Consumer Behavior in Web Based Commerce An Empirical Study

International Journal of Electronic Commerce
ISSN: 1086-4415 (Print) 1557-9301 (Online) Journal homepage: https://www.tandfonline.com/loi/mjec20
Consumer Behavior in Web-Based Commerce: An

Empirical Study
Marios Koufaris , Ajit Kambil & Priscilla Ann Labarbera
To cite this article: Marios Koufaris , Ajit Kambil & Priscilla Ann Labarbera (2001) Consumer
Behavior in Web-Based Commerce: An Empirical Study, International Journal of Electronic
Commerce, 6:2, 115-138, DOI: 10.1080/10864415.2001.11044233
To link to this article: https://doi.org/10.1080/10864415.2001.11044233
Published online: 23 Dec 2014.
Submit your article to this journal
Article views: 946
View related articles
Citing articles: 34 View citing articles
Full Terms & Conditions of access and use can be found at

https://www.tandfonline.com/action/journalInformation?journalCode=mjec20
Consumer Behavior in Web-Based Commerce: An
Empirical Study
Marios Koufaris, Ajit Kambil, and Priscilla Ann LaBarbera

ABSTRACT: Electronic commerce challenges companies to design electronic systems and
interactions that retain customers and increase sales. This exploratory study examines the
impact of consumer experience and attitudes on intention to return and unplanned pur-
chases on-line. It also examines how certain consumer and Web site factors influence the
on-line consumer experience. The study finds that perceived control and shopping enjoy-
ment can increase the intention of new Web customers to return, but seemingly do not
influence repeat customers to return. It also finds that a Web store that utilizes value-
added search mechanisms and presents a positively challenging experience can increase
customers’ shopping enjoyment. Further, the more often customers return to a Web store,
the more their shopping enjoyment is determined by their product involvement. Customers
with low need specificit y (i.e., who do not know what they are looking for) are more likely
to use value-added search mechanisms. Finally, neither perceived control nor shopping
enjoyment has any significant impact on unplanned purchases.
KEY WORDS AND PHRASES: Consumer behavior, customer retention, electronic com-
merce, unplanned purchases, Web customers.
The World Wide Web has revolutionized commerce, with on-line retail spend-
ing projected to reach almost $120 billion by the year 2003 [35]. In business-to-
consumer electronic commerce there have been two transformations: (1) the
transformation of the consumer into a computer user, and (2) the transforma-
tion of the physical store into a phenomenon that is information technology
intensive—in other words, a Web site. The new double role of consumer/
computer user resulting from the first transformation is inherently complex
and still not well understood. The second transformation concerns store tech-
nology. In the physical commercial world, the information technology used in
a store is most often invisible to the consumer, whereas in e-commerce it has
been moved to the foreground. To the consumer, the Web site, an information
system, is a full representation of the store. Therefore, it is necessary to look at
the interaction between consumer and store through the filter of user-technol-
ogy interaction. Given these transformations, researchers need to combine the
research streams of information systems, psychology, and marketing to study
and understand on-line consumer behavior.
Empirical research has tested some of the effects of electronic marketing on
consumer behavior. There are, for example, indications that traditional mar-
keting promotions are not always successful in on-line commerce [33]. While
convenience and control are at the top of consumers’ lists of benefits from on-
line shopping [10, 30], enjoyment of the on-line shopping experience may also
be an important determinant of customer loyalty [52]. The perceived risks
associated with on-line shopping influence attitudes toward on-line purchas-
ing, as does the perceived ease of using the Web site [24]. The impact of per-
ceived ease of use, however, seems to vary depending on the type of task the
International Journal of Electronic Commerce / Winter 2001–2002, Vol. 6, No. 2, pp. 115–138.
Copyright © 2002 M.E. Sharpe, Inc. All rights reserved.
1086-4415/2002 $9.50 + 0.00.
116 KOUFARIS, KAMBIL, AND LABARBERA
consumer is undertaking. The effect is more significant when consumers are

using a Web site to inquire about products rather than purchase them [20].
Consumers who use the Internet search more extensively before making a
purchase than non-users [50]. “Adaptive” Web sites that change their inter-
faces in accordance with user behavior improve user performance but not
overall satisfaction [59].
The present exploratory study builds on emerging information systems and
marketing research to examine the factors that lead to increased customer
loyalty and unplanned purchases on-line. It explores how the level of per-
ceived control and shopping enjoyment affect the behavior of on-line con-
sumers. Finally, it examines the question of how individual consumer factors
(need specificity, customer tenure, product involvement) and Web store fac-
tors (search mechanisms used, positive challenges ) influence consumer atti-
tudes and behavior on the Web.
Theoretical Framework
Environmental and individual characteristics are known to influence emo-
tional responses that, in turn, affect behavioral responses. In the area of envi-
ronmental psychology, three key emotional variables that may determine
behavior are pleasure, dominance, and arousal [36]. These variables are also
important determinants of consumer behavior off- and on-line. In a consumer
context, pleasure can be defined as shopping enjoyment, and dominance as the
equivalent of the perceived control felt by customers. Arousal deals with more
physical sensations, such as being jittery, sluggish, or relaxed. In this study, it
was not possible to directly measure physical responses to the on-line shopping
experience, and instead the research concentrated on the first two variables.
The theoretical framework for this study, as can be seen in Figure 1, in-
cludes shopping enjoyment and perceived control as attitudinal variables that
directly affect on-line consumer behavior. In accordance with both environ-
mental psychology and traditional consumer behavior, it also examines the
direct or moderating impact of individual and Web store (environmental) fac-
tors on consumer attitudes and behavior [17, 36].
Dependent Variables
Unplanned Purchases
While customers usually have specific purchases in mind when they enter a
store, they often end up making unplanned purchases, either on impulse or
because of an in-store promotion that acts as a reminder or suggestion [58].
Such purchases provide significant profit opportunities for firms [29]. In the
physical world, unplanned purchases, and especially pure impulse ones, are
strongly determined by the time and money available for shopping, enjoy-
ment of the shopping experience, the level of impulse-buying tendency, and
exposure to marketing promotions [3, 5].
INTERNATIONAL JOURNAL OF ELECTRONIC COMMERCE 117
Need Specificity Product Involvement Customer Tenure

(new vs. repeat)
Shopping UNPLANNED
Search Mechanism PURCHASES
Used Enjoyment
Perceived
Perceived INTENTION TO
Positive Challenges Control
Control RETURN
Figure 1. Theoretical Framework for Consumer Attitudes and Behavior

on the Web
In some situations on the Web, conditions that enable a consumer to save

time could lead to more unplanned purchases. The ability to search more prod-
ucts, services, and information may lead to longer purchasing decisions with
a negative impact on unplanned purchases. Money availability is as impor-
tant on-line as off-line. Shopping enjoyment, however, is a challenge to Web-
based stores. They cannot provide the full experience of walking through a
store surrounded by carefully chosen displays, listening to mood-enhancing
music, seeing all the products (as well as touching them, smelling them, and
sometimes tasting them), trying the products out, and seeing other custom-
ers. The limited experience of buying on-line may decrease enjoyment and
reduce unplanned sales.
The ability to target specific customers on the Web can lead to more effi-
cient marketing promotions that may increase unplanned purchases. How-
ever, on-line customers have unprecedented control over what they see and
do. The availability of search engines, intelligent agents, recommender sys-
tems, and the user-friendliness of the Web browser gives Web consumers con-
siderable power [4, 23, 48]. This means that consumers have more choice over
the material they view and the advertising and marketing communications
they are exposed to [14, 49, 55, 56]. Small marketing promotions have no sig-
nificant effect on traffic and sales [32]. Since most Web companies allow cus-
tomers to go directly to their products without being exposed to marketing
promotions, the effect may be a reduction in unplanned purchases [9]. Be-
cause unplanned purchases are a key driver of profit, an understanding of
the factors driving them is crucial.
Customer Intention to Return

