ArticlePDF Available

Abstract and Figures

This study investigates consumers’ usage of online recommendation sources and their influence on online product choices. A 3 (websites) × 4 (recommendation sources) × 2 (products) online experiment was conducted with 487 subjects. Results indicate that subjects who consulted product recommendations selected recommended products twice as often as subjects who did not consult recommendations. The online recommendation source labeled “recommender system,” typical of the personalization possibilities offered by online retailing, was more influential than more traditional recommendation sources such as “human experts” and “other consumers”. The type of product also had a significant influence on the propensity to follow product recommendations. Theoretical and managerial implications of these findings are provided.
Content may be subject to copyright.
Journal of Retailing 80 (2004) 159–169
The influence of online product recommendations on
consumers’ online choices
Sylvain Senecala,, Jacques Nantela,1
aHEC Montreal, University of Montreal, 3000 Chemin de la Cote-Sainte-Catherine, Montreal, Que., Canada H3T 2A7
This study investigates consumers’ usage of online recommendation sources and their influence on online product choices. A 3 (websites) ×
4 (recommendation sources) ×2 (products) online experiment was conducted with 487 subjects. Results indicate that subjects who consulted
product recommendations selected recommended products twice as often as subjects who did not consult recommendations. The online
recommendation source labeled “recommender system,” typical of the personalization possibilities offered by online retailing, was more
influential than more traditional recommendation sources such as “human experts” and “other consumers”. The type of product also had a
significant influence on the propensity to follow product recommendations. Theoretical and managerial implications of these findings are
© 2004 by New York University. Published by Elsevier. All rights reserved.
Keywords: Online product; Recommendation; Consumers
Among all possible advantages offered by electronic com-
merce to retailers, the capacity to offer consumers a flexible
and personalized relationship is probably one of the most
important (Wind & Rangaswamy, 2001). Online personal-
ization offers retailers two major benefits. It allows them
to provide accurate and timely information to customers
which, in turn, often generates additional sales (Postma &
Brokke, 2002). Personalization has also been shown to in-
crease the level of loyalty consumers hold toward a retailer
(Cyber Dialogue, 2001;Srinivasan, Anderson, & Ponnavolu,
2002). While there are several ways to personalize an online
relationship, the capacity for an online retailer to make rec-
ommendations is certainly among the most promising (The
e-tailing Group, 2003). Online, recommendation sources
range from traditional sources such as other consumers
(e.g., testimonies of customers on retail websites such as to personalized recommendations provided
by recommender systems (West et al., 1999). To date, no
study has specifically investigated and compared the rela-
tive influence of these online recommendation sources on
Corresponding author. Tel.: +1 419 530 2422.
E-mail addresses: (S. Senecal), (J. Nantel).
1Tel.: +1 514 340 6421.
consumers’ product choices. Therefore, the main objective
of this study is to investigate the influence of online product
recommendations on consumers’ online product choices.
In addition, we explore the moderating influence of vari-
ables related to recommendation sources and the purchase
Literature review
Research on the use and influence of recommendations
on consumers has typically been subsumed under personal
influence or word-of-mouth (WOM) research. In addition, as
noted by Rosen and Olshavsky (1987), research on opinion
leadership and reference groups also relates to the study of
recommendations and to influence in general.
Recommendation sources are considered primarily as
information sources. Andreasen (1968) proposes the fol-
lowing typology of information sources: (1) Impersonal
Advocate (e.g., mass media), (2) Impersonal Independent
(e.g., Consumer Reports), (3) Personal Advocate (e.g.,
sales clerks), and (4) Personal Independent (e.g., friends).
Although research on personal influence and WOM fo-
cuses on the latter two information sources, it is notewor-
thy that impersonal independent information sources such
as Consumer Reports can also serve as recommendation
sources. Moreover, the Internet can provide consumers with
0022-4359/$ – see front matter © 2004 by New York University. Published by Elsevier. All rights reserved.
160 S. Senecal, J. Nantel/Journal of Retailing 80 (2004) 159–169
an additional type of impersonal information source. For
instance, electronic decision-making aids such as recom-
mender systems are impersonal information sources that
provide personalized information to consumers (Ansari,
Essegaier, & Kohli, 2000). In an effort to extend Andreasen’s
(1968) typology to computer-mediated environments, we
assert that information sources can be sorted into one of
four groups: (1) Personal source providing personalized
information (e.g., “My sister says that this product is best
for me.”); (2) Personal source providing non-personalized
information (e.g., “A renowned expert says that this product
is the best.”); (3) Impersonal source providing personalized
information (e.g., “Based on my profile, the recommender
system suggests this product.”); (4) Impersonal source pro-
viding non-personalized information (e.g., “According to
Consumer Reports, this is the best product on the market.”).
In consumer research, studies on personal influence, so-
cial influence, or WOM, can be categorized as studies
investigating personal sources providing personalized or
non-personalized information. Furthermore, studies dealing
with reference groups encompass such sources as well as
impersonal sources that provide non-personalized informa-
tion. Thus, a new area has emerged in consumer research,
arising mainly from information technologies such as the
Internet: that of impersonal sources that provide person-
alized information (Alba et al., 1997; Ansari et al., 2000;
Häubl & Trifts, 2000;Maes, 1999;Urban, Sultan, & Qualls,
1999;West et al., 1999).
Research on information sources suggests that personal
and impersonal information sources influence consumers’
decision-making (Ardnt, 1967; Duhan et al., 1997; Gilly
et al., 1998;Olshavsky & Granbois, 1979;Price & Feick,
1984). For instance, Price and Feick (1984) found that con-
sumers planned to use the following information sources
for their next durable good purchase: (1) Friends, relatives,
and acquaintances, (2) Salespeople, (3) Publications such
as Consumer Reports. However, if much is known about
the relative likelihood of consumers to consider recommen-
dations in the course of their decision making process, lit-
tle is known about how recommendations, especially in a
computer-mediated environment, impact consumers’ prod-
uct choices.
Determinants of recommendation influence
The current study focuses on three determinants that could
influence the impact of computer-mediated recommenda-
tions on consumers’ online product choices: the nature of the
product recommended, the nature of the website on which
the recommendation is proposed, and the type of recommen-
dation source.
Prior research has shown that the type of product affects
consumers’ use of personal information sources and their
influence on consumers’ choices (Bearden & Etzel, 1982;
Childers & Rao, 1992;King & Balasubramanian, 1994).
Nelson (1970) suggests that goods can be classified as pos-
sessing either search or experience qualities. Search quali-
ties are those that “the consumer can determine by inspec-
tion prior to purchase,” and experience qualities are those
that “are not determined prior to purchase” (Nelson, 1974,
p. 730). Since it is difficult or even impossible to evalu-
ate experience products before purchase, consumers should
rely more on product recommendations for these products
than for search products. In support of this view, King and
Balasubramanian (1994) found that consumers assessing a
search product (e.g., a 35-mm camera) are more likely to
use own-based decision-making processes than consumers
assessing an experience product, and that consumers evalu-
ating an experience product (e.g., a film-processing service)
rely more on other-based and hybrid decision-making pro-
cesses than consumers assessing a search product.
The nature of the website can also influence the im-
pact of a given recommendation. Based on previous web-
site classifications (Hoffman, Novak, & Chatterjee, 1995;
Spiller & Lohse, 1998), Senecal and Nantel (2002) sug-
gest that recommendation sources can be used and promoted
by three different types of websites: sellers (e.g., retailer
or manufacturer websites such as, commer-
cially linked third parties (e.g., comparison shopping web-
sites such as, and non-commercially linked
third parties (e.g., product or merchant assessment websites
such as More independent websites
such as non-commercially linked third parties that facilitate
consumers’ external search effort by decreasing search costs
are assumed to be preferred by consumers (Alba et al., 1997;
Bakos, 1997;Lynch & Ariely, 2000). By providing more al-
ternatives to choose from and more objective information,
independent websites should be perceived as more useful by
consumers. In addition, prior research on attribution theory
suggests that consumers discredit recommendations from
endorsers if they suspect that the latter have incentives to
recommend a product (for reviews, refer to Folkes, 1988;
Mizerski, Golden, & Kernan, 1979). According to the dis-
counting principle of the attribution theory (Kelley, 1973),
which suggests that a communicator will be perceived as
biased if the recipient can infer that the message can be at-
tributed to personal or situational causes, consumers would
attribute more non-product related motivations (e.g., com-
missions on sales) to recommendation sources that are pro-
moted by commercially linked third parties and sellers than
independent third party websites. Consequently consumers
would follow product recommendations in a greater propor-
tion when shopping on more independent than on less inde-
pendent websites.
