factors for web mining adoption of b2c firms: taiwan experience
TRANSCRIPT
Electronic Commerce Research and Applications 3 (2004) 266–279
www.elsevier.com/locate/ecra
Factors for web mining adoption of B2C firms:Taiwan experience
Echo Huang *, Tzu-Chuan Chou 1
Department of Information Management, National Kaohsiung First University of Science and Technology,
2 Juoyue Road, Nantz District, Kaohsiung 811, Taiwan, ROC
Received 18 July 2003; received in revised form 9 February 2004; accepted 6 April 2004
Available online 15 June 2004
Abstract
This paper focuses on the web mining adoption of small and medium size enterprises (SMEs) in terms of organi-
zational innovation theories. It first examines the current literature of information systems studies and suggests that the
context of web mining adoption needs to be taken into account. This paper proposes an analytical model employing a
number of internal, external factors, and the stages of web mining adoption. The model explores the relationships
influencing the stages of web mining adoption. Empirical testing is based on a sample of 68 B2C firms from Taiwanese
SMEs. The results show that firm’s Internet strategy, internationalized strategy, and business complexity along with
competitive pressure influence on the stage of web mining adoption. The implications of findings for the management of
web mining adoption and suggestions for the future study are discussed.
� 2004 Elsevier B.V. All rights reserved.
Keywords: B2C firms; Electronic commerce; Web mining
1. Introduction
Similar to conventional industries, B2C firms
not only need to be profitable to survive, but also
need to be adaptable to the turbulent business
environment. However, attracting new online
customers or retaining existing one are not easy
[10]. In this circumstance, a historical archive of
customer information and data mining techniques
* Corresponding author. Tel.: +886-7-6011000x4119.
E-mail address: [email protected] (E. Huang).1 Tel.: +886-7-6011000x4121.
1567-4223/$ - see front matter � 2004 Elsevier B.V. All rights reserv
doi:10.1016/j.elerap.2004.04.002
are valuable to overcome this problem by drawing
and analyzing information on customer behaviorand activities, and providing analysis of customer
preference. In particular, web mining in the anal-
ysis of user behavior on the Internet has been in-
creasing rapidly to understand users’ common
behavior [33]. Web mining provides greater
purchasing and customer service options by in-
corporating data warehouses and knowledge
management projects. It is one of the best strate-gies to differentiate from competitors and enables
B2C firms to discover resource, extract informa-
tion and uncover general patterns [32], resulting in
a quicker respond to their customers and, in turn,
ed.
E. Huang, T.-C. Chou / Electronic Commerce Research and Applications 3 (2004) 266–279 267
a more satisfied and loyal customer. Accordingly,
the management of adoption of web mining is now
of a critically important issue.
Although the strategic potential of web mining
is now well recognized, a number of phenomena
show more efforts are need. First, previous studiesmainly focused either on the technical perspective
or on the application development of web mining
(e.g. [6,20]), and have little efforts on the factors
that affect the process of web mining deployment.
Adoption and diffusion of innovative technologies
have remained critical concerns in IT research. IT
investment and adoption decisions are more diffi-
cult than many other investment decisions [37] andmanagement now faces a dilemma concerning the
strategic use of IT. Second, most previous research
has concentrated on large business context, ig-
noring small and medium size enterprises (SMEs).
This is understandable, since IT has been the
privilege of large businesses due to the huge in-
vestment required in the past. However, SMEs
potentially constitute the most dynamic firms in anemerging economy and are the life-blood of
modern economics [15]. The importance of SMEs
in economic growth has made them a central ele-
ment in much recent policy making to promote
and facilitate the operation of the innovation
process within SMEs [19].
Third, studies (e.g. [20]) have revealed many
reasons for data mining such as to improve cus-tomer service, to build a long-term client rela-
tionship, to reduce marketing cost, and to increase
sales. Too often, the business is surfing in pre-
vailing views enthusiastically such as the adoption
of web mining without asking: ‘‘why some orga-
nizations do not?’’ That is, they do not explain
why some organizations adopt it earlier than
others, and why some may never adopt thisinnovation.