Ensuring that customers return is a primary goal of almost all companies,
including Web-based companies [46, 51]. Nonetheless, customer loyalty is a
challenge to Web-based companies. For example, in a survey by Yahoo Store
(store.yahoo.com), a store-hosting service, more than 85 percent of stores re-
ceived fewer than 10 percent of their orders from repeat customers [2]. In
other words, very few customers returned for a second purchase after buying
for the first time. There are three possible explanations for store loyalty [15]:
1. Resource constraints: Store loyalty results from a lack of consumer

resources, such as time or money.
2. Non-shopping lifestyle: Store loyalty results from lack of consumer
interest in shopping due to other commitments or personality differ-
ences. Customers tend to always shop at the same stores because of
inertia rather than loyalty.
3. Discretionary loyalty: Store loyalty results from the convenience of
shopping at a large store that offers one-stop shopping, such as a
department store.
In light of this, store loyalty for Web-based companies can decrease be-
cause consumers do not have to spend substantial amounts of money or time
for transportation from store to store. They can buy from any company avail-
able on-line while sitting in front of their computer. Also, discretionary loy-
alty is no longer necessary. The center for one-stop shopping is now the
customer’s computer, where switching from one store to another is easy, not
requiring significant physical effort, though it may require some learning of
the new Web site. For consumers who dislike shopping, consumer loyalty on
the Web may not decrease significantly, since even the small switching costs
on the Web can prove too much for them. All of these factors suggest that the
design of satisfactory experiences will be crucial to store loyalty.
Customers who view shopping as a chance to get out, socialize, and have
fun may find a Web store unappealing and therefore choose not to return [40,
52]. A very important consideration determining whether customers will re-
turn to a Web site is its relevant and non-frivolous content [52]. If customers
do not have a valuable and intrinsically enjoyable experience on-line, they
may return to the more enriching and enjoyable physical world.
Attitudinal Variables
The two attitudinal variables of the model presented in this paper, perceived
control and shopping enjoyment, have also been investigated in the context of
user attitudes and drivers of system use, and especially in the context of flow.
The founder of flow, Mihaly Csikszentmihalyi, calls it “the holistic sensation
that people feel when they act with total involvement” [11, 12, 13]. The sense
of perceived control and enjoyment that people feel when they are involved
in an activity are two of the main components of flow.
Trevino and Webster found that flow, partly measured as perceived control
and intrinsic enjoyment, was positively related to the attitudes, communica-
tion effectiveness, and amount of communication of users of computer sys-
tems [60]. Webster, Trevino, and Ryan found that flow was positively related
to perceptions of the flexibility and modifiability of the software used [63]. It
was also positively related to more experimentation in the use of the software
as well as to actual and expected computer use. Ghani and Deshpande exam-
ined flow in the context of individuals who used computers in their daily
work. They found that flow had a significant impact on exploratory use of the
computer that in turn had a significant effect on the extent of computer use
[21].
Novak et al. introduced flow in the study of on-line marketing and tested a
structural equation model of flow using an on-line survey of Web users [43].
They found that skill was a significant antecedent of flow but only indirectly
through perceived control. Challenge was an antecedent of flow through
arousal, and focused attention was also a significant antecedent. Overall, play-
fulness can be an important indicator of flow because it is predicted by the
antecedents of flow (skill, challenge, and focused attention) and leads to the
consequences of flow (positive experience, exploratory behavior, and greater
expected Web use) [62].
Consumer Control
Consumers regard time, particularly leisure time, as a scarce resource, and
their cognitive information-processing capability is constant and limited [17,
37, 47]. On-line customers have tremendous access to information about prod-
ucts and services from anywhere in the world and from sources other than the
product seller. The combination of less time available for shopping, limited
cognitive resources for information processing, and the explosion of informa-
tion on the Web has led to customer demands for more control, less effort, and
greater efficiency during shopping. Convenience and effort reduction are now
most important to on-line customers [10, 30].
Web stores have responded to the desire for customer control and conve-
nience with systems that enable consumers to easily find what they need,
learn more about it, and quickly make a purchase decision. Internal search
engines, hierarchical classifications of company products, and intelligent agents
are provided by many Web stores to support users’ need for control in access-
ing information. Sites offer quick, automated purchasing through “one-click”
buying and the use of shopping carts. All these site features enable Web cus-
tomers to enjoy higher levels of control and convenience [4, 23].
Shopping Enjoyment
Shopping can go beyond the utilitarian experience of fulfilling product needs.
It can be used to alleviate loneliness, eliminate boredom, fulfill fantasies, or
escape from everyday life [40]. Shopping can be intrinsically enjoyable [18],
and enjoyment of the shopping experience is an important determinant of
consumer behavior [7]. On the Web, shopping enjoyment is positively and
significantly related both to attitudes toward shopping on the Web and inten-
tions toward shopping on the Web [16, 30].
Shopping on-line is a different experience from shopping in a physical re-
tail store. One major point of difference deals with” store atmospherics” [17].
This term describes the physical aspects of a store, such as colors, music type,
music volume and tempo, and layout of products. Store atmospherics have a
direct effect on customer mood and behavior [15]. The atmospherics of Web-
based businesses are at best limited to the confines of a computer monitor that
usually displays only two-dimensional pictures and text. Even with the addi-
tion of three-dimensional images and musical accompaniment, Web stores
today cannot fully simulate the ambiance of a physical store. In the competi-
tion between click and mortar, the system design of the e-tailing experience
must compensate for the loss of traditional in-store atmospherics.
Impact of Attitudinal Variables on Consumer Behavior

According to the available literature on flow, perceived control and enjoy-
ment can both have a positive impact on the use of computer-mediated envi-
ronments [43, 60, 63]. A similar effect may be expected in on-line shopping. A
more positive shopping experience, where customers experience high levels
of shopping enjoyment, will undoubtedly increase the number of unplanned
purchases. A similar effect has been recorded in the physical world [5]. How-
ever, perceived control will reflect the actual control that on-line consumers
have over their actions. Therefore, as discussed earlier, they are less likely to
be exposed to advertising and promotions, leading to a decrease in their un-
planned purchases. Hence, one may hypothesize:
H1a: Perceived control will be negatively related to unplanned purchases.
H1b: Shopping enjoyment will be positively related to unplanned purchases.
Similarly the literature shows that flow is positively related to expected

future computer interactions [63]. This suggests that perceived control and
enjoyment may be positively related to expected return visits to the Web
site. Indeed, Rice’s survey of 87 Web sites found that an enjoyable visit de-
termined whether a customer would return to the site [52]. So too Jarvenpaa
and Todd found that shopping experience is positively related to shoppers’
attitudes and intentions toward shopping on the Web [30]. Thus one may
hypothesize:
H2a: Perceived control will be positively related to consumer intention to