In light of research on consumers’ use of relevant others
in their pre-purchase external search efforts (Olshavsky &
Granbois, 1979;Price & Feick, 1984;Rosen & Olshavsky,
1987) and in consideration of the emergence of online in-
formation sources providing personalized recommendations
(Ansari et al., 2000), Senecal and Nantel (2002) assert that
online recommendation sources can be sorted into three
broad categories: (1) other consumers (e.g., relatives, friends
S. Senecal, J. Nantel/Journal of Retailing 80 (2004) 159–169 161
and acquaintances), (2) human experts (e.g., salespersons,
independent experts), and (3) expert systems such as recom-
mender systems. We posit that these online recommendation
sources will have different levels of influence on consumers’
online product selection. Brown and Reingen (1987) sug-
gest that information received from sources that have some
personal knowledge about the consumer have more influ-
ence on the latter than sources that have no personal knowl-
edge about the consumer. Thus, a recommendation source
providing personalized information to consumers (e.g., rec-
ommender system) should be more influential than a recom-
mendation source providing non-personalized information
(e.g., other consumers).
The fact that both factors, the origin (source) of a recom-
mendation as well as the type of website on which it is made,
have an impact on the likelihood it has to be followed may
find its explanation in Kelman’s (1961) work on source cred-
ibility. Kelman (1961) suggests that credibility is a product
of expertise and trustworthiness. Expertise can be viewed
as the perceived ability of an information source to know
the right answer and trustworthiness as the perceived infor-
mation source’s motivation to communicate this expertise
without bias (McGuire, 1969). Although moderated by con-
textual factors (for a review, refer to Sternthal, Phillips, &
Dholakia, 1978), source expertise and trustworthiness have
been found to be positively correlated with consumers’ atti-
tude toward the brand, and consumers’ behavioral intentions
and behaviors (Gilly et al., 1998;Harmon & Coney, 1982;
Lascu, Bearden, & Rose, 1995;Tybout, 1978).
Based on the preceding review of the literature we pos-
tulate that personal information sources as well as imper-
sonal information sources providing product recommenda-
tions (Price & Feick, 1984) will influence consumers in
computer-mediated environments such as the Internet and
the World Wide Web. We thus formulate the following gen-
eral hypothesis.
H1. Consumers who consult an online information source
recommending a given brand will select that brand in a
greater proportion than consumers who do not consult an
online recommendation source.
As for the impact that such a recommendation will have on
consumers’ choice, we formulate three additional hypothe-
ses. First, we posit that the nature of the product for which
a recommendation is provided will influence the likelihood
that it will be followed. Based on prior research on the re-
lationship between product type and personal information
source influence (Bearden & Etzel, 1982;Childers & Rao,
1992;King & Balasubramanian, 1994), we put forward the
following hypothesis.
H2. Online recommendations for experience products will
be followed in a greater proportion than online recommen-
dations for search products.
Second, based on Alba et al. (1997),Bakos (1997) and
Lynch and Ariely (2000), we propose that online product
recommendations from more independent websites are more
influential than those from less independent websites. We
therefore put forth the following hypothesis.
H3. Online product recommendations consulted on
“non-commercially linked third party” websites will be
followed in a greater proportion by consumer than if con-
sulted on “commercially linked third party” websites, and
online product recommendation consulted on the latter type
of websites will be followed in a greater proportion than if
consulted on “seller” websites.
Finally, we believe, based on the literature which has dealt
with the issue of consumers’ use of relevant others in their
pre-purchase external search efforts, that personalized rec-
ommendations will have a greater influence on consumers
than non-personalized ones (Brown & Reingen, 1987). Thus
follows hypotheses four.
H4. Recommendations from information sources offering
personalized recommendations (e.g., recommender system)
will be followed in a greater proportion by consumers
than recommendations from information sources providing
non-personalized recommendations.
In addition to this set of hypotheses, which pertains
to the variables that moderate the influence of an online
recommendation, we formulate a set of three hypotheses
which consider potential reasons for which various online
recommendation sources may differ in their influence on
consumers’ choices. First, we expect that the recommen-
dation source “other consumers” will be perceived as less
expert than “human experts” and “recommender systems”.
However, based on the discounting principle of attribution
theory (Kelley, 1967), the recommendation source “other
consumers” should be perceived as more trustworthy than
human experts and recommender systems since the lat-
ter two recommendation sources are more susceptible to
non-product related attributions. Second, since consumers
may also attribute non-product related motivations more
easily to recommendation sources promoted by websites
that are not clearly independent, we predict that the type of
website will have an impact on the perception of the recom-
mendation source’s trustworthiness. For instance, a human
expert who recommends a product on a seller website may
be perceived by consumers as less trustworthy than if that
person recommended the same product on an independent
third party website. Thus, the following hypotheses are
162 S. Senecal, J. Nantel/Journal of Retailing 80 (2004) 159–169
H5a. The online recommendation sources “human experts”
and “recommender system” will be perceived as possess-
ing more expertise than the online recommendation source
“other consumers.”
H5b. The online recommendation sources “human experts”
and “recommender system” will be perceived as less trust-
worthy than the online recommendation source “other
H6. Consumers’ trust in the recommendation source will be
greater when shopping on a “non-commercially linked third
party” type of website than on a “commercially linked third
party” type of website, which in turn will be perceived as
more trustworthy than the “seller” type of website.
A convenience sample of 487 subjects was generated
on the basis of 25742 e-mails sent to three populations of
web users. The initial e-mail stated that two researchers
from a large business school were conducting a study on
electronic commerce, and that participants had a chance
of winning one of the products about which the experi-
ment was designed. Potential participants did not know
in advance the types of products that were to be tested.
They were told that they would be asked to participate in
two sessions in order to complete the study. Overall, 630
subjects participated in the first session and 487 subjects
participated in both sessions. Subjects who participated
in both sessions did not significantly differ from subjects
who only participated in the first session with regards
to their socio-demographic profile (X2
agegroup(4)=2.138, X2
income(7)=11.683; all p-values > .05), their product class
familiarity (Fwine(1,629)=0.316, Fcalculator(1,630)=
0.951; all p-values > .05), and their product class subjective
knowledge (Fwine(1,629)=0.758, Fcalculator(1,630)=
0.000; all p-values > .05). Subjects took an average of 6–8
days to complete both sessions of the online experiment,
which included a 5-day delay between email invitations of
sessions 1 and 2. Of the 487 participants, 173 were from
a specialized firm list (response rate: 0.6%), 59 originated
from an e-commerce research center list (response rate:
12.0%), 203 were from an undergraduate student list (re-
sponse rate: 7.7%) and 52 could not be traced since they
used a different email address than the ones on the lists.
Subjects participated in both sessions of the study from the
location where they usually use the Internet. The major-
ity of subjects were between the ages of 18 and 29 years
(84%). Fifty percent were female, one third were working
full time (31%); 26% of subjects were full-time students
and another 31% were part-time workers and students. On
average, subjects had been using the Internet for 4.5 years
and currently used it 18 hours per week.