Accordingly, the adoption behavior of web
mining has not yet been convincingly demon-
strated. This study attempts to address these con-
cerns by viewing web mining as a technological
innovation and examining factors that facilitate
their initiation, adoption, and implementation. A
research model is then proposed and tested. Bothinternal and external contextual factors of web
mining adoption are investigated. Using data from
a national wide survey, impact of internal and
external e-commerce characteristics on the initia-
tion and adoption of web mining was examined.
This empirical test of the proposed research model
is drawn on a sample of 68 Taiwanese B2C firms in
terms of the stages of adoption. The next sectiondiscusses the background and the state-of-the-art
of the development of web mining. This is fol-
lowed by a detailed description of the proposed
model and research methodology. Subsequently,
the paper presents the data analysis and hypothe-
ses testing. The final section concludes by dis-
cussing future research directions.
2. Research background
In terms of resource perspective, data captured
in different operational databases over time could
further be extracted, transported and integrated
together in data warehouses for building decision
support systems [8]. Data mining is a subset ofknowledge discovery in databases, data ware-
houses, and data marts [8]. Web mining inherits
the characteristics of data mining, but also has its
own characteristics. Web mining generally refers
to data mining on the Internet and is a technology
of the interaction between data mining and the
World-Wide Web [54]. That is, Web mining is the
application of data mining techniques to discovermeaningful patterns, profiles, and trends from
Web sites. Web mining is critical for EC due to the
large number of visitors to EC sites [34]. However,
the term Web mining is being used in two different
ways. First, web content mining is the process of
discovering information from millions of Web
documents. Second, web usage mining is the pro-
cess of analyzing Web access logs on one or moreWeb localities.
Web mining is an increasingly important and
very active research field, which adapts advanced
machine learning techniques for understanding the
complex information flow of the World-Wide Web
[28]. This is especially true since markets and
competitive structures have grown increasingly
complicated and volatile. The potential of webmining to help people navigate, search, and visu-
alize the contents of the web is enormous [32].
268 E. Huang, T.-C. Chou / Electronic Commerce Research and Applications 3 (2004) 266–279
Accordingly, web mining facilitates business to
learn more about their customers and provide
personalized offerings. A brief review of the fol-
lowing reported cases provides strong evidence for
this statement.
• Amazon has created a favor when it revealed ithas been keeping track of customer book buy-
ing behavior and has composed best-seller lists
by ZIP codes, workplace, and colleges [43].
• Real Networks has used its popular software to
monitor and report user listening behaviors
[22].
• Yahoo analyzes which advertisements and
products on its web have the most appeal so itmay alter its price and sales strategies accord-
ingly [24].
Based on Daft’s [9] dual core model, the
adoption web mining is one form of administrative
innovation in today digital economy. Fichman [14]
proposed two different scopes of IS innovations in
distinguishing the organization’s willingness/abili-
ty to adopt. The first type of innovation imposes alow knowledge burden and few new user interde-
pendencies. The organization’s willingness is the
primary determinant to adopt. In contrast, the
second type of innovation imposes a high knowl-
edge burden or high user interdependencies. In this
case, the organization’s ability is the primary de-
terminant to adopt. Swanson [46] classified IS in-
novation within an organization into three types:(1) innovations affecting mainly the organization’s
IS processes, (2) innovations affecting the business
and IS processes in the organization and (3) in-
novations additionally affecting the actual prod-
ucts made by the organization as well as
integration with other businesses. Web mining
adoption is similar to Fichman’s Type II model
and Swanson’s Type II and III models.Apparently, Internet B2C firms utilize the
massive data they collected online in a variety of
way, including tailoring products to specific cus-
tomers. For example, a lower price may be offered
in the form of a rebate coupon if we have not
purchased something recently; alternatively, a
popular item may have a higher price since the
high demand suggests this particular good is lessprice elastic. However, the innovation diffusion
process usually starts out slowly, but takes off
rapidly after an initial period Rogers [39]. The
development of web mining is still at an infant
stage. For example, different systems generate
different data, stored in separate, incompatible,
and difficult to access databases. Moreover, the
web is too unstructured for web mining [32], andcompanies may have problems to overcome these
challenges because of fragmented data. Also, it
becomes difficult enough to retrieve data for
analysis, much less to obtain a broader overview of
the processes by which a company interacts with
its agents customers, and vendors. This fragmen-
tation contributes to major business inefficiencies,
and cripples a company’s ability to detect andidentify problems, let alone resolving them. Simi-
larly, potential synergies between different product
lines were not exploited. In addition, different
market segments, each with its own regulatory
environment, require extensive information for
effectively policy management and planning. These
problems become more serious in a SMEs context
since SMEs always have problems with both for-mulating and acquiring new knowledge and skills
[50].