return.
H2b: Shopping enjoyment will be positively related to consumer intention

to return.
Individual and Web Store Factors

Individual Factors
Customer Tenure (New vs. Repeat)
It is hypothesized here that customer tenure (i.e., how many times a customer
has been to a Web store in the past) will have a moderating effect on the rela-
tionship between attitudes and behavior. Since repeat customers have already
returned at least once to a store, their intention to return again may be deter-
mined by factors other than those that apply to new customers. For new cus-
tomers, the novelty of the Web site can be very important, but for repeat
customers this may not be the case. Other factors, such as service, product
availability, and personal involvement (discussed later), may be more impor-
tant to a repeat customer. Unplanned purchases may also vary based on the
tenure of the customer. New customers tend to be more exploratory in a new
store and are more susceptible to marketing messages and promotions. In this
study, the moderating effect of customer tenure on consumer attitudes and
behavior will be examined.
H3: Customer tenure will have a significant moderating effect on the

relationship between attitudes and behavior.
Need Specificity
Another individual factor is the specificity of the consumer’s needs: in other
words, how well individual consumers know what they want when they visit
a store. The scope and number of products consumers have in mind when
they begin their purchasing search has a significant impact on their behavior
[27, 28, 53]. Since the study deals mainly with commodities, product com-
plexity is low, and this makes it easier to measure the specificity of the con-
sumer need.
A product is characterized by a set of attributes—in the case of a book, for
instance, its title, author, edition, publisher, subject matter, and so on. Cus-
tomer need specificity is defined as the number of such attributes, with a fixed
or limited range of values, that the customer uses in searching for a product.
Some attributes are functionally dependent on others. For example, the author’s
name is functionally dependent on the book title, since for a specific title there
is only one (or more) specific author(s). Thus, if a customer has a fixed value
for the book title when conducting a search, need specificity will be high, since
other attributes that are functionally dependent on the title will also have fixed
values. The resulting set of alternative products after the customer has ap-
plied the search criteria is similar to evoked sets and consideration sets [27,
28, 53].
Product Involvement
Product involvement and its measurement have been the subject of consider-
able research and debate since the Personal Involvement Inventory (PII) was
proposed and analyzed by Zaichkowsky [64]. While there have been many
variations on the definition of involvement [22, 38, 45, 64], it is generally ac-
cepted that the concept comprises (a) one’s motivational state (i.e., arousal,
interest, drive) toward an object where (b) that motivational state is activated
by the relevance or importance of the object [39]. Much has been written about
involvement with advertisements as well as with the purchase process [1, 57],
but this article will concentrate on involvement with the product. It is im-
portant to note that product involvement is not constant for a product or
type of product but varies for each customer. For example, the importance
of and interest in music CDs ( i.e., product involvement) is different for each
consumer.
Web Store Factors

Search Mechanisms
Site search features, such as internal search engines, hierarchical classifica-
tions, and intelligent agents, give Web customers high levels of control and
convenience [4, 23]. In a survey by Internet World, 89 percent of 163 leading
Web design firms said that they used a search mechanism of some type to
lengthen and enhance their users’ experience [19]. The present study exam-
ines the effects of using different types of search mechanisms on the consumer
experience. There have been few attempts to create a framework or typology
of Web-based customer decision-support systems (CDSS) [44]. For the pur-
poses of this study, a preliminary two-dimensional framework of search mecha-
nisms used in Web-based stores has been developed.
The first dimension deals with the type of information used by the search
mechanism. There are two types of information: non-value-added and value-
added. Non-value-added consists of information that is publicly available.
This is usually objective information that describes the product sold. When
dealing with books, for example, such information can be the book’s title,
author, and genre (mystery, children’s, etc.).
Value-added consists of all the information generated by the Web store and
not publicly available. In the book example, this would include in-house book
reviews and subjective book categories (e.g., tearjerker, brain twisters). Value-
added information can be generated from the company (e.g., weekly bestseller
lists of products), a third party (e.g. , hyperlinks to on-line reviews of products
by independent parties), or the customers themselves (e.g., customer reviews
available to other customers). Whatever its source, value-added information
can influence the search for products and services by customers. External search
may be undertaken simply because it is fun (e.g., window shopping) [8]. Also,
customers often engage in ongoing search activities that involve information
gathering independent of specific needs or purchase decisions [8]. For such
customers, value-added information may prove interesting and helpful. The
existence of value-added information at a commercial Web site can be an im-
portant incentive for people to shop on-line and provides a key source of dif-
ferentiation [30].
One of the most important consumer benefits of the Web is the access it
provides to information and the concomitant opportunity for enhanced con-
sumer decision-making through complex, non-linear, non-directed queries [26].
This enhanced decision-making process is often achieved through the use of
value-added information on the products and services available. On the other
hand, many consumers engage in little external pre-purchase search even for
major items [6, 17, 41]. In such cases the abundance of value-added informa-
tion may be more of a nuisance than a help. The present study seeks to deter-
mine when and for whom value-added information can have a positive effect
on attitudes and behavior.
The second dimension of the study framework deals with the technology
used by the search mechanism. Three groups of search technologies are iden-
tified: (1) keyword-type search engines, (2) categorical classification, and (3)
other technologies, such as recommender systems and product reviews. Fig-
ure 2 shows the framework and some examples of search mechanisms in each
cell. Matching different types of search mechanisms with different types of
customers can affect their shopping experience and behavior.
Challenges
Along with individual skills, the positive challenges presented by an activity
are the most important predictors of flow [1, 11, 21, 23, 43, 60, 63]. Web skills
probably do not affect a Web customer’s experience and behavior. Customers
seeking to make purchases on the Web have already developed the basic skills
necessary for this activity. However, the positive challenges presented by a
Web-based store could be an important factor in determining the customer’s
experience. Using a Web site to purchase products can be a positively chal-
lenging activity, because customers are required to use their skills and abili-
ties in navigating the Web site, learning the interface, processing information,
and making decisions to find and buy the right products or services. The
negative challenges posed by slow download times, dead links, and incom-
plete checkout processes are not of interest in this study. It will look at how
positive challenges from shopping on the Web can affect consumer attitudes
and behavior.
Impact of Individual and Web Store Factors on Attitudinal

Variables
One expects to find that customer tenure has a significant moderating effect
on the relationship between product involvement and shopping enjoyment.
Involvement with the product(s) sold by a Web-based company affects con-
sumer attitudes and behavior. Involvement, measured simply as importance
of the Web to the consumer, also had a strong effect on the primary anteced-
ents of flow, namely skills, challenges, and focused attention [43]. A consumer
visiting a Web site for the first time is more likely to experience enjoyment
independent of involvement with the product sold. The novelty of the new
Web store, the initial exploration of its offerings, and the initial challenge of
finding what is needed could lead to the consumer’s enjoying the visit re-
gardless of personal involvement with the product. However, the more often
a customer returns to a Web store the more this involvement will determine
the experience. The higher the customer’s involvement with the product, the
more likely the consumer will be to continue experiencing shopping enjoy-
ment while visiting the Web site. One does not expect to see a relationship
between involvement and perceived control.
Technology Type
Information Type
Keyword type Categorical classification Other

Value-added Search by Gift-suggestion categories Recommender systems
keywords Top sellers
(categorical Product reviews
indexing)
Non-value- Search by Product categories New products
added keywords (full-
text indexing)
Figure 2. A Framework for Classification of Web-store Search