Experiment overview
In the first session of the experiment, subjects were simply
asked to complete an online questionnaire. In the second ses-
sion, subjects were asked to perform online shopping tasks
on a specific website. During that second session a 3×4×2
online experiment was conducted. The first between-subject
factor was the website manipulation. Subjects were assigned
to one of three types of websites: retailer, third party com-
mercially linked to retailers, or non-commercially linked
third party. The second between-subject factor manipulated
the source of recommendation. Subjects were assigned to
one of the four following conditions: other consumers, hu-
man experts, recommender system, or no recommendation
source. Finally, the last factor, a within-subject factor, was
the product manipulation. During their first online shopping
task, subjects were randomly assigned to either a search or
experience product, and they were assigned to the remaining
product type for their second shopping task.
Experiment description
To motivate subjects to participate without mentioning the
precise goal of the experiment (i.e., the influence of recom-
mendation on product choice), a cover story was used. Sub-
jects were told that a two-session study was being conducted
to assess the commercial potential of various products that a
foreign company (i.e., Maximo) was interested in introduc-
ing to local markets via their website. In addition, partici-
pants were informed that they would be asked in the second
session of the study to select three products, and that they
had a one in three chance of winning one of the products
selected. This procedure was used to maximize the involve-
ment of subjects with their online shopping tasks. Subjects
were informed that the average product value was $45. The
first session questionnaire measured their knowledge and
familiarity with the computer mouse, calculator and wine
product classes,2their Internet usage, and some demograph-
ics. At the end of the questionnaire, subjects were asked to
provide their email address and were told that they would
be contacted in the following days for the second session.
Five days after the first session, subjects were sent an
email providing a hyperlink to the second session website.
Once on the website, they were asked to logon by entering
their email address. Following a brief introduction to the ex-
perimental website to remind them of the goal of the study
(i.e., cover story), they were randomly assigned to one of
three Maximo websites. Once on the website, subjects were
2In the first session, subjects were asked to complete a subjective
knowledge measurement scale (Flynn & Goldsmith, 1999) and a famil-
iarity measurement scale (Park, Mothersbaugh, & Feick, 1994) for each
product category.
S. Senecal, J. Nantel/Journal of Retailing 80 (2004) 159–169 163
Table 1
Brands used in the experiment
Computer Mice Calculators Red Wine
Kensington Mouse In A Box Optical Pro Canon P23-DH Ceuso Custera 1998
Targus Optical Stroller Mini Mouse Casio HR-8LaLes Longeroies 1998
Microsoft Intellimouse Optical Texas Instruments TI5019 Callabriga 1995a
Logitech Wheel Mouse OpticalaCasio HR100LC Coteaux du Languedoc, Les Hauts de Lunes 1996
aProduct recommended by all recommendation sources.
instructed to read a description of the company to clearly
understand which of the three types of websites they were
visiting. During the experiment, the information and its pre-
sentation (e.g., website layout) were held constant across the
different treatments levels. Thus, all three Maximo websites
were graphically identical. Subjects were then advised that
within the next few minutes they would be asked to shop on
Maximo’s website and select three products from three dif-
ferent product classes. Although Maximo was presented as
a real European company with a professional looking web-
site it was in fact a fictitious company. However, all products
used in the experiment were actual brands available online
(see Table 1).
As recommended by Nosek, Banaji, and Greenwald
(2002), the first online shopping task was a warm-up task.
Its goal was to familiarize subjects with the structure and
functionalities of Maximo’s website. Subjects were shown
four computer mice and asked to choose one. They were
able to evaluate mice based on their attributes and they
were also randomly assigned to one of the four recommen-
dation source treatment levels. Hence, most subjects had
the opportunity to consult a recommendation page. Sub-
jects were free to consult or not the recommendation page
(i.e., Click or not on the “Our recommendation” button).
Remaining subjects were assigned to the control group
condition, i.e., they were randomly assigned to one website
treatment level and to the “no recommendation” condition
of the recommendation source factor. On the recommenda-
tion page, the recommendation source (e.g., human experts)
was described to the subject and it recommended one of
four products within the product class. Note that the same
product was recommended by all recommendation sources
(see Table 1). After this initial product assessment, subjects
were asked to choose one of the four mice presented.
The warm-up task was followed by a second online shop-
ping task. Subjects were randomly assigned to a product
class (i.e., calculator or wine3). Product classes were ran-
domized to control for any order effect. The second shop-
ping task essentially followed the same procedure as the
warm-up shopping task. Subjects were assigned to the same
recommendation source treatment level and were asked to
3The data collection was performed in Canada where the legal age for
drinking is 18 years old. In addition, no significant relationship was found
between subjects’ professional situation (e.g., part time worker, full time
worker, full time students, etc.) and their subjective knowledge of wine
(F(6,473)=1.322; p> .05).
select one product out of four within the product class (see
Table 1). The product recommended by all recommendation
sources was again the same. Following the second product
choice, subjects not assigned to the control group condi-
tion and who had consulted the recommendation page (i.e.,
subjects who clicked on the “Our recommendation” button)
were asked to complete a recommendation source credibil-
ity measurement scale.
Following this second shopping task, subjects were asked
to perform the third and final shopping task. As part of this
task, subjects were exposed to four products of the remaining
product class (i.e., calculator or wine). The third shopping
task essentially followed the same procedure as the second
shopping task. Subjects were assigned to the same recom-
mendation source as that of previous shopping tasks. Follow-
ing their final product selection, subjects who consulted the
recommendation page were again asked to evaluate the rec-
ommendation source’s credibility. After having completed
all three shopping tasks, subjects were asked to complete
a short final questionnaire in which they were prompted to
guess the main objective of the experiment.4They then ac-
cessed a debriefing page explaining the actual goal of the
experiment and were logged out of the second session. The
debriefing page explained the real goal of the experiment
(i.e., influence of recommendations on product choices), re-
assured subjects about their chance to win one of the prod-
ucts they selected, indicated that the collected data would
remain confidential, and that all researchers performing the
study had signed a confidentiality agreement. Finally, sub-
jects were provided the University Ethics Comittee phone
number in order for them to call if they had any questions
or comments on the study.
The website treatment
Subjects were assigned to one of the three following web-
site treatment levels presented in Table 2. We purposely used
a fictitious company in order to control for any past experi-
The recommendation source treatment
Four treatment levels were used for the recommenda-
tion source manipulation: other consumers, human experts,
recommender system, and no recommendation. During the
4The following open-ended question was asked to subjects at the end
of the second session: “To your knowledge, what is the main goal of this
164 S. Senecal, J. Nantel/Journal of Retailing 80 (2004) 159–169
Table 2
The website treatment levels
Treatment Level Description Provided to Subjects
Retailer “Maximo is a large European store selling products on the Internet. It is currently assessing the
feasibility of offering new products on its website to consumers in your area. Therefore, it is very
interested in learning your product preferences. In the region, Maximo competes with The Bay,
Staples and Wal-Mart, which also offer their products on the Internet.”
3rd party commercially linked to retailers “Maximo is a large European buying group. It is currently assessing the feasibility of offering
new products on its website to consumers in your area. Therefore, it is very interested in learning
your product preferences. Being an intermediary between consumers and a limited number of
partner-retailers offering their products on the Internet, Maximo offers the best products available
at its partner-retailers. In the region, Maximo has the following partners: The Bay, Staples and
Non-commercially linked 3rd party “Maximo is a large European independent organization offering a product comparison service on
the Internet. It is currently assessing the feasibility of offering new products on its website to
consumers in your area. Therefore, it is very interested in learning your product preferences.
Being independent, Maximo selects for you the best products available on all sites offering
products on the Internet. Hence, Maximo offers a service similar to that of Consumer Reports.”
experiment, if subjects were assigned to a recommendation
source, and if they elected to see the product recommen-
dation (i.e., clicked on the “Our recommendation” button),
they were exposed to a recommended product and a descrip-
tion of the recommendation source. Based on pretest results
of consumers’ preferences, the second best preferred prod-
uct was always proposed by recommendation sources. For
the “other consumers” treatment level, the recommendation
source for the mouse product class was described as follows.