Several important trends redirect attention to a
more contextualized approach in the study of or-
ganizations [57]. Also, Burns and Stalker [4]
identified the importance of organizational context
to researchers have studied the many ways in
which different organizational features can facili-tate innovation. Accordingly, in order to study the
adoption behavior of web mining of SMEs, this
study employs the contextualized approach. In the
review of innovation adoption research by Fich-
man [14], two major categories, innovation pro-
cesses and innovation variance, were identified.
The first concerns with examining the process of
adoption of an innovation, and the second therelationship between innovativeness, external en-
vironment, and organizational performance. This
study investigates what factors are considered by
senior managers in terms of new technology,
market pressure, and whether there is a trend
across organizations (based on different size, age,
line of business, and strategies). Impact of two sets
of antecedent factors on the adoption of webmining is described. External characteristics con-
sidered are perceived competitive pressure and
E. Huang, T.-C. Chou / Electronic Commerce Research and Applications 3 (2004) 266–279 269
technological benefits; internal organizational
characteristics are industry type, size, age, degree
of business complexity, Internet strategy, interna-
tion strategy, and web site content strategy. The
detail of research model and hypotheses are dis-
cussed in the following section.
3. Research model and hypotheses
In planning for an IT change, managers con-
sider both external and internal forces. There is a
growing consensus among researchers that IS in-
novations in an organization differs from otherinnovations [14,46]. For example, Kwon and
Zmud [27] proposed five factors that influence
adoption such as innovation, organizational, en-
vironmental, task, and individual characteristics.
Tornatzky and Klein [48] suggested three innova-
tion characteristics influence adoption, which are
relative advantage, compatibility, and complexity.
Yang [53] proposed four external variables thatinfluence CASE tools adoption in Taiwan: per-
sonal, organizational, IS department, and system
development variables. Based on these studies,
nine contextual factors are considered in our re-
search model shown in Fig. 1.
Internal Factors•Industry-type•Firm’s Size•Firm’s age•Business Complexity•Internet Strategy•International strategy•Website content strategy
Internal Factors•Industry-type•Firm’s Size•Firm’s age•Business Complexity•Internet Strategy•International strategy•Website content strategy
External Factors•Perceived competitive pressure
•Perceived technological benefits
External Factors•Perceived competitive pressure
•Perceived technological benefits
Innovation Context of B2C Firms
Fig. 1. A model of web
3.1. Internal factors
Research in IT adoption identifies many orga-
nizational factors that influence adoption such as
top management support, size, and IT expertise.Firm size has long been an issue of IT adoption
and has been found to predict technology adop-
tion [11,52]. For example, Kuan and Chau [25]
suggest small business face substantially greater
risks in IS implementation than large business do
because of inadequate resources an limited edu-
cation about IS. Thong and Yap [47] found that
use of IT in a small business was influenced by sizeof the business and its CEO characteristics. Hence,
size is a positive factor associated with adoption
[47,49] and large companies can more easily ab-
sorb risks and cost of implementing web mining
[6]. On the other hand, technology adoption is not
only a form of innovation, but also a process of
organizational learning [52]. That is, based on
learning to learn perspective, effective adoptionskills and strategies are the basis of effective
learning from previous adoption. Apparently, an
mature firm is likely to have more IT adoption
experience than a nascent firm.
Moreover, Tornatzky and Klein [48] suggest
complexity is one of the significant factors
Web mining Adoption Stage
Web mining Adoption Stage
mining adoption.
270 E. Huang, T.-C. Chou / Electronic Commerce Research and Applications 3 (2004) 266–279
correlated with adoption across studies. Com-
plexity refers to the degree of difficulty associated
with management and maintenance of a B2C
business [39]. The complexity of business products/
locations creates greater dependence for a suc-
cessful implementation, and therefore, increasesthe preference for adoption. Hence, complexity is
positively associated with adoption. Also, Drew
[12] suggests that SMEs in different industry
sectors may adopt different strategies for
e-commence. The IT industry related firms have
more IT domain knowledge and a higher extent of
digital. Since these firms are more information
intensive, they will depend on new techniques incollecting customer information. In this circum-
stance, the study predicts IT industry related firms
are likely to adopt web mining techniques.