Mechanisms
H4: As the number of past customer visits increases, the level of customer
involvement with the product will have a stronger positive relationship
with the customer experience of shopping enjoyment.
In general, the more enriched and satisfying the shopping experience is,
the more likely customers are to experience high levels of perceived control
and enjoyment. The use of value-added search mechanisms should enrich the
shopping experience, since searching for product information can be an in-
trinsically enjoyable experience [8]. Value-added information can also improve
consumer decision-making through complex, non-linear, and non-directed
queries, and can be an important incentive for people to shop on-line by giv-
ing them more control over their purchasing activities [26, 30]. Therefore:
H5a: The use of value-added search mechanisms will be positively related

to perceived control.
H5b: The use of value-added search mechanisms will be positively related

to shopping enjoyment.
Similarly, if customers perceive their shopping experience as positively

challenging, they can be expected to have similarly positive experiences. In
the flow literature, challenges increase flow when paired with equal skills [11,
12, 13]. Since it is assumed that Web customers already possess the necessary
skills, albeit basic ones, to use a Web-based store, one expects that higher posi-
tive challenges will increase their level of perceived control and enjoyment.
H6a: The level of challenges of a Web-based store will be positively related

to perceived control.
H6b: The level of challenges of a Web-based store will be positively related

to shopping enjoyment.
Consumer need specificity influences the type of search mechanisms used

by customers. More specifically, customers with high need specificity will use
non-value-added search mechanisms more often. Since they know exactly what
they want, they would like to be able to get to it without having to go through
a lot of subjective or tangential information. Customers with low need speci-
ficity (i.e., those unsure of what they are looking for) will opt for value-added
search mechanisms that can help them better define their needs.
H7: Customer need specificity will be positively related to the use of non-
value-added search mechanisms and negatively related to the use of value-
added search mechanisms.
Study Design
The Web-based Company: Kozmo.com™
The company selected for the study was Kozmo.com™ (www.kozmo.com), a
well-known e-commerce player. At the time it was a small startup operating
only in New York City as a purely Web-based video rental and delivery store.
Customers could access its movie collection from its Web site and rent movies
on-line. Movies were physically delivered within one hour. They could be
returned at selected drop-off boxes throughout the city or could be picked up
for a nominal fee. The company’s Web site featured search mechanisms to
help customers find movies. After the completion of the study Kozmo.com™
expanded both the range of products it offered as well as the number of cities
in which it operated. Finally, however, it succumbed to the fate of many other
dot coms and shut down in the spring of 2001.
Customers’ intentions to return are influenced by many factors, including
the quality of the product, the availability of products at the store, and the
quality of the service. The study eliminated as many of these alternative ex-
planations as possible. Videos are a commodity where product quality across
stores varies only slightly. Kozmo.com™ bought its movies directly from the
distributors and therefore their quality was uniformly good. The quality of
service was consistently high during the time of this study because the com-
pany still only served a small market in New York. Finally, Kozmo.com™
only rented movies on video, eliminating any possible effects on consumer
attitudes and behavior based on different types of products.1
Sample
An on-line questionnaire was made available to the customers of Kozmo.
com™, which at the time had a customer base of about 1,300. Due to the small
customer base and the length of the survey, it was decided to make the ques-
tionnaire available to all customers rather than a random sample. This re-
sulted in self-selection among the respondents, but the advantage of a larger
response rate outweighed the disadvantages of self-selection. Since the sur-
vey was made available to every customer who reached the checkout page, it
was being provided to a very large portion of the entire population of subjects
instead of just a representative sample. As an incentive, a free video rental
was given to each customer who fully answered the questionnaire.
Survey Deployment
As they reached the checkout page, customers were presented with a banner
that read “Free Rental. Click Here.” Customers who clicked on the banner
were invited to fill out the survey for a free video rental. They were only
allowed to fill out the survey once, and had to complete the entire survey in
order to receive the free rental. The on-line survey was coded so that custom-
ers who had already filled it out could not access it again and customers who
did not fill out the whole survey could not submit it. The subjects remained
anonymous, identified only by a unique ID number.
The survey asked about the consumer’s experience during that specific
visit. This methodology encourages reports that are more valid, because data
based on recollections of past experiences can be unreliable. The Web-based
form was simple. Pre-testing showed that subjects took less than 10 minutes
to complete it and had no problems with either the user interface or question
comprehension. The questionnaire is described in the appendix.
Results
After the rejection of subjects who used Kozmo.com™ only for renting DVDs,
the sample comprised 332 responses collected in four days. There were 50
new and 282 repeat customers in the sample. A Levene’s test for equality of
variances and a t-test for equality of means on the demographic variables of
gender and age showed no differences between the two subsamples despite
the difference in size. Given that one of the dependent variables was binary
and the existence of the moderating variable of customer tenure, the complex-
ity of the model made the use of structural equation modeling (SEM) difficult.
The subsample of new customers with 50 subjects was too small for SEM [31].
Also, the use of two-item scales introduced problems of underidentification
and instability of parameter estimates [61]. Therefore, the hypotheses were
tested with separate regression models.
A binary variable for unplanned purchases was created that took the value
of 0 if customers rented as many movies as they originally intended and 1 if
they rented more. For a dichotomous dependent variable (unplanned pur-
chases), logistic regression is most useful for predicting the presence or ab-
sence of the outcome. A logistic regression was run to test whether perceived
control and shopping enjoyment had any impact on that variable. Also tested,
by including the relevant interaction terms seen in Table 1, was whether cus-
tomer tenure was a significant moderating variable in this relationship. The
logistic regression model was not significant, indicating no support for hy-
potheses H1a and H1b.
The significance of customer tenure as a moderating variable was tested by
running three regression models for the hypotheses: a pooled model that in-
corporated all subjects, a constant shift model that assumed different inter-
cepts for new and repeat customers, and a full model that assumed different
intercepts and slopes for new and repeat customers. If the pooled model had
the best fit, this would mean that tenure had no significant effect. If either the
Variable B Standard Error Wald
Constant –1.4165o 0.738

Perceived control 0.009 0.125 0.006
Shopping enjoyment 0.015 0.126 0.014
Tenure –3.156 2.828 1.246
Tenure*Perceived control –0.100 0.376 0.071
Tenure*Shopping Enjoyment 0.538 0.520 1.070
–2 Log Likelihood = 329.603
Goodness of Fit = 330.591
Cox & Snell—R2 = 0.013
Nagelkerke—R2 = 0.020
Summary of logistic regression analysis for variables predicting customer unplanned purchases for all
customers (N = 332).
Chi-Square df Significance
Model 4.321 5 0.5041

Block 4.321 5 0.5041
Step 4.321 5 0.5041
Classification table for unplanned purchases. The Cut Value is .50
Predicted Percent Correct
0 1
Observed
0 265 0 100%
1 67 0 0%
where ** p < 0.01, * p < 0.05,° p < 0.1 and Tenure = 1 for new customers and Tenure = 0 for repeat
customers
Table 1. Logistic Regression Model for Customer Unplanned Purchases

with Customer Tenure as Moderating Variable.
constant shift or the full model had a better fit with customer tenure having a
significant coefficient, this would mean that tenure had a significant moderat-
ing effect.
The best regression model was the full model (see Table 2). Perceived con-
trol is a significant predictor of intention to return for both new and return
customers (H2a). Also, customer tenure has a significant moderating effect on
the relationship (H3) because the regression line for new customers has a lower
constant and a higher coefficient than the one for repeat customers. While the
model has good fit (F = 12.979, p < 0.001), its R2 is low at 0.168. The diagnostics
show heteroscedasticity and non-normal distribution of standardized residu-
als. A look at the distribution of the variable “intention to return” (see Table 3)
shows that 96.1 percent of repeat customers indicated they would return to
the Web site. Therefore, hypotheses H2a and H2b could not be tested for re-
peat customers.
Variable B Standard Error b
Constant 5.900** 0.247