“This recommendation is based on other consumer se-
lections. In fact, based on the choices of consumers in your
area, we have determined the following preferences:
Product Consumers who have
selected the product (%)
Kensington’s Mouse In A
Box Optical Pro 9
Targus’ Optical Scroller
Mini Mouse 2
Microsoft’s Intellimouse
Optical 18
Logitech’s Wheel Mouse
Optical 71
When subjects were assigned to the “human experts” con-
dition, the recommendation source was presented as follows:
“This recommendation is based on an evaluation by our
team of experts. Our advisors, experts in this product class,
highly recommend this product over the others.” The “rec-
ommender system” treatment level was described as follows:
“This recommendation results from the analysis of the an-
swers to the questionnaire that you completed a few days ago
during the first phase of the study. Our computer system ana-
lyzed your answers and, based on your personalized profile,
the system highly recommends this product over the others.”
Thus, subjects were led to believe that the recommendation
from the recommender system was personalized based on
their answers to the questions of the first session. Subjects
assigned to the “no recommendation” treatment level did not
have the opportunity to consult a recommendation source
during their sopping tasks, i.e., no “Our recommendation”
button was present on the Maximo website they visited.
Product type treatment
The product type was manipulated by using two different
product classes. Based on pretest results, the search product
class used for the experiment was the calculator, and wine
was used for the experience product class. Since it is the
only within factor of the experiment, after their warm-up
task, subjects were randomly assigned to either the calculator
or the wine product class on their first shopping task and
assigned to the other product class for their second shopping
The first dependent variable was the influence of the rec-
ommendation source on consumers’ online product choices.
The influence was measured by a dichotomous variable.
Each product choice was categorized as either a decision to
follow or not to follow the product recommendation. The
remaining dependent variables were related to the credi-
bility of recommendation sources. The measurement scale
developed by Ohanian (1990) was used to assess recom-
mendation sources’ expertise and trustworthiness. Results
from a pretest (n=39) and from the experiment (n=
487) both show that the measurement scale is reliable. The
Cronbach’s alphas for the expertise dimension ranged from
0.88 to 0.91 and from 0.84 to 0.88 for the trustworthiness
Manipulation checks
Following Perdue and Summers (1986), all manipulation
checks were performed during pretests. A series of four
pretests were necessary to achieve effective online manipula-
tions. After each pretest, necessary iterations were made on
S. Senecal, J. Nantel/Journal of Retailing 80 (2004) 159–169 165
manipulations (e.g., Website descriptions) to ensure their ef-
fectiveness. Results from the last pretest are presented below.
For this pretest, a non-student sample of 33 consumers was
used. The pretest followed the same procedure as the main
experiment (i.e., online questionnaires and 3×4×2 mixed
design online experiment) except that it was conducted in
only one session. Pretest subjects were not included in the
final sample.
Product manipulation
First, subjects were asked to evaluate the nature of a set
of product classes (calculator, camping cooler, computer
mouse, water filter system, bottle of wine, and 35mm cam-
era). For each product class, subjects were asked whether
products could either be evaluated: (1) before purchase; (2)
mostly before purchase; (3) mostly after purchase; (4) only
after purchase. Results of the pretest indicated that the wine
product class was perceived as the most “experience” prod-
uct (mean =3.2, median =3) and the calculator product
class was perceived as the most “search” product (mean =
1.4, median =1). Furthermore, the difference between
the evaluations of the two product classes was significant
(t(27)=−7.48, p<.001). The computer mouse was the
most balanced product category (mean =2.1, median =2).
Website manipulation
Following their first online shopping task, subjects in-
volved in the pretest, were asked to identify the type of
website on which they were shopping. They had to select
one of three different descriptions representing either a re-
tailer, a third party commercially linked to retailers, or a
non-commercially linked third party. Subjects assigned to
one type of website (e.g., retailer) mentioned more often that
they were shopping on that specific type of website than sub-
jects assigned to the other types of websites (X2
4.54, p<.05; X2
Dependent 3rd party(1)=8.42, p<.005;
Independent 3rd party(1)=12.44, p<.001; n=33). Thus,
pretest results suggested that the website manipulation was
Recommendation source manipulation
After their second shopping task, pretest subjects who had
consulted a product recommendation were asked to identify
which type of recommendation source they had consulted.
They had to select one of three different descriptions repre-
senting a recommender system, human experts, or a group of
other consumers. Subjects assigned to one recommendation
source (e.g., other consumers) more often mentioned that
they received a recommendation from that specific source
than subjects assigned to the other recommendation sources
Other consumers(1)=7.47, p<.01; X2
6.86, p<.01; X2
Recommender system(1)=5.57, p<.05, n=
16). Again, pretest results suggested that the recommenda-
tion source manipulation was effective.
Hypothesis guessing
Of the 487 participants, 20.3% correctly guessed the goal
of the experiment. However, the frequency of choosing a
recommended product did not differ between subjects who
correctly guessed the goal of the experiment and subjects
who did not correctly guess the main goal of the experiment
(Wine: X2(1)=3.640, p> .05; Calculator: X2(1)=2.703,
p> .05). Furthermore, no significant difference was found
between the two groups regarding the perceived recommen-
dation source credibility (Wine: F(1,264)=0.056, Calcu-
lator: F(1,261)=0.03; all p-values > .05). Note that sub-
jects who consulted a product recommendation had a better
chance of guessing the experiment’s goal since they were
asked to complete a source credibility measurement scale
after their shopping tasks. Thus, observations from all par-
ticipants were used to test the hypotheses.
Test of hypotheses
H1 to H4 all suggest relationships between a categorical
independent variable (i.e., exposition to a recommendation,
type of product, type of website, or recommendation source)
and a dichotomous dependent variable (i.e., selection or non
selection of a recommended product). In addition, subjects’
responses were likely to be correlated since they had to per-
form two product choices (i.e., correlation between the de-
cision to follow the wine product recommendation and the
calculator product recommendation). Because of the nature
of the data and because of the repeated treatment used in
this study, a Generalized Estimating Equations (GEE) pro-
cedure was used in order to test H1 to H4. Introduced by
Liang and Zeger (1986), the GEE approach is an extension
of generalized linear models designed to handle categorical
repeated measurements arising from within-subject designs
(for more details on the GEE methodology, see Liang, Zeger,
& Qaqish, 1992;Stokes, Davis, & Koch, 2001;Zeger, Liang,
& Albert, 1988). GEE analyses were thus performed using
the GENMOD procedure of the SAS system version 8.0. Fi-
nally, since H5 and H6 dealt with categorical independent
variables (i.e., type of recommendation source and type of
website) and dependent variables that were continuous and
measured twice during the experiment (i.e., perceived trust
and expertise), a MANCOVA for repeated measures was
The two online shopping tasks performed by 487 sub-
jects generated 974 observations. Of the 974 observations,
200 observations were collected from the control group
that was not assigned to any recommendation sources and
412 observations reflect decisions to consult a product
166 S. Senecal, J. Nantel/Journal of Retailing 80 (2004) 159–169
Table 3
Total number of product choices performed by subjects who consulted a product recommendation and percentages of recommended product selected
Recommendation source/
website manipulations Other consumers Human experts Recommender system Total
Retailer 33 54 60 147
Ca: 20% C: 35% C: 43% C: 35%
Wa: 39% W: 52% W: 77% W: 59%
Ab: 30% A: 43% A: 60% A: 47%
Dependent 3rd party 43 45 42 130
C: 30% C: 36% C: 42% C: 36%
W: 40% W: 65% W: 65% W: 57%
A: 35% A: 49% A: 55% A: 46%
Independent 3rd party 51 47 37 135
C: 36% C: 32% C: 31% C: 33%
W: 55% W: 52% W: 57% W: 55%
A: 43% A: 43% A: 46% A: 44%
Total 127 146 139 412
C: 30% C: 34% C: 40% C: 35%
W: 45% W: 56% W: 68% W: 57%
A: 37% A: 45% A: 55% A: 46%
aC (W): Proportion of subjects who selected the recommended calculator (wine).
bA: Weighted average of C and W.
recommendation. Of the latter number of observations,
188 observations represent decisions to follow a product
Influence of recommendation on product choice
H1 stipulated that consumers who consult a product rec-
ommendation were more likely to select the recommended
product than consumers who do not consult a recommenda-
tion source. A first GEE analysis using observations from
all subjects (n=974) was performed to test this hypothe-
sis. One dichotomous independent variable (exposition to a
product recommendation) and one dichotomous dependent
variable (selection of the recommended product) were used
in the GEE analysis. Results strongly support H1 (X2(1)=
52.3, p<.001). Overall only 22.5% of product choices
made by subjects who did not see a product recommenda-
tion, either because they were assigned to the control group
or because they did not click on the recommendation button
during their shopping task, favored the recommended prod-
uct compared to a proportion of 45.6% of product choices
made by subjects who consulted a product recommendation.
Thus, online product recommendations greatly influenced
subjects’ product choices.
In order to test H2 to H4, a second GEE analysis using
only observations from subjects who had consulted a prod-
uct recommendation (n=412) was performed to test the
relationships between the type of product (H2), the type of
website (H3), or the type of recommendation source (H4)
and product choices (Table 3). Thus, three categorical inde-
pendent variables (type of product, type of website, and type
of recommendation source) and one dichotomous depen-
dent variable (selection of the recommended product) were
used in the GEE analysis. In addition, the following vari-
ables were used as covariates in the analysis: professional
status (i.e., student/non-student), order of exposure to prod-
ucts, product class familiarity, and product class subjective
As illustrated in Table 4, results of the GEE analysis
showed no significant two-way or three-way interactions, but
they revealed main effects for the product type and the rec-
ommendation source manipulation. The analysis of the co-
variates revealed that subjects’ professional status (X2(1)=
0.15, p> .05), subjective knowledge of the product class
((X2(1)=0.05, p> .05), and the order in which they were
exposed to the two products (X2(1)=3.36, p> .05), did not
influence their decision to follow or not the product recom-
mendation. The only covariate with a significant influence
on subjects’ decision to follow or not the product recommen-
dation was their product class familiarity (X2(1)=8.13,
p<.005). Subjects who perceived themselves as more fa-
miliar with the products selected the recommended product
Table 4
GEE analysis results for recommendation influence
Main effects and interactions df X2
Product 1 20.27a
Website 2 0.35
Recommendation 2 11.61a
Product ×Website 2 0.06
Product ×Recommendation 2 0.94
Website ×Recommendation 4 6.21
Product ×Website ×Recommendation 4 1.78
Contrast Tests for Recommendation
Rec. System vs. Other Consumers 1 11.69a
Rec. System vs. Experts 1 4.28b
Other Consumers vs. Experts 1 2.05
S. Senecal, J. Nantel/Journal of Retailing 80 (2004) 159–169 167
more often than subjects who perceived themselves as less
familiar with the product category. The differences between
the two groups were significant for the wine product class
(F(1,196)=18.777, p<.05) but not for the calculator
product class (F(1,203)=0.007, p> .05).
H2 stipulated that consumers would be more influenced
by recommendations for experience products than for
search products. Results provided strong support for H2
(X2(1)=20.27, p<.005). Recommendations for wine
were more influential than recommendations for calcula-
tors. H3 suggested that product recommendation influence
is greater on more independent websites. Results did not
support H3: no relationship was found between the type
of website and subjects’ propensity to follow product rec-
ommendations (X2(2)=0.35, p> .05). H4 stipulated that
recommender systems are more influential than other con-
sumers and human experts. In support of H4, contrast tests
showed that the recommender system was more influential
than other consumers (X2(1)=11.69, p<.005) and human
experts (X2(1)=4.28, p<.05). As illustrated in Table 3,
more subjects followed the product recommendation when
the product recommendation came from the recommender
system. An additional contrast test also revealed that no
significant difference existed between human experts and
other consumers relative to their influence on subjects’
choices (X2(1)=2.05, p> .05).
Recommendation source credibility
In order to test H5 and H6, a MANCOVA for repeated
measures using only observations from subjects who had
consulted a product recommendation (n=412) was per-
formed in order to test the relationships between the type of
recommendation source (H5) and the type of website (H6)
and the perceived trust and expertise of recommendation
sources. The order in which subjects were exposed to the
two different products was used as a covariate in the anal-
ysis. Results showed no significant two-way or three-way
interactions, but revealed a main effect for the recommenda-
tion source manipulation. As predicted by H5a, differences
were observed between perceived expertise of recommenda-
tion sources (F(2,204)=14.089, p<.001). Contrast tests
revealed that other consumers were perceived as less expert
than human experts (Mean (M) =4.4 and 5.2 respectively;
Contrast Estimate (C.E.) =−0.771, p<.001) and recom-
mender systems (M=4.7; C.E.=−0.356, p<.05). It
is noteworthy that human experts were perceived as more
expert than recommender systems (C.E.=0.415, p<.05).
As stated in H5b, differences in recommendation sources’
trustworthiness were also observed (F(2,204)=3.679, p<
.05). Contrast tests showed that the recommendation source
“other consumers” was perceived as significantly more
trustworthy than the recommendation source “recommender
system” (M=5.1 and 4.6, respectively; C.E.=0.438,
p<.01) but as trustworthy as human experts (M=4.9;
C.E.=0.192, p> .05), thus providing partial support to
H5b. Interestingly, human experts were perceived as trust-
worthy as recommender systems (C.E.=0.245, p> .05).
Finally, H6 was not supported. No significant differences
were found between the trustworthiness of recommendation
sources among the different types of websites (M=5.0
(retailer), 4.9 (Third party commercially linked to retailers),
4.9 (Non-commercially linked third party); F(2,204)=
1.310, p> .05).
Discussion and conclusion
Results strongly support our contention that consumers
are influenced in their online product choices by online
recommendations. However, all online recommendation
sources are not equally influential. The recommender sys-
tem was found to be the most influential recommendation
source even if it was perceived as possessing less expertise
than human experts and as being less trustworthy than other
consumers. In addition, recommendations for the experience
product were significantly more influential than recommen-
dations for the search product. The type of website on which
recommendation sources were used did not affect their
perceived trustworthiness and did not influence consumers’
propensity to follow the product recommendation.
This paper’s main theoretical implication is related to the
influence of recommender systems on consumers’ online
choices. With the emergence of the Internet, consumers now
have access to new impersonal sources of influence that can
provide personalized product information and recommen-
dations. Results show that this type of information source
indeed influences consumers’ online product choices, and
that it is more influential than conventional recommendation
sources. Thus, this study contributes to an emergent con-
sumer research area, namely the use and influence of imper-
sonal information sources providing personalized informa-
tion (e.g., recommender systems and intelligent agents) on
consumers’ decision-making processes.
This paper also has implications for marketers. First, re-
sults show that online recommendation sources influence
consumers’ online choices: products were selected twice as
often if they were recommended. Second, this influence is
moderated by the type of recommendation source and the
type of product, but it is not moderated by the type of web-
site. Thus, results suggest that a specific recommendation
source will be as effective on a retailer website (e.g., Ama-
zon) as it will be on an independent third party website such
as Consumer Reports. It seems that consumers focus much
more on the recommendation source itself than on the type
of website on which the recommendation source is used.
Since research on online recommendation sources is
emergent, this study points to many research avenues. First,
since impersonal information sources are used by and in-
fluence consumers, an effort should be made to develop
and/or adapt existing tools related to information source
influence. For instance, it would be useful to develop a
susceptibility to relevant others influence measurement
168 S. Senecal, J. Nantel/Journal of Retailing 80 (2004) 159–169
scale that would both include personal information sources
(Bearden, Netemeyer, & Teel, 1989) and these new infor-
mation sources. Second, results of the present study suggest
that recommender systems are perceived as less trustworthy
than other consumers. As suggested by Urban et al. (1999),
trust is not instantaneous and increases over multiple suc-
cessful interactions. Since our purpose was not to perform
a longitudinal study, our results must be interpreted with
caution. It is plausible that a customer’s trust in a specific
recommender system would increase over time if he/she
were satisfied with products previously recommended by
that system. Similarly, it would be of interest to explore
if customer loyalty to a specific website (Srinivasan et al.,
2002) acts as a moderator of the influence of online recom-
mendation sources. In addition, results of the present study
do not provide an explanation as to why recommender sys-
tems are more influential than human experts based on the
traditional source credibility approach. Additional studies
should investigate the effect of information personaliza-
tion on perceived source trust and expertise. Third, in the
present study only a limited product assortment was used
(i.e., four products) in each product class. It would be very
interesting to study how product assortment (see Simonson,
1999 for a review) affects the influence of online recom-
mendation sources. For instance, an increase in the number
of alternatives presented or the presence of a clearly domi-
nant alternative not recommended may affect the credibility
and influence of online recommendation sources. Fourth,
it would be interesting to examine if the decision to fol-
low (or not to follow) an online product recommendation
affects consumers’ satisfaction with their online shopping
experience (Szymanski & Hise, 2000). Finally, additional
variables need to be investigated to better understand why
consumers follow online product recommendations. Vari-
ables such as online shopping familiarity and experience,
past experience with online recommendations, age, and time
starvation (Lohse, Bellman, & Johnson, 2000;Wood, 2002)
may help explain why consumers use recommendations in
their online decision-making process.
This study has some limitations that should be kept in
mind before applying the results to real market situations.
First, although three different sampling frames were used for
this study, they were all used to generate convenience sam-
ples. Thus, as with most online studies, due to the possible
self-selection bias and low response rates it is not possible to
confirm that our participants are representative of the pop-
ulation of Internet shoppers. Second, results of the present
study are limited to only one search product (i.e., calcula-
tor) and one experience product (i.e., wine). Thus, additional
studies conducted with different samples and different prod-
ucts would contribute to the generalization of the present.
Finally, this study only investigated consumers’ online prod-
uct choices; it did not investigate online purchases. Thus,
additional variables such as product price, product availabil-
ity or delivery time could also affect how consumers are
influenced by online product recommendations.
The authors would like to thank the co-editors and the
three reviewers for their helpful comments. This research
was supported by the RBC Financial Group Chair of
E-Commerce and by the Center for Interuniversity Research
and Analysis of Organizations (CIRANO).
Alba, J., Lynch, J., Weitz, B., Janiszewski, C., Lutz, R., Sawyer, A., et al.
(1997). Interactive home shopping: Consumer, retailer, and manufac-
turer incentives to participate in electronic marketplaces. Journal of
Marketing,61(3), 38–53.
Andreasen, A. R. (1968). Attitudes and customer behavior: A decision
model. In H. H. Kassarjian & T. S. Robertson (Eds.), Perspectives
in consumer behavior (pp. 498–510). Glenview, IL: Scott, Foresman
and Company.
Ansari, A., Essegaier, S., & Kohli, R. (2000). Internet recommendation
systems. Journal of Marketing Research,37(3), 363–375.
Ardnt, J. (1967). Role of product-related conversations in the diffusion
of a new product. Journal of Marketing Research,4(3), 291–295.
Bakos, Y. J. (1997). Reducing buyer search costs: Implications for elec-
tronic marketplaces. Management Science,43(2), 1676–1692.
Bearden, W. O., & Etzel, M. J. (1982). Reference group influence on
product and brand purchase decisions. Journal of Consumer Research,
9(2), 183–194.
Bearden, W. O., Netemeyer, R. G., & Teel, J. E. (1989). Measurement
of consumer susceptibility to interpersonal influence. Journal of Con-
sumer Research,15(4), 473–481.
Brown, J. J., & Reingen, P. H. (1987). Social ties and word-of-mouth
referral behavior. Journal of Consumer Research,14(3), 350–362.
Childers, T. L., & Rao, R. (1992). The influence of familial and peer-based
reference groups. Journal of Consumer Research,19(2), 198–212.
Cyber Dialogue (2001). The personalization consortium’s online consumer
personalization survey.
Duhan, D. F., Johnson, S. D., Wilcox, J. B., & Harell, G. D. (1997). In-
fluences on consumer use of word-of-mouth recommendation sources.
Journal of the Academy of Marketing Science,25(4), 283–295.
Folkes, V. S. (1988). Recent attribution research in consumer behavior:
A review and new directions. Journal of Consumer Research,14(4),
Flynn, L. R., & Goldsmith, R. E. (1999). A short, reliable measure of
subjective knowledge. Journal of Business Research,46(1), 57–66.
Gilly, M. C., Graham, J. L., Wolfinbarger, M. F., & Yale, L. J. (1998). A
dyadic study of personal information search. Journal of the Academy
of Marketing Science,26(2), 83–100.
Harmon, R. R., & Coney, K. A. (1982). The persuasive effects of source
credibility in buy and lease situations. Journal of Marketing Research,
19(2), 255–260.
Häubl, G., & Trifts, V. (2000). Consumer decision-making in online shop-
ping environments: The effects of interactive decision aids. Marketing
Science,19(1), 4–21.
Hoffman, D. L., Novak, P. T., & Chatterjee, P. (1995). Commercial sce-
narios for the web: Opportunities and challenges. Journal of Computer
Mediated Communication,1(1).
Kelley, H. H. (1967). Attribution Theory in Social Psychology.InD.
Levine (Ed.), Nebraska symposium on motivation (pp. 192–241). Lin-
coln, NE: University of Nebraska Press.
Kelley, H. H. (1973). The process of causal attribution. American Psy-
chologist,28, 107–128.
Kelman, H. C. (1961). Processes of opinion change. Public Opinion
Quarterly,25, 57–78.
S. Senecal, J. Nantel/Journal of Retailing 80 (2004) 159–169 169
King, M. F., & Balasubramanian, S. K. (1994). The effects of exper-
tise, end goal, and product type on adoption of preference formation
strategy. Journal of the Academy of Marketing Science,22(2), 146–
Lascu, D.-N., Bearden, W. O., & Rose, R. L. (1995). Norm extremity
and personal influences on consumer conformity. Journal of Business
Research,32(3), 201–213.
Liang, K.-L., & Zeger, S. L. (1986). Longitudinal data analysis using
generalized linear models. Biometrika,73(1), 13–22.
Liang, K.-L., Zeger, S. L., & Qaqish, B. (1992). Multivariate regression
analyses for categorical data. Journal of the Royal Statistical Society,
54, 3–40.
Lohse, L. G., Bellman, S., & Johnson, E. J. (2000). Consumer behavior
on the Internet: Findings from panel data. Journal of Interactive
Marketing,14(1), 15–29.
Lynch, J. G., & Ariely, D. (2000). Wine online: Search costs affect
competition on price, quality, and distribution. Marketing Science,
19(1), 83–103.
Maes, P. (1999). Smart commerce: The future of intelligent agents in
cyberspace. Journal of Interactive Marketing,13(3), 66–76.
McGuire, W. J. (1969). The nature of attitudes and attitude change.
In G. Lindzey & E. Aronson (Eds.), The handbook of social psy-
chology (pp. 137–314). Reading, MA: Addison-Wesley Publishing
Mizerski, R. W., Golden, L. L., & Kernan, J. B. (1979). The attribution
process in consumer decision making. Journal of Consumer Research,
6(2), 123–140.
Nelson, P. (1970). Information and consumer behavior. Journal of Political
Economy,78(2), 311–329.
Nelson, P. (1974). Advertising as information. Journal of Political Econ-
omy,83(4), 729–754.
Nosek, B. A., Banaji, M. R., & Greenwald, A. G. (2002). E-research:
Ethics, security, design, and control in psychological research on the
Internet. Journal of Social Issues,58(1), 161–176.
Ohanian, R. (1990). Construction and validation of a scale to measure
celebrity endorsers’ perceived expertise, trustworthiness, and attrac-
tiveness. Journal of Advertising,19(3), 39–52.
Olshavsky, R. W., & Granbois, D. H. (1979). Consumer decision-making—
fact or fiction. Journal of Consumer Research,6(2), 93–100.
Park, C. W., Mothersbaugh, D. L., & Feick, L. (1994). Consumer
knowledge assessment. Journal of Consumer Research,21(2), 71–
Perdue, B. C., & Summers, J. O. (1986). Checking the success of ma-
nipulations in marketing experiments. Journal of Marketing Research,
23(4), 317–326.
Postma, O. J., & Brokke, M. (2002). Personalisation in practice: The
proven effects of personalisation. Journal of Database Marketing,
9(2), 137–142.
Price, L. L., & Feick, L. F. (1984). The role of recommendation sources
in external search: An informational perspective. In T. Kinnear (Ed.),
Advances in consumer research (Vol. 11, pp. 250–255). Provo, UT:
Association for Consumer Research.
Rosen, D. L., & Olshavsky, R. C. (1987). The dual role of informational
social influence: Implications for marketing management. Journal of
Business Research,15(2), 123–144.
Senecal, S., & Nantel, J. (2002). Online influence of relevant oth-
ers: A framework (Working Paper). RBC Financial Group Chair of
E-Commerce, HEC Montreal, University of Montreal.
Simonson, I. (1999). The effect of product assortment on buyer prefer-
ences. Journal of Retailing,75(3), 347–370.
Spiller, P., & Lohse, J. (1998). A classification of Internet retail stores.
International Journal of Electronic Commerce,2(2), 29–56.
Srinivasan, S. S., Anderson, R., & Ponnavolu, K. (2002). Customer loyalty
in E-commerce: An exploration of its antecedents and consequences.
Journal of Retailing,78(1), 41–50.
Sternthal, B., Phillips, L. W., & Dholakia, R. (1978). The persuasive
effect of source credibility: A situational analysis. Public Opinion
Quarterly,42(3), 285–314.
Stokes, M. E., Davis, C. S., & Koch, G. G. (2001). Categorical data
analysis using the SAS System. Cary, NC: SAS Institute Inc.
Szymanski, D. M., & Hise, R. T. (2000). e-Satisfaction: An initial exam-
ination. Journal of Retailing,76(3), 309–322.
The e-tailing Group (2003). 2nd Annual Merchant Survey.
Tybout, A. M. (1978). Relative effectiveness of three behavioral influence
strategies as supplements to persuasion in a marketing context. Journal
of Marketing Research,15(2), 229–242.
Urban, G., Sultan, F., & Qualls, W. (1999). Design and evaluation of a
trust based advisor on the Internet (Working Paper 40). e-Commerce
Research Forum, MIT.
West, P. M., Ariely, D., Bellman, S., Bradlow, E., Huber, J., Johnson,
E., et al. (1999). Agents to the rescue? Marketing Letters,10(3),
Wind, J., & Rangaswamy, A. (2001). Customerization: The next revolution
in mass communication. Journal of Interactive Marketing,15(1), 13–
Wood, S. L. (2002). Future fantasies: A social change perspective of
retailing in the 21st century. Journal of Retailing,78(1), 77–83.
Zeger, S. L., Liang, K.-L., & Albert, S. (1988). Models for longitudinal
data: A generalized estimating equation approach. Biometrics,44,
... Overall, eWOM is more influential and effective compared to more official and controlled means of communication (Bakos & Dellarocas, 2011;Duan, Gu, & Whinston, 2008;Liu, 2020), and can be seen as less biased, as e-commerce platforms allow several buyers to share different opinions simultaneously (Lee, Park, & Han, 2008;Senecal & Nantel, 2004). In addition, eWOM is relatively easy to interpret and operationalize (Floyd et al., 2014;Park & Kim, 2008). ...
Full-text available
This study investigates the contemporary role of electronic word-of-mouth (eWOM) in business exchanges through buyers' signaling of observable and unobservable supplier characteristics on the Alibaba e-commerce platform. Utilizing a qualitative pre-study of 20 interviews (five buyer–supplier dyads with two interviews per firm) and more than 8000 buyer reviews on Alibaba, we identify characteristic patterns of this type of B2B eWOM. Signaling Theory and Social Exchange Theory underpin the empirical investigation. To enable further conceptualization, the study distinguishes between online B2B reviews based on the extent to which they are controlled by organizational partners. Unlike some other forms of B2B reviews, reviews on Alibaba are uncontrolled and comprise a form of eWOM. Findings indicate that the relational patterns of B2B eWOM shared on Alibaba can be aggregated into three categories: human touch, responsiveness, and resilience. Besides these new categories, the importance of product/service quality has been confirmed. Through Alibaba reviews, buyers' signals are sent not only to suppliers as feedback, but also to other prospective buyers to influence their purchase decisions. Our study aims to contribute to the B2B literature on eWOM and signaling in business relationships. By showing that human touch occurs even in online-only buyer–supplier relationships, the study provides evidence that bonding, the development of mutuality, and relationship intimacy in buyer–supplier relationships does not always require in-person contact. Managerial implications are offered with a focus on the signaling of unobservable qualities, such as human touch, with the help of B2B eWOM.
... However, because consumers have little first-hand evidence to form their own evaluation of the frontline service worker, they will have to rely more on indirect experience (positive or negative) to assess the service performance. In particular, for offerings that usually would require more direct experience with the provider to be evaluated, consumers are more likely to align their attitudes and behaviors with the evidence provided by indirect experience (Huang et al., 2009;Park and Lee, 2009;Senecal and Nantel, 2004). In other words, under these circumstances, consumers tend to go along with the consensus of the majority, a manifestation of the consensus heuristic (Axson et al., 1987) and use this as a social norm to justify their behavior (Fogel and Zachariah, 2017;Sridhar and Srinivasan, 2012). ...
Full-text available
Purpose. Academic research has supported the belief that consumers undertip minority race service workers due to implicit racial biases. However, there has been less focus in examining possible moderating factors. This paper fills this gap by analyzing the role of direct and indirect experience in tipping frontline service workers from a minority background. Given the prominence of customer ratings on digital service platforms and the perception that African-Americans are discriminated against, we look at the interplay of interaction length (direct experience) and customer ratings (indirect experience) on the relationship between race and tipping. Design/methodology/approach. An expectancy disconfirmation framework was developed and tested with a sample of 360 US participants in an online experiment. The experiment followed a 2 x (race: African-American versus Caucasian) x 2 (direct experience: limited versus extensive) x 3 (indirect experience: absent versus positive versus negative customer rating) design. Findings. We found consumers who have extended direct experience (longer service interaction) and no indirect experience (absent customer ratings) tipped African-Americans more than Caucasians. Interestingly, this effect is reduced in the presence of indirect experience (customer ratings). Lastly, where the consumer lacks direct experience (shorter service interaction) but is exposed to positive indirect experience (positive customer ratings), consumers tip African-Americans more. Originality/value. This is the first paper that examines the role of direct and indirect experience in the relationship between race and tipping. Based on our findings, we provide several contributions, including recommendations to reduce inequalities arising from implicit racial bias on digital service platforms.
This paper examines the effects of loyalty expressions (i.e., repurchase intentions vs. recommendations) on review persuasiveness. Specifically, we propose that repurchase intentions have a stronger positive effect on review persuasiveness compared to recommendations because of reviewer credibility. We test the above proposition using both an empirical dataset and multiple experimental studies. In addition, we examine frequency of purchase as a boundary condition for our proposition. Accordingly, we find that for frequent purchases, repurchase intentions (vs. recommendations) increases credibility, which, in turn, augments review persuasiveness. For infrequent purchases, however, we observe that recommendations (vs. repurchase intentions) enhance review persuasiveness, which occurs because of increased credibility. This research offers contributions to theory in the areas of online reviews, loyalty, source credibility, and cue-diagnosticity, as well as to practice regarding how firms should seek to elicit loyalty expressions (i.e., repurchase intentions vs. recommendations) when soliciting reviews.
Because e-WOM is one of the useful digital marketing elements for any organization, a better understanding of its process will help individuals take more advantage of this concept. e-WOM enables individuals to form relationships with firms, brands, and other customers, which leads to benefits for both consumers and companies. It plays a significant role in a firm’s performance. The present study implements a different approach to reviewing by combining two bibliometric methods, multidimensional scaling analysis (MDS) and hierarchical cluster analysis (HCA), via Bibexcel software to have a deeper investigation of the process. Considering the 468 journal papers on e-WOM allowed us to study the intellectual streams and significant perceptions underpinning e-WOM. By dividing the study timeframe into three periods, we realized that there have always been three main concepts in this field: consumer behavior, sales, and the tourism and hotel industry. Further, by proposing a framework, we have expanded these concepts accompanied by the role of artificial intelligence and robots in the process of e-WOM. Consequently, new concepts “r-WOM”, “automated user engagement”, and “smart selling” are introduced and demonstrated as a consequence of using technology-based tools in the process of e-WOM. Finally, the future scope of this field has been designed. We contribute to the literature by offering theoretical and managerial implications.
This study builds on the efforts to systematize the effects of online reviews, trust, and attitude on behavioral intentions. Specifically, we report the direct and indirect effects of variables on behavioral intentions through different models that compare the two types of restaurants (casual vs. fine dining). Our findings provide ample evidence that trust and attitude are directly related to behavioral intentions when online reviews do not directly influence behavioral intentions for both restaurant types. The findings also show that the online review-attitude-behavior link of Model 4 is more stable and robust than in Model 3. In particular, the indirect role of online reviews on behavioral intentions enhances the effects of the attitude-behavior relationship in both casual and fine dining restaurants.
Information and communication technologies have empowered the consumers with ease of access to information about products and services in form of electronic word of mouth (eWOM). eWOM plays a critical role in consumer purchase decisions in the form of online reviews. The study uses the S-O-R framework to examine the impact of online reviews on online hotel booking. The data were collected using purposive sampling through an online self-administered questionnaire and analyzed using the PLS-SEM technique. The results of the study show that review valence significantly impacts review credibility whereas review length does not. Furthermore, credible reviews lead to high purchase intention. The positive impact of purchase intention on purchase in online hotel booking context is a novel finding. We suggest managers to include more positive valanced reviews to develop trust and credibility while take proper measures to reduce risk.
Purpose Prior hospitality studies have reviewed review trustworthiness and perceived price as predictors of restaurant selection. However, the impacts of these two factors may vary by sales promotion and customer types. This study aims to determine whether sales promotions and customer type are the key elements that facilitate behavioral intentions by moderating the linkage between perceived price and behavioral intentions as well as the linkage between online review trustworthiness and behavioral intentions. Design/methodology/approach Analysis of the responses of 533 individuals familiar with the Michelin Guide for restaurants in Seoul provided evidence supporting a sales promotion theory wherein promotions signal benefits in consumers’ minds. Findings The findings show that when perceived price is positive and the trustworthiness of online reviews is high, repeat customers prefer mixed coupons to price discounts. Notably, the results indicate that when the trustworthiness of online reviews is high, first-time customers also prefer mixed coupons to price discounts. Furthermore, the findings suggest that negative evaluations of perceived price increase the impact of mixed coupons by signaling to first-time customers that given restaurants’ offerings provide monetary benefits regardless of their intentions to revisit said restaurants. Research limitations/implications The study findings provide insights that should help managers better understand various levels of promotion. Managers can design their pricing strategies to strengthen customers’ motivations to visit their restaurants – the very thing customers often seek in sales promotions. Originality/value This study provides indisputable evidence for a sales promotion theory, wherein promotions signal benefits in consumers’ minds; however, it also shows that first-time and repeat customers do not respond equally to sales promotions.
The digital marketing transformation of the Internet has significantly experienced a paradigm shift, i.e., a transformation from a passive source of information to an interactive and engaging participatory web. This study demonstrates the ability of electronic word-of-mouth (eWOM) as a participatory web tool that enables enterprises to achieve profitable growth, resilience, business sustainability, and competitiveness, through developing operational strategy. This study adopts a conclusive descriptive cross-sectional survey research design, which allows the collection of quantitative data through structured questionnaires. The data were obtained from Egyptian social media users through a convenience sampling method. To test the hypotheses regression analysis was conducted. Results indicate that eWOM positively influences the brand image and purchase intention, which consequently enables the enterprises to achieve business sustainability. Accordingly, enterprises wanting to achieve strategic competitiveness must integrate social media into their marketing mix which would generate positive eWOM. Using convenience sampling might result in the inability to generalize the findings. This study is designed to study the effect of eWOM using social media platforms in general, however, future studies should replicate this study to specified types of different social media platforms, to identify which platform generates the highest impact. The proposed conceptual model tests a relationship that connects eWOM dimensions, namely, credibility, quality, and quantity, to purchase intention and brand image. There is lack of research in the Egyptian context on the implications of eWOM on enterprise competitiveness and sustainability
The payment method is a key factor influencing online consumers’ purchase decisions. However, little is known about its underlying neural basis, which could help to reveal the mechanism by which payment methods affect online purchase decisions. Combined with the event-related potentials (ERPs), a neuroscience technique with the advantage of measuring implicit psychological variables to reveal the mechanism behind behaviors, this study uncovers consumers’ discrepant perception between pay-online and pay-on-delivery in different purchase contexts through an online purchase task. Behavioral results showed that purchase intention is higher for pay-online than pay-on-delivery, regardless of product type. At the brain level, we found consumers induce higher perceived risk (indicated in larger N2 amplitudes) and smaller negative emotion (mirrored by larger P3 amplitudes) for pay-online than pay-on-delivery, especially when shopping for search products. However, this effect disappeared when purchasing experience products. Moreover, the larger perceived risk for experience products than search products may lead consumers to ignore the difference between the two payment methods. This study helps online sellers optimize payment services for specific products.
It is common to observe a vector of discrete and/or continuous responses in scientific problems where the objective is to characterize the dependence of each response on explanatory variables and to account for the association between the outcomes. The response vector can comprise repeated observations on one variable, as in longitudinal studies or genetic studies of families, or can include observations for different variables. This paper discusses a class of models for the marginal expectations of each response and for pairwise associations. The marginal models are contrasted with log‐linear models. Two generalized estimating equation approaches are compared for parameter estimation. The first focuses on the regression parameters; the second simultaneously estimates the regression and association parameters. The robustness and efficiency of each is discussed. The methods are illustrated with analyses of two data sets from public health research.
Attitude and opinion data provide a basis for inferring the meaning of opinions held by individuals and groups and also for predictions about their future behavior. Such inferences and predictions, if they are to be made effectively, require a theoretical foundation which explains the processes by which people adopt and express particular opinions. Here is a theory of three processes by which persons respond to social influence.
The relative effectiveness of three influence strategies in gaining acceptance of a new service advocated by either a high or low credibility source was determined. Although the influence strategies did not differ in their overall effectiveness, the optimal strategy varied as a function of level of source credibility. These results were obtained in both personal selling and mass-media-like contexts. The theoretical, methodological, and practical implications of these findings are discussed.
The authors discuss several issues in the timing, construction, and analysis of manipulation and confounding checks in marketing experiments. A review of 34 experiments involving latent independent variables reported in the "Journal of Marketing Research" over the past decade suggests that most researchers are familiar with the concept of manipulation checks but few systematically evaluate potential sources of confounding in experimental manipulations. Three alternative approaches for assessing the construct validity of experimental manipulations also are discussed.
The purpose of this study was to develop a scale for measuring celebrity endorsers' perceived expertise, trustworthiness, and attractiveness. Accepted psychometric scale-development procedures were followed which rigorously tested a large pool of items for their reliability and validity. Using two exploratory and two confirmatory samples, the current research developed a 15-item semantic differential scale to measure perceived expertise, trustworthiness, and attractiveness. The scale was validated using respondents' self-reported measures of intention to purchase and perception of quality for the products being tested. The resulting scale demonstrated high reliability and validity.