.e., single-, multiple-,
and non-users) but also on the basis of the antecedent
attitudes and beliefs that govern this behavior. We have
developed a measure for attitudinal loyalty for low
involvement products. We have also demonstrated that
the attribute–brand pair affirmation is the greatest for
single users followed by multiple users and non-users in
that order.
However, our study has several limitations. In the
survey, behavioral loyalty measures were based on the
respondents’ recall of their purchase history, and hence
may not be completely accurate.6 For example, a respon-
6
We understand from the company supplying the data that the panelists
maintain a record of their purchase history for more than a year. Hence,
we do not expect a lot of inaccuracy in the purchase history data.
43
dent may accurately recall the attributes they recognize
about different brands, but not as accurately the number of
purchases of that brand in a year. However, as stated
earlier, free recall measures used in the survey question-
naire overcome the possible demerits of this shortcoming
(Barnard and Ehrenberg, 1990). Future studies should try
to collect purchase data from the panel of respondents
during or immediately after the purchase using scanners
and then collect attitudinal data from the same panel
through a survey.
The survey used a two-point (yes/no) scale to measure if
respondents felt that a given brand had each of the 30
attributes. Unfortunately, this type of scale cannot
accurately capture the degree of confidence in the response.
In future, a seven-point Likert-type scale should be used to
provide richer data for better understanding of the
attitude–behavior relationship.
In conclusion, our research may be seen as a starting
point in exploring the relationship between behavioral
loyalty and attitudinal loyalty. Further support for our
work across different product categories and markets will
be helpful in validating the bi-dimensional (behavioral and
attitudinal) characteristics of brand loyalty. From a
managerial perspective, our research findings will provide
the brand manager with valuable insight into how
consumer beliefs about brand attributes influence their
brand purchase behavior.
Acknowledgements
We thank Julie Wolfe, Kunal Gupta, and Soumita
Bandyopadhyay for editorial support and background
research. The first author gratefully acknowledges the
financial support of the School of Business and Economics,
and the Center for Regional Excellence of Indiana
University Northwest, and the Research and the University
Graduate School (RUGS) of Indiana University, Bloo-
mington.
References
Alba, J.W., Marmorstein, H., 1987. Frequency information as a
dimension of consumer knowledge. Journal of Consumer Research
14, 14–25.
Allenby, G.M., Rossi, P.E., 1991. Quality perceptions and asymmetric
switching between brands. Marketing Science 10 (3), 185–204.
Amine, A., 1998. Consumers’ true brand loyalty: the central role of
commitment. Journal of Strategic Marketing 6, 305–319.
Baldinger, A.L., Rubinson, J., 1996. Brand loyalty: the link between
attitude and behavior. Journal of Advertising Research 36 (6), 22–34.
Bandyopadhyay, S., Gupta, K., 2004. Comparing double jeopardy effects
at the behavioral and attitudinal Levels. Journal of Product and Brand
Management 13 (3), 180–191.
Bandyopadhyay, S., Gupta, K., Dube, L., 2005. Does brand loyalty
influence double jeopardy? A theoretical and empirical study. Journal
of Product and Brand Management 14 (7), 414–423.
Barnard, N.R., Ehrenberg, A.S.C., 1990. Robust measures of consumer
brand beliefs. Journal of Marketing Research 27, 477–484.
Bennett, R., Rundle-Thiele, S., 2002. Measuring attitudinal loyalty.
Journal of Brand Management 9 (3), 193–209.
 ARTICLE IN PRESS
44 S. Bandyopadhyay, M. Martell / Journal of Retailing and Consumer Services 14 (2007) 35–44
Berne, C., Mugica, J.M., Yague, M.J., 1997. Intrinsic variety-seeking as a
moderator of loyalty. In: Arnott, D., Bridgewater et al. (Eds.), 26th
European Marketing Academic Conference, vol. 3, The University of
Warwick, Warwick Business School, pp. 1452–1462.
Bettman, J., Park, W., 1980. Effects of prior knowledge and phase of the
choices process on consumer decision process: a protocol analysis.
Journal of Consumer Research 7, 234–248.
Bird, M., Ehrenberg, A., 1970. Consumer attitudes and brand usage.
Journal of the Market Research Society 12, 233–247.
Blattberg, R.C., Sen, S.K., 1974. Market segmentation using models of
multidimensional purchasing behavior. Journal of Marketing 38 (4),
20–28.
Bloemer, J., de Ruyter, K., Wetzels, M., 1999. Linking perceived service
quality and service loyalty: a multi-dimensional perspective. European
Journal of Marketing 33 (11/12), 1082–1106.
Chaudhuri, A., Holbrook, M.B., 2001. The chain of effects from brand
trust and brand affect to brand performance: the role of brand loyalty.
Journal of Marketing 65 (2), 81–93.
Coulson, J.S., 1966. Buying decisions within the family and the
consumer–brand relationship. In: Newman, J. (Ed.), On Knowing
the Consumer. Wiley and Sons, New York.
Cunningham, R.M., 1956. Brand loyalty—what where how much.
Harvard Business Review 34, 116–128.
Day, G.S., 1969. A two-dimensional concept of brand loyalty. Journal of
Advertising Research 9, 29–35.
Dall’Olmo Riley, F., Ehrenberg, A.S.C., Castlberry, S.B., Barwise, T.P.,
Barnard, N.R., 1997. The variability of attitudinal repeat rates.
International Journal of Research in Marketing 14, 437–450.
Dick, A.S., Basu, K., 1994. Consumer loyalty: towards an integrated
conceptual approach. Journal of the Academy of Marketing Science 22
(2), 99–113.
East, R., Gendall, P., Hammond, K., Lomax, W., 2005. Customer loyalty:
singular, additive or interactive? Australasian Marketing Journal 13
(2), 10–26.
Ehrenberg, A.S.C., 2000. Repeat buying—facts, theory and applications.
Journal of Empirical Generalizations in Marketing Science 5,
392–770.
Ehrenberg, A.S.C., Goodhardt, G.J., Barwise, P.B., 1990. Double
jeopardy revisited. Journal of Marketing 54, 82–91.
Fishbein, M., Ajzen, I., 1975. Belief Attitude Intention and Behavior: An
Introduction to Theory and Research. Addison-Wesley, Reading, MA.
Fournier, S., 1998. Consumers and their Brands: developing relationship
theory in consumer research. Journal of Consumer Research 24 (4),
343–373.
Ganesh, J., Arnold, M.J., Reynolds, K.E., 2000. Understanding the
customer base of service providers: an examination of the differences
between switchers and stayers. Journal of Marketing 64, 65–87.
Hasher, L., Zacks, R., 1979. Automatic and effortful processes in memory.
Journal of Experimental Psychology: General 108, 356–388.
Hasher, L., Zacks, R., 1984. Automatic processing of fundamental
information. American Psychologist 39, 1372–1388.
Jacoby, J., 1971. Brand loyalty: a conceptual definition. Proceedings of the
79th American Psychological Association Convention, 655–656.
Jacoby, J., Kyner, D.B., 1973. Brand loyalty versus repeat purchasing
behavior. Journal of Marketing Research February, 1–9.
Kahn, B.E., Kalwani, M.U., Morrison, D.G., 1986. Measuring variety
seeking and reinforcement behaviors using panel data. Journal of
Marketing Research 23, 89–100.
McConell, J.D., 1968. The development of brand loyalty: an experimental
study. Journal of Marketing Research 5, 13–19.
Mellens, M.M., Dekimpe, G., Steenkamp, J.B.E., 1996. A review of
brand-loyalty measures in marketing. Tijdschrift voor Economic en
Management 41, 507–533.
Mittal, B., Lasser, W.M., 1998. Why do customers switch? The dynamics
of satisfaction versus loyalty. The Journal of Services Marketing 12 (3),
177–194.
Newman, J., 1966. On Knowing the Consumer. Wiley, New York.
Oliver, R.L., 1997. Satisfaction: a behavioral perspective on the consumer.
Irwin/McGraw-Hill, New York.
Oliver, R.L., 1999. Whence consumer loyalty? Journal of Marketing 63
(Special Issue), 33–44.
Petty, R., Cacioppo, J., 1984. The effects of involvement on responses to
argument quantity and quality: central and peripheral routes to
persuasion. Journal of Personality and Social Psychology 46, 69–81.
Raju, P., 1980. Optimum stimulation level, its relationship to personality,
demographics, and exploratory behavior. Journal of Consumer
Research 7, 272–282.
Rundle-Thiele, S.R., 2005. Elaborating customer loyalty: exploring loyalty
to wine retailers. Journal of Retailing and Consumer Services 12 (5),
333–344.
Rundle-Thiele, S.R., Bennett, R., 2001. A brand for all seasons: A
discussion of loyalty approaches and their applicability for different
markets. Journal of Product and Brand Management 10 (1), 25–37.
Russo, J., Dosher, B., 1983. Strategies for multiattribute binary choice.
Journal of Experimental Psychology: Learning, Memory, and Cogni-
tion 9, 676–696.
Sharp, A., Sharp, B., Wright, M., 2002. Questioning the value of the
‘‘True’’ brand loyalty distinction. International Journal of Research in
Marketing 19 (1), 81–90.
Sheth, J.N., Mittal, B., Newman, B.I., 1999. Consumer Behavior &
Beyond. Dryden Press, Sydney.
Uncles, M.D., Hammond, K., Ehrenberg, A.S.C., Davis, R.E., 1994. A
replication study of two brand-loyalty measures. European Journal of
Operational Research 76 (2), 375–384.
Uncles, M.D., Dowling, G.R., Hammond, K., 2003. Customer loyalty and
customer loyalty programs. Journal of Consumer Marketing 20 (4),
294–314.
Wansink, B., Sonka, S., Park, S., 2001. Category-based segmentation
methods: category seekers vs. heavy users. Food and Brand Lab
Working Paper, University of Illinois.
Zeithaml, V.A., Berry, L.L., Parasuraman, A., 1996. The behavioral
consequences of services quality. Journal of Marketing 60 (2), 31–46.