Perceived control 0.135** 0.042 0.203**
Shopping enjoyment 0.026 0.043 0.037
Tenure –2.005** 0.552 –.808**
Tenure*Perceived Control 0.328** 0.104 0.728**
Tenure*Shopping enjoyment –0.018 0.116 –0.040
R2 = 0.168, F = 12.979 (p < 0.01)
Summary of Simultaneous Regression Analysis for Variables Predicting Customer Intention to Return for All
Customers (N = 327)—Full Model.
where ** p < 0.01and Tenure = 1 for new customers and Tenure = 0 for repeat customers
Repeat customers New customers

Variable Coefficient Coefficient
Constant 5.900 3.895

Perceived control 0.135 0.463
Shopping enjoyment n.s. n.s.
Final linear regression models for repeat and new customers
Table 2. Full Regression Model for Customer Intention to Return with

Customer Tenure as Moderating Variable.
Customer Intention
to Return (Scale 1– 7) Frequency Percentage
1 7 2.5
2 2 0.7
5 2 0.7
6 25 8.9
7 246 87.2
Total 282 100.0
Table 3. Summary of Responses for Customer Intention to Return for

Repeat Customers (N = 282).
Instead, the study looked only at the new customers. Table 4 shows the
linear regression results for new customers with intention to return as the
dependent variable and perceived control and enjoyment as the independent
variables. After removing two outliers for the new customer subsample, the
results indicate support for H2a and weaker support for H2b.
Virtually all the repeat customers reported that they would return to Kozmo.
com™. This fact is perhaps to be explained by the incentive they were given
to participate. The free video rental encouraged them to return and use it.
Although a limitation of the study design, the incentive was constant for all
subjects, and therefore one would expect its effect to be relatively constant
Variable B Std. Error b

Perceived control 0.211* 0.085 0.382*
Shopping enjoyment 0.166° 0.092 0.279°
R2 = 0.359, F = 12.618 (p < 0.01)
Summary of Simultaneous Regression Analysis for Variables Predicting Customer Intention to Return for New
Customers (N = 48).
where * p < 0.05 and ° p < 0.1
Table 4. Determinants of Customer Intention to Return for New

Customers.
Expected frequency
of return visits Frequency Percentage
Never 0 0
Once per year 1 0.3
Once per six months 0 0
Once per three months 8 2.4
Once per month 48 14.5
Once per two weeks 138 41.6
Once per week or more 137 41.3
Total 332 100.0
Table 5. Expected Frequency of Return Visits for All Customers.
Constant 1.823 ** 0.494

Involvement 0.451 ** 0.067 0.342 **
Challenges 0.165 ** 0.053 0.160 **
Skills 0.051 0.052 0.050
Value-added used 0.187 ** 0.051 0.193 **
Non-value-added used -0.014 0.074 -0.010
Summary of simultaneous regression analysis for variables predicting shopping enjoyment for all customers
(N = 332)—Pooled model.
R2 = 0.200, F = 16.297 (p < 0.01) **p < 0.01
Table 6. Determinants of Shopping Enjoyment.
Need specifici ty Value-added Non-value-added
Need specificit y -0.211** 0.053

Value-added used -0.211** 0.323 **
Non-value-added used 0.053 0.323 **
where ** p < 0.01
Table 7. Correlation Between Need Specificity and Use of Search

Mechanisms.
0.6
0.5 0.483
0.415
0.4
Correlation
0.341 0.347
0.3 0.301
0.2
0.1
0
0 1-5 6-10 11-15 16+
Past Visits
Summary of simultaneous regression analysis for past visits predicting correlation of involvement and
enjoyment for all customers (N = 332).

Past visits 0.037 0.015 0.814°
R2 = 0.663, F = 5.9 (p = 0.093)
where ° p < 0.1
Figure 3. Correlation Between Involvement and Enjoyment Based on

Number of Past Visits (all correlations are at least significant at the
p < 0.05 level).
across all subjects, and thus the differential effect of the attitudinal variables
could still be measured. In order to verify that customers were not planning to
simply return once, use the free video rental, and never return again, they were
asked how frequently they expected to visit Kozmo. com™ in the future. As
Table 5 shows, the vast majority expected to return many times to the Web site.
In order to test H4 the sample was split further into five groups based on
the customers’ tenure at the Web site. For each group, the correlation between
product involvement and shopping was calculated (see Figure 3). The upward
linear trend seems to support the hypothesis. A linear regression with the cor-
relation as the dependent variable and the tenure group as the independent
variable shows that the slope is borderline significant. In other words, there is
some support for the hypothesis that as the number of past customer visits
increases, the level of customer involvement with the product will have a stron-
ger positive relationship with the customer experience of shopping enjoyment.
The pooled, constant shift, and full regression models were also run with
perceived control as the dependent variable and involvement, challenges, skills,
value-added search mechanisms used, and non-value-added search mecha-
nisms as the independent variables. All the models had low R2 values from
0.043 to 0.086 and non-significant F-statistics, indicating no support for H5a
and H6a. The same three linear regression models were run with shopping
enjoyment as the dependent variable. The pooled model had the best fit, indi-
cating that customer tenure did not have a significant moderating effect (see
Table 6). Positive challenges and use of value-added search mechanisms are
positively related to shopping enjoyment on-line (H5b, H6b). Involvement is
also a significant factor consistent with the results of H4. As expected, skills
were not a significant predictor.
Table 7 shows the correlations between need specificity and search mecha-
nism use for all customers. The results indicate partial support for H7, since
need specificity is negatively correlated with the use of value-added search
mechanisms but not significantly related to the use of non-value-added search
mechanisms. This may be because almost everyone uses non-value-added
search mechanisms anyway (such as New Releases, which was available on
the initial page of Kozmo.com™).
The final model with significant relationships can be seen in Figure 4.
Discussion of Findings
The findings in this study indicate that there are differences between the ex-
perience and behavior of new customers and repeat customers. For new cus-
tomers, enjoyment and perceived control are important factors that determine
their intention to return. However, almost all repeat customers became loyal
customers. This has serious implications in the highly competitive e-commerce
environment. Given increasing customer-acquisition costs, companies must
find the right formula of design and functionality that will increase new cus-
tomer retention and loyalty. The present study found that a Web-based com-
pany must provide new customers with both an enjoyable experience and high
levels of perceived control to entice them to return. System functionality that
increases perceptions of control and engaging design that increases perceptions
of enjoyment are important to new customers. In the case of Kozmo.com™
once new customers decided to return to the store, they became very loyal.
The results on unplanned purchases were surprising. While an enjoyable
experience was expected to lead to more unplanned purchases (rentals in this
case), and high levels of perceived control to fewer ones, no such relationship
was found. Perhaps these factors would have been important if other catego-
ries of products had been cross-sold with the movies. After the completion of
the study Kozmo. com™ increasingly began to cross-sell other, related im-
pulse goods, such as ice cream, pharmacy goods, and electronics. Further re-
search is required to understand how on-line environments can be best designed
to increase unplanned purchases within and across product categories.
The results also show that customers’ product involvement can have a sig-
nificant impact on their on-line experience. For new customers, involvement
is less important, and the novelty of the Web site can prove sufficiently inter-
esting even if the customer is not as involved in the product or service. How-
ever, for repeat customers, novelty quickly wears off, and product involvement
is more important in increasing shopping enjoyment. This relationship seems
to increase with the number of past visits to the Web store.
As for the other factors that increase the perceived control and enjoyment
experienced by customers, the results were mixed. Value-added search mecha-
Customer tenure
Need specificit y Product involvement (new vs. repeat)
–0.211** 0.342**
UNPLANNED
PURCHASES
0.193**
Value-added Shopping
search mechanism used enjoyment 0.279°
0.160**
Perceived 0.382* INTENTION TO

Positive control RETURN
challenges
– – – – – – Weak or statistically non-significant

—————— Statistically significant
where ** p < 0.01, * p < 0.05, ° p < 0.1
Figure 4. Final Model

Note: Relationships bet ween Shopping enjoyment, Perceived control, and Intention to return are
for new customers only.
nisms and positive challenges increase shopping enjoyment for all customers,
but there was no such relationship for perceived control.
These results provide important guidelines for Web stores on how to pro-
vide information that supports customer search for products. They suggest
that constructing unique content for Web sites to support customer search
and providing multiple search mechanisms that make the product search a
positive challenge will increase enjoyment. This can be a key source of service
differentiation among companies and play an important role in making new
customers return to the Web site.
As expected, the study also shows that value-added search mechanisms
are used more when customer needs are less specific. In other words, the less
certain customers are about what they are looking for, the more they will use
search mechanisms that provide value-added information. In consequence,
Web stores should be designed to be responsive to different levels of need
specificity, providing search mechanisms for customers with all kinds of needs.
This study has some limitations. First, there is the problem of self-selection.
Since the survey was not randomly assigned to subjects, the results may have
been biased toward customers who were willing to answer surveys. The sur-
vey was made available to all customers because the customer base of Kozmo.
com™ was rather small (about 1,300) and it was necessary to get enough
responses for the results to be analyzed. Since the survey was made available
to every customer who reached the checkout page, it was provided to a very
large portion of the entire population of subjects instead of just a representa-

tive sample.
A second limitation is the possibility that the results may be limited to cus-
tomers of video rental stores, and thus cannot be generalized. This is a com-
mon risk with empirical studies of single companies. For example, product
involvement may be lower, on average, for movies on video than for a more
complex product such as cars. Involvement may have a different impact on
consumer experience and behavior for other products. On the positive side,
since the product, movies on video, is a commodity, the effect of variables,
such as product quality and price, was minimized. Also minimized was the
effect of service quality, since the service of Kozmo.com™ was fairly uniform
in quality because of its centralized location in Manhattan. This makes for
confidence in the relationships between the attitudinal variables and the cus-
tomer behavior variables. Interestingly, the top-selling products on the Web
are such commodities as books, music CDs, and software.
Overall, the results of this study provide important insights into the on-line
consumer experience, with implications for system design. It is critical for
information systems researchers to understand how system design affects and
increases the efficiency of sales processes. This exploratory study develops a
model of consumer behavior that addresses such issues and tests some factors
thought to affect customer retention and unplanned purchases.
Electronic commerce empowers consumers by giving them unprecedented
information access and communications. As customer switching costs go down,
the challenges of attracting, retaining, and realizing profits from customer in-
teractions increase. Carefully designed studies that integrate marketing per-
spectives on consumer behavior and information systems research on user
behaviors will be critical to understanding consumer behaviors on-line. In-
deed, each of the major variable categories considered in the present study
(i.e., individual consumer factors, Web site design factors, consumer/user at-
titudes and decision processes, and behavior) presents opportunities for mar-
keting and information systems researchers. Further research on these topics
from a multidisciplinary perspective will be critical to the effective design of
systems that strategically influence customer behavior.
NOTE
1. Kozmo.com™ also rented movies on DVD, but customers who only wanted
to rent DVDs were not considered in the sample.
REFERENCES
1. Andrews, J.C., and Durvasula, S. Suggestions for manipulating and

measuring involvement in advertising message content. In R.H. Holman
and M.S. Solomon (eds.), Advances in Consumer Research, vol. 18. Ann Arbor,
MI: Association for Consumer Research, 1991, pp. 194–201.
2. Anonymous. Repeat customers are scarce for most sites. Internet World,
4, 30 (1998), 14.
3. Assael, H. Consumer Behavior and Marketing Action. Boston: PWS-KENT,

1992.
4. Baty, J.B., and Lee, R.M. Intershop: Enhancing the vendor/customer
dialectic in electronic shopping. Journal of Management Information Systems,
11, 4 (1995), 9–31.
5. Beatty, S.E., and Ferrell, M.E. Impulse buying: Modeling its precursors.
Journal of Retailing, 74, 2 (1998), 169–191.
6. Beatty, S.E., and Smith, S.M. External search effort: An investigation
across several product categories. Journal of Consumer Research, 14 (June
1987), 83–95.
7. Blakney, V.L., and Sekely, W. Retail attributes: Influence on shopping
mode choice behavior. Journal of Managerial Issues, 6, 1 (1994), 101–119.
8. Bloch, P.H.; Sherrell, D.; and Ridgway, N. Consumer search: An ex-
tended framework. Journal of Consumer Research, 13 (June 1986), 119–126.
9. Burke, R.R. Do you see what I see? The future of virtual shopping.
Journal of the Academy of Marketing Science, 25, 4 (1997), 352–360.
10. Clawson, P. Study: Consumers want interactive TV. Electronic Media
(August 23, 1993), 24–25.
11. Csikszentmihalyi, M. Play and intrinsic rewards. Journal of Humanistic
Psychology, 15, 3 (1975), 41–63.
12. Csikszentmihalyi, M. Beyond Boredom and Anxiety. San Francisco: Jossey-
Bass, 1997.
13. Csikszentmihalyi, M., and Csikszentmihalyi, I.S. Optimal Experience:
Psychological Studies of Flow in Consciousness. Cambridge: Cambridge
University Press, 1988.
14. Draft, H.C. The destiny of direct. Direct Marketing, 55, 8 (December
1992), 28–32.
15. East, R. Consumer Behavior: Advances and Applications in Marketing.
London: Prentice Hall, 1997.
16. Eighmey, J. Profiling user responses to commercial Web sites. Journal of
Advertising Research, 37, 3 (May–June 1997), 59–66.
17. Engel, J.F.; Blackwell, R.D.; and Miniard, P.W. Consumer Behavior. Chi-
cago: Dryden Press, 1990.
18. Forman, A.M., and Sriram, V. The depersonalization of retailing: Its
impact on the “lonely” consumer. Journal of Retailing, 67, 2 (December 1991),
226–243.
19. Gardner, E. 1998 sees fall in prices but rise in expectations (part of Design:
A Special Supplement). Internet World, 4, 30 (September 21, 1998), wd2–wd3.
20. Gefen, D., and Straub, D. The relative importance of perceived ease of
use in IS adoption: A study of e-commerce adoption. Journal of the Associa-
tion for Information Systems 1, 8 (2000) (jais.isworld.org).
21. Ghani, J.A., and Deshpande, S.P. Task characteristics and the experience
of optimal flow in human-computer interaction. Journal of Psychology, 128, 4
(1994), 381–391.
22. Ghani, J.A.; Supnick, R.; and Rooney, P. The experience of flow in
computer-mediated and in face-to-face groups. In Proceedings of the Twelfth
International Conference on Information Systems. New York, 1991, pp. 229–
237.
23. Greenwald, A.G., and Leavitt, C. Audience involvement in advertising:

Four levels. Journal of Consumer Research, 11, 1 (1984), 581–592.
24. Heijden, H. v. D.; Verhagen, T.; and Creemers, M. Predicting online
purchase behavior: Replications and tests of competing models. In Proceed-
ings of the 34th Hawaii International Conference on System Sciences. Maui, HI:
IEEE Computer Society Press, 2001.
25. Hoffman, D.L., and Novak, T.P. Marketing in hypermedia computer-
mediated environments: Conceptual foundations. Journal of Marketing, 60, 3
(1996), 50–68.
26. Hoffman, D.L.; Novak, T.P.; and Chatterjee, P. Commercial scenarios for
the Web: Opportunities and challenges. Journal of Computer-Mediated Com-
munication, 1, 3 (1995) (www.ascusc.org/jcmc/vol1/issue3/hoffman.html).
27. Howard, J.A. Consumer Behavior: Application of Theory. New York:
McGraw-Hill, 1963.
28. Howard, J.A. Consumer Behavior in Marketing Strategy. Englewood Cliffs,
NJ: Prentice Hall, 1994.
29. Inman, J.J.; McAlister, L.; and Hoyer, W.D. Promotion signal: Proxy for a
price cut? Journal of Consumer Research, 17 (June 1990), 74–81.
30. Jarvenpaa, S.L., and Todd, P.A. Consumer reactions to electronic shop-
ping on the World Wide Web. International Journal of Electronic Commerce, 1, 2
(1997), 59–88.
31. Kline, R.B. Principles and Practice of Structural Equation Modeling. New
York: Guilford Press, 1998.
32. Lohse, G.L., and Spiller, P. Electronic shopping: Quantifying the effect of
customer interfaces on traffic and sales. Communications of the ACM, 41, 7
(1998), 81–87.
33. Maignan, I., and Lukas, B.A. The nature and social uses of the Internet:
A qualitative investigation. Journal of Consumer Affairs, 31, 2 (1997), 346–371.
34. McQuarrie, E.F., and Munson, J.M. A revised product involvement
inventory: Improved usability and validity. In J.F. Sherry and B. Sternthal
(eds.), Advances in Consumer Research, vol. 19. Provo, UT: Association for
Consumer Research, 1992, pp. 108–115.
35. McQuivey, J.L.; Delhagen, K.; Levin, K.; and Kadison, M.L. Retail’s
growth spiral. Forrester Report, 1, 8 (1998).
36. Mehrabian, A., and Russel, J.A. An Approach to Environmental Psychology.
Cambridge, MA: MIT Press, 1974.
37. Miller, G.A. The magical number seven, plus or minus two: Some limits
on our capacity for processing information. Psychological Review, 63 (March
1956), 81–97.
38. Mitchell, A.A. The dimensions of advertising involvement. In K.B.
Monroe (ed. ), Advances in Consumer Research. Valdosta, GA: Association for
Consumer Research, 1981, pp. 25–30.
39. Mittal, B. A theoretical analysis of two recent measures of involvement.
In T.K. Srull (ed.), Advances in Consumer Research, vol. 16. Provo, UT: Asso-
ciation for Consumer Research, 1989, pp. 697–702.
40. Morris, B. As a favorite pastime, shopping ranks high with most Ameri-
cans. Wall Street Journal, July 30, 1987, 13.
41. Newman, J.W. Consumer external search: Amount and determinants. In
A. Woodside et al. (eds.), Consumer and Industrial Buying Behavior. New York:
Elsevier, 1977, pp. 79–84.
42. Novak, T.P.; Hoffman, D.L.; and Yung, Y.-F. Modeling the structure of the
flow experience among Web users. In Abstract for INFORMS Marketing
Science and the Internet Mini-Conference. Cambridge, MA: MIT, 1998.
43. Novak, T.P.; Hoffman, D.L.; and Yung, Y.-F. Measuring the flow con-
struct in online environments: A structural modeling approach. Working
paper. Owen Graduate School of Management, Vanderbilt University,
1998.
44. O’Keefe, R., and McEachern, T. Web-based customer decision support
systems. Communications of the ACM, 41, 3 (1998), 71–75.
45. Park, C.W., and Mittal, B. A theory of involvement in consumer behav-
ior: Problems and issues. In J.N. Sheth (ed.), Research in Consumer Behavior.
Greenwich, CT: JAI Press, 1985, pp. 201–232.
46. Pine, J.I.B.; Peppers, D.; and Rogers, M. Do you want to keep your
customers forever? Harvard Business Review, 73, 2 (1995), 692–703.
47. Quelch, J.A., and Klein, L.R. The Internet and international marketing.
Sloan Management Review, 37, 3 (spring 1996), 60–75.
48. Raman, N.V. A qualitative investigation of Web-browsing behavior. In
M. Brucks and D. MacInnis (eds.), Advances in Consumer Research. Valdosta,
GA: Association for Consumer Research, 1997, pp. 511–516.
49. Raman, N.V., and Leckenby, J.D. Advertising in interactive media: Some
issues, some suggestions. In Proceedings of the 1995 Conference of the American
Academy of Advertising, Waco, TX, 1995, pp. 247–251.
50. Ratchford, B.T.; Talukdar, D.; and Lee, M. A model of consumer choice of
the Internet as an information source. International Journal of Electronic
Commerce, 5, 3 (spring 2001), 7–21.
51. Reicheld, F.F., and Sasser, W.E., Jr. Zero defections: Quality comes to
services. Harvard Business Review, 68, 5 (1990), 105–111.
52. Rice, M. What makes users revisit a Web site? Marketing News, 31, 6
(March 7, 1997), 12.
53. Roberts, J.H., and Lattin, J.M. Consideration: Review of research and
prospects for future insights. Journal of Marketing Research, 34 (August 1997),
406–410.
54. Rook, D.W. The buying impulse. Journal of Consumer Research, 14 (Sep-
tember 1987), 189–199.
55. Rust, R.T., and Oliver, R.W. The death of advertising. Journal of Advertis-
ing, 23, 4 (1994), 71–78.
56. Shell, A. A map of the information highway. Public Relations Journal, 50
(January 1994), 27.
57. Slama, M.E., and Tashchian, A. Selected socioeconomic and demo-
graphic characteristics associated with purchasing involvement. Journal of
Marketing, 49 (winter 1985), 72–82.
58. Stern, H. The significance of impulse buying today. Journal of Marketing,
26 (April 1962), 59–62.
59. Te’eni, D., and Feldman, R. Performance and satisfaction in adaptive
Websites: An experiment on searches within a task-adapted Website. Journal
of the AIS, 2, 3 (2001) (jais.isworld.org).
60. Trevino, L.K., and Webster, J. Flow in computer-mediated communica-

tion: Electronic mail and voice mail evaluation and impacts. Communication
Research, 19, 5 (1992), 539–573.
61. Venkatesh, V., and Davis, F.D. A theoretical extension of the Technology
Acceptance Model: Four longitudinal field studies. Management Science, 46,
2 (2000), 186–204.
62. Webster, J., and Martocchio, J. J. Microcomputer playfulness: Develop-
ment of a measure with workplace implications. MIS Quarterly, 16, 2 (1992),
201–226.
63. Webster, J.; Trevino, L.K.; and Ryan, L. The dimensionality and correlates
of flow in human-computer interaction. Computers in Human Behavior, 9, 4
(1993), 411–426.
64. Zaichkowsky, J.L. Measuring the involvement construct. Journal of
Consumer Research, 12, 3 (1985), 341–352.
Appendix: The Questionnaire
Unplanned purchases: Customers were asked how many movies they intended to
rent when they visited the video store and how many movies they actually rented
on that day. The difference between the two numbers is the number of unplanned
purchases.
Need specificity: Customers selected the criteria they used in their search for a movie.
The item “I had no preferences . . .” had the lowest specificity and was given a value
of 0, and the item “I knew exactly . . .” had the highest specificity and was given the
value of 5. In all other cases, the sum of the items checked (with possible values between
1 and 4) was the measure of need specificity.
Number of past visits: Customers self-reported how many times they had visited the
video store in the past on a five-point Likert scale (ranging from “never” to “16+
times”).
Search mechanisms used: A multiple-response question was used to determine what search
mechanisms the customer used. Due to lack of resources, the study depended on self-
reports rather than Web log files to determine which search features customers used.
Shopping enjoyment (seven-point Likert scale ranging from “strongly disagree” to “strongly
agree”): In order to measure shopping enjoyment two of the items used by Ghani et al.
[1] were adapted to reflect the customer experience at the video store:
“I found my visit to Kozmo.com™interesting.”

“My visit to Kozmo. com™ was fun.”
Perceived Control (seven-point Likert scale ranging from “strongly disagree” to “strongly
agree”): Perceived control was measured with an item used by Ghani et al. [1]. It was
adapted to reflect the customer’s experience at the video store:
“During my visit to Kozmo. com™ I felt in control.”
Challenges (seven-point Likert scale ranging from “strongly disagree” to “strongly

agree”): The challenges presented by the video store’s Web site as self-reported by the
customers were measured with three items adapted from Novak, Hoffman, and Yung
[42]:
“Using the Kozmo. com™ site challenged me to perform to the best of my
ability.”
“Using the Kozmo. com™ site provided a good test of my skills.”
“I find that using the Kozmo. com™ site stretched my capabilities to the
limit.”
Involvement: The study used the ten-item Revised Personal Involvement Inventory
proposed by McQuarrie and Munson [34]. Subjects were asked to indicate their
involvement with movies on video on the following seven-point scale:
Important .................................................... Unimportant

Irrelevant ..................................................... Relevant
Means a lot to me ....................................... Means nothing to me
Unexciting ................................................... Exciting
Dull .............................................................. Neat
Matters to me .............................................. Doesn’t matter to me
Boring ..........................................................Interesting
Fun ............................................................... Not fun
Appealing ................................................... Unappealing
Of no concern to me .................................. Of concern to me
MARIOS KOUFARIS (marios_koufaris@baruch.cuny.edu) is an assistant professor of

computer information systems at Baruch College, City University of New York. He
has a Ph.D. in information systems from the Stern School of Business of New York
University, a B.Sc. in decision sciences from the Wharton School of Business, and a
B.A. in psychology from the College of Arts and Sciences, both at the University of
Pennsylvania. His research interests include electronic commerce and specifically con-
sumer behavior in Web-based commerce, hypermedia design, and socio-psychologi-
cal factors in information systems.
AJIT KAMBIL (ajit.kambil@accenture.com) is a senior research fellow and associate

partner at Accenture Institute for Strategic Change, and was previously an assistant
professor at the Stern School of Business at New York University. His research focuses
on electronic commerce, the design of networked organizations and the alignment of
IT and business strategies. He has consulted various companies on developing an
electronic commerce strategy and helps senior line and IT managers develop strate-
gies for realizing the value of information technology.
PRISCILLA ANN LABARBERA was associate professor of marketing at New York

University’s Stern School of Business from 1976 until her recent passing. She re-
ceived her M.B.A. and Ph.D. degrees from Michigan State University, and pioneered
the application of time compression to radio and television commercials, which has
since become the industry standard in the production of broadcast commercials. In
addition to her publications on advertising effectiveness, she was known for her
research in the areas of advertising self-regulation, consumer behavior and the reli-
gious subculture, social responsibility, entrepreneurship, and long-range planning/
future environments.

Consumer Behavior in Web Based Commerce An Empirical Study

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Consumer Behavior in Web Based Commerce An Empirical Study

Uploaded by

Copyright:

Available Formats

International Journal of Electronic Commerce

ISSN: 1086-4415 (Print) 1557-9301 (Online) Journal homepage: https://www.tandfonline.com/loi/mjec20

Consumer Behavior in Web-Based Commerce: An

Marios Koufaris , Ajit Kambil & Priscilla Ann Labarbera

To link to this article: https://doi.org/10.1080/10864415.2001.11044233

Published online: 23 Dec 2014.

Submit your article to this journal

Article views: 946

View related articles

Citing articles: 34 View citing articles

Full Terms & Conditions of access and use can be found at

Marios Koufaris, Ajit Kambil, and Priscilla Ann LaBarbera

consumer is undertaking. The effect is more significant when consumers are

Need Specificity Product Involvement Customer Tenure

Figure 1. Theoretical Framework for Consumer Attitudes and Behavior

In some situations on the Web, conditions that enable a consumer to save

Customer Intention to Return

1. Resource constraints: Store loyalty results from a lack of consumer

Impact of Attitudinal Variables on Consumer Behavior

H1a: Perceived control will be negatively related to unplanned purchases.

H1b: Shopping enjoyment will be positively related to unplanned purchases.

Similarly the literature shows that flow is positively related to expected

H2a: Perceived control will be positively related to consumer intention to

H2b: Shopping enjoyment will be positively related to consumer intention

Individual and Web Store Factors

H3: Customer tenure will have a significant moderating effect on the

Web Store Factors

Impact of Individual and Web Store Factors on Attitudinal

Keyword type Categorical classification Other

Figure 2. A Framework for Classification of Web-store Search

H5a: The use of value-added search mechanisms will be positively related

H5b: The use of value-added search mechanisms will be positively related

Similarly, if customers perceive their shopping experience as positively

H6a: The level of challenges of a Web-based store will be positively related

H6b: The level of challenges of a Web-based store will be positively related

Consumer need specificity influences the type of search mechanisms used

Variable B Standard Error Wald

Constant –1.4165o 0.738

Model 4.321 5 0.5041

Classification table for unplanned purchases. The Cut Value is .50

Predicted Percent Correct

Table 1. Logistic Regression Model for Customer Unplanned Purchases

Variable B Standard Error b

Constant 5.900** 0.247

Repeat customers New customers

Constant 5.900 3.895

Final linear regression models for repeat and new customers

Table 2. Full Regression Model for Customer Intention to Return with

Table 3. Summary of Responses for Customer Intention to Return for

Variable B Std. Error b

Constant 4.541 0.414

where * p < 0.05 and ° p < 0.1

Table 4. Determinants of Customer Intention to Return for New

Table 5. Expected Frequency of Return Visits for All Customers.

Variable B Std. Error b

Constant 1.823 ** 0.494

Table 6. Determinants of Shopping Enjoyment.

Need specifici ty Value-added Non-value-added

Need specificit y -0.211** 0.053

Table 7. Correlation Between Need Specificity and Use of Search

Variable B Std. Error b

Constant 0.266 0.051

Figure 3. Correlation Between Involvement and Enjoyment Based on