Hypothesis 1. (H1). The size of a company is
positively related to the adoption of web mining.
Hypothesis 2. (H2). The age of a company ispositively related to the adoption of web mining.
Hypothesis 3. (H3). The complexity of a company
is positively related to the adoption of web mining.
Hypothesis 4. (H4). The IT industry related firms
with a higher extent of digital are more likely to
adopt web mining.
The adoption of IT is really one of alignment,
and organizations that are aware of IT’s new role
have usually made efforts to incorporate IT in
their strategic thinking [13]. As Barua et al. [2]
indicate, a firm may have to invest in IT, re-
gardless of its underlying cost structure, in re-
sponse to a competitor’s investment. Thus,strategic considerations are critical to the evalua-
tion of webmining. Chou et al. [7] also suggest
that management may fail to link the strategic
purpose of IT with the organization’s strategy,
and this will lead to reduce the effectiveness of
decision-making. Three different but related
strategies, Internet strategy, international strategy
and website content strategy are particular im-portant for B2C firms to taken into account of
web mining adoption.
However, users now span the globe and repre-
sent a wide variety of economic, political, and
cultural perspectives. Firms’ innovative strategies
exploit features of the Internet and WWW, such as
real-time communication, cost-effectiveness,
ubiquity and global reach, information-richnessand multimedia capability, and a user-friendly
interface, to create new value propositions for
customers [22]. Apparently, an information-
orientated Internet strategy will concern online
customer information much more than transac-
tion-oriented Internet strategy. Thus, firms with
information-orientated Internet strategy are likely
to adopt web mining.The exploration of international marketing ac-
tivity on the Internet and the associated emergence
of the global information superhighway have a
profound effect on the conduct of international
business in the new millennium [18]. Global
strategy is the way by which a business competes in
the global market [56]. The Internet can provide
companies with a low cost gateway to globalmarkets by helping to overcome many of the
barriers to internationalization commonly experi-
enced by companies [30]. However, firms still need
more effort to understand their foreign customers
to have effective market segmentation and cus-
tomer targeting strategies. In these circumstances,
firms with the strategy toward internationalize and
provide multiple languages web content need tolearn more from their customers and therefore are
more likely to adopt web mining.
Hypothesis 5. (H5). An information-oriented In-
ternet strategy is more likely to adopt web mining.
Hypothesis 6. (H6). Firms with the strategy to-
ward internationalized are more likely to adoptweb mining.
Hypothesis 7. (H7). Firms with multiple languages
web content are more likely to adopt web mining.
3.2. External factors
In reviewing the literature, studies (e.g. [21])offer an analysis of organizational innovation in
terms of institutional theories. In terms of insti-
E. Huang, T.-C. Chou / Electronic Commerce Research and Applications 3 (2004) 266–279 271
tutional theory, survival arises out the conformity
to external rules and norms. A company’s desire to
stay ahead of its competition is a major factor in
adoption of IT [26]. That is, adoptions of new
techniques such as web mining are a form of or-
ganizational response to the external environ-mental change. Indeed, an organization must
adapt to changing environmental conditions by
altering its organizational characteristics such as
structure or process. Competition characteristics
therefore affect the adoption initiation and pro-
cess. For example, companies that dominate a
particular market tend to be leaders; they are ei-
ther responsible for IT innovations or are quick toadopt them once introduced by competitors. Thus,
a higher extent of competition will positively in-
fluence the willingness to adopt of an innovation.
This effect is particularly evident if the innovation
leads to a competitive advantage. Previous re-
search in IT adoption has shown that it has be-
come a strategic necessity to use these technologies
to compete in the market place.
Hypothesis 8. (H8). Firms perceived a higher level
of competitive pressure are more likely to adopt
web mining.
Previous researches argued that a rational
adoption decision in an organization would involve
evaluating the advantage of the new technology.Although, an application’s technical complexity
may hinder implementation success [48], several
studies have shown benefits introduced by new
technology impact its adoption. In studying EDI
adoption of small business, Kuan et al. [26] also
suggest that adopter firms perceived greater bene-
fits than non-adopter firms. Rogers [39] defined
relative advantage of an innovation as the degree towhich the innovation is perceived as better than the
idea it supersedes. Apparently, perceived benefit is
a significant variable that positively relate to
adoption of innovation. Web mining technologies
provide many benefits to the adopters in terms of
better customer service; more timely information
available for decision making, greater online in-
teraction, and a more focused marketing cam-paign. In a competitive marketplace, these benefits
are significant motivations for adoption.
Hypothesis 9. (H9). Firms perceived a higher level
of technological benefits are more likely to adopt
web mining.
4. Methodology
Investigating the issues presented above in-
volved empirical work undertaken in Taiwan B2C
firms. The constructs were operationalized in the
form of a questionnaire. Senior managers of IT or
EC Department from Top 500 companies listed in
the 2000 issue of Digital Week are selected as re-spondents. This particular audience is chosen as
these firms are generally leaders in IT applications,
and the managers are knowledgeable about their
firms’ MIS applications. Respondents were asked
to evaluate propositions based on web mining
project developed and implemented which they
had experience. A four-page survey was sent to
these IT managers. Six weeks later, 64 usableforms were returned. Repeated callbacks to non-
respondents were carried out twice. After another
six more weeks, data collection was concluded.
Four usable questionnaires were received from the
second wave of mailing, giving a total of 68 re-
spondents, and a 13.6% response rate. Such a low
response rate may be indicative that subjects were
unwilling to respond to unsolicited surveys, or thatcompanies have set a policy for rejecting survey
questionnaires.
4.1. Measurement
The variable, their operationalization and the
sources of variables are presented in Appendix A.
Seven single items with a nominal scale were usedto collect the respondents’ characteristics. Web
mining adoption stages are classified into un-
known, no adoption, initiation, adoption, and
implementation. Industry types are distinguished
by either IT or non-IT firms. Firm size is measured
by the number of employee and by total gross
sales. Complexity is classified into high volume
transactions/processing (with more than two di-verse types of business transactions) or low volume
transactions/processing (with only one type of
business transactions). Internet strategy is either
Table 1
Sample characteristics
Number of firms Percentage
Industry
e-Tailing 5 7.4
Airline 2 2.9
Traveling 1 1.5
Insurance 2 2.9
Banking 5 7.4
Information 32 47.1
Entertainment 7 10.3
Others 14 20.6
History
1–3 years 34 50
272 E. Huang, T.-C. Chou / Electronic Commerce Research and Applications 3 (2004) 266–279
transaction- or information-oriented. Similarly,
two groups are created based on variables of in-
ternational strategy and website content. Interna-
tional strategies are classified into yes/no two
groups. Website content are divided into two
group-multiple languages versions and single lan-guage version. Two external factors, perceived
competitive pressure and perceived technology
benefits, were measured using multi-item indica-
tors that aimed to capture the underlying theo-
retical domain of the constructs. Most items were
measured using a seven-point Likert-type scales
ranging from strongly disagree to strongly agree.
4–7 years 16 23.58–10 years 3 4.4
11–1 5 years 6 8.8
Over 15 years 9 13.2
Annual sales revenue
Under 1000 m 12 17.6
1001–3000 m 11 16.2
3001 m–1 billion 20 29.4
1–10 billion 14 20.6
Over 10 billion 11 16.2
Employee
Under 10 18 26.5
11–50 17 25.0
51–100 11 16.2
101–150 1 1.5
151–250 21 30.9
Table 2
B2C technology adoption
Stage Number of firms Percentages
Unknown 6 8.8
No adoption 5 7.4
Initiation 24 35.3
Adoption 5 7.4
Implementation 28 41.2
5. Data analysis and hypotheses testing
This section presents the analysis results. First,
the nature of the data is explored. Second, factor
analysis is employed to confirm the dimensionality
of two external actors. Third, the regression
analysis and ANOVA are employed in testingproposed hypotheses.
Table 1 shows more than seven industries were
collected and information industry is the largest
sector, followed by entertainment industry. The
sample also includes firms of various size and hi-
story. The number of employees of the collected
samples is all less than 250. This indicates that the
collected data is suitable for the study of SMEs[12].
Table 2 shows the current status of web mining
in our sample. The leading category is implemen-
tation (41.2%), followed by initiation (35.3%),
while the smallest category is no adoption (7.4%).
When considering the major purposes of web
mining, web mining can be used for various pur-
poses: to increase sales, product management anddevelopment, sales support arrangement, channel
relationship management, support promotion
program, support marketing research, price man-
agement, customer relationship management. The
‘‘perceived benefits’’ factor presents whether the
sample firms realize these benefits in adoption web
mining. Table 3 shows the results of companies’
incentives toward web mining. All items showsover 70% of the firms do agree (from slightly to
strongly agree) with web mining can achieve these
benefits, but these is no significant difference be-tween each benefits. One possible explanation is
that the development of web mining is still at an
infant stage, and the adoption objectives for web
mining projects are likely to be blurred. There are
two possible reasons for this phenomenon. First,
the adopter firms fail to identify investment ob-
jectives or/and adoption objectives may attract less
attention. At this point, it is not clear which reasonis correct, but, without a clear identification of the
Table 4
The characteristics of web mining adopter
Characteristics Unknown No adoption Initiation Adoption Implementation
Industry
Non-IT 4 5 9 3 14
IT 2 0 15 1 14
History
1–3 years 3 3 14 2 12
4–7 years 0 1 4 0 11
8–10 years 0 0 0 3 0
11–15 years 0 0 4 0 2
Over 15 years 3 1 2 0 3
Annual sales revenue
Under 1000 m 2 1 7 0 2
1001–3000 m 0 2 0 2 7
3001 m–1 billion 1 0 7 0 12
1–10 billion 2 1 6 1 4
Over 10 billion 1 1 4 2 3
Business complexity
Direct marketing
Yes 1 0 10 1 15
No 5 5 14 4 13
Mail order
Yes 2 1 12 3 28
No 4 4 11 2 0
Logistics
Yes 1 0 6 1 13
No 5 5 18 4 15
Entry mode
Information 5 4 17 5 7
Transaction 1 1 7 0 21
Table 3
B2C companies’ incentives of web mining
Degree of perceived benefits Strongly
disagree
Disagree Slightly
disagree
No opinion Slightly
agree
Agree Strongly
agree
Increase Sales 0 3 (4.4%) 2 (2.9%) 13 (19.1%) 15 (22.1%) 18 (26.5%) 17 (25.0%)
Product management and
development
0 2 (2.9%) 1 (1.5%) 14 (20.6%) 15 (22.1%) 19 (27.9%) 17 (25.0%)
Support salesperson
arrangement
0 2 (2.9%) 1 (1.5%) 15 (22.1%) 14 (20.6%) 20 (29.4%) 16 (23.5%)
Channel relationship
management
0 2 (2.9%) 2 (2.9%) 14 (20.6%) 14 (20.6%) 20 (29.4%) 16 (23.5%)
Support promotion program 0 2 (2.9%) 2 (2.9%) 14 (20.6%) 14 (20.6%) 20 (29.4%) 16 (23.5%)
Support marketing research 0 2 (2.9%) 2 (2.9%) 14 (20.6%) 14 (20.6%) 20 (29.4%) 16 (23.5%)
Price management 0 2 (2.9%) 2 (2.9%) 14 (20.6%) 1 (1.5%) 20 (29.4%) 16 (23.5%)
Customer relationship
management
0 2 (2.9%) 4 (5.9%) 11 (16.2%) 17 (25.0%) 22 (32.4%) 12 (17.6%)
E. Huang, T.-C. Chou / Electronic Commerce Research and Applications 3 (2004) 266–279 273
274 E. Huang, T.-C. Chou / Electronic Commerce Research and Applications 3 (2004) 266–279
values of objectives, the resulting decisions are
likely to be sub-optimal.
Characteristics of adopters in our survey are
shown in Table 4. Our research indicated that IT
and non-IT firms do not show significant differ-
ence in web mining adoption. Similarly, the size ofa company is not a significant factor. However, the
complexity of a company’s business influences the
adoption preference. Transaction-oriented com-
panies are more likely to implement data mining
for their web sites. Most information-oriented
companies tend to initiate new technology. The
industry-specific factor shows that information
and computer companies tend to be adopters incomparison with other industries. Young compa-
nies, established for less than 5 years, are more
likely to adopt new technology, which may be
explainable in terms of fewer system conversion
required. Companies engaged in mail-order busi-
ness also show a higher adoption of web mining, as
do companies with global strategies.
Table 5
Results of regression analyses predicting adoption of web
mining
Independent variables T-value p-value b
Industry type 0.099 0.921 0.046
Firm age 0.531 0.597 0.114
Firm size 1.032 0.307 0.096
Complexity )3.355 0.001�� 0.262
Internet strategy 5.049 0.000��� 0.293
International strategy 2.224 0.030� 0.274
Content information )0.706 0.483 0.265
Technological benefit )0.147 0.883 0.196
Advertisement )0.653 0.516 0.114
Transaction )0.844 0.403 0.199
Planning )1.655 0.104 0.110
Information )0.536 0.594 0.101
Globalization 2.178 0.034� 0.151
R2 ¼ 0:666, adjusted R2 ¼ 0:641
F-value (p-value) ¼ 8.129 (<0.000)
df¼ 67
Notes: �p < 0:05; ��p < 0:01; ���p < 0:001.
6. Factor analysis
Since the two construct external factors, per-
ceived competitive pressure and perceived tech-
nological benefits, combined measures from a
number of different studies, it is necessary to
confirm their validity and reliability, to ensure themeasurement is accurate. Factor analysis of multi-
item indicators can be used to evaluate if the the-
orized items for a construct converge. The stan-
dard critieria of eigen-value greater 1.0, factor
loadings greater than 0.3, and a well-explained
factor structure were used in the analysis [55]. The
reliability of the constructs was assessed using
Cronbach’s a. Results indicated all the constructshave adequate alpha value (>0.65)and are well
within the thresholds suggested by Nunnally [31].
In terms of competitive pressure, Bartlett’s test
for sphericity was highly significant (Bar-
lett’s¼ 1187.64, p < 0:00), suggesting significant
correlations among at least some of the pressures.
The measure of sampling adequacy for each item
was acceptable, and the overall measure of sam-pling adequacy was 0.77, well within the accept-
able range [17]. We used the principal components
extraction method with oblimin rotation (d ¼ 0).
We determined the number of factors on the basis
of the minimum eigenvalue (greater than one) and
the screen plot. The factor analysis revealed five
underlying dimensions, which collectively explain
67.5% of the total variance in adoption.Five factors are extracted from this factor
analysis. The first factor is advertisement, which
represents how a company perceives the value of
Internet advertising; the second is transaction,
which describes a company’s technical limitation
of order processing, inquiry response, and security
mechanism; the third, integrated planning, repre-
sents a desire to achieve a more flexible and dy-namic web site; the fourth, comprehensive
information, describes customer requirements of
accessing online order, process, shipping, return-
ing, and maintenance information; the last, global
marketing, represents a desire to develop a global
marketing strategy. In term of perceived technol-
ogy benefits, only one underlying dimension is
extracted, collectively explaining 78% of the totalvariance in initiation and adoption.
Table 6
Compare mean of external factors
Stage Benefits Advertisement Transaction Planning Information Globalization
Unknown Mean 4.5833 5.0000 5.7500 5.6667 5.2783 3.8889
SD 0.84 0.54 0.75 0.76 1.21 1.66
No adoption Mean 2.9180 5.1000 4.7500 4.4000 5.4000 2.0833
SD 1.05 2.27 1.10 1.55 1.94 1.31
Initiation Mean 5.7371 5.5208 5.8854 5.0972 5.1675 4.8611
SD 1.01 1.45 0.98 1.37 1.39 0.91
Adoption Mean 4.5180 5.4000 5.5000 5.0667 4.2000 4.2667
SD 0.99 0.74 0.77 1.1 1.04 1.81
Implementation Mean 5.3061 5.8036 5.7679 5.5833 4.8911 5.2262
SD 1.04 0.72 0.68 1.20 1.31 1.36
F-value 9.117 0.859 1.979 1.291 0.725 0.0196
p-value 0.000��� 0.494 0.108 0.253 0.578 0.000���
Note: ���p < 0:001.
E. Huang, T.-C. Chou / Electronic Commerce Research and Applications 3 (2004) 266–279 275
7. Hypothesis testing
After establishing items loaded appropriately
on their expected constructs, the hypotheses were
correlated using linear regressions. Results in
Table 5 support three hypotheses (H3, H5, and
H6), i.e., complexity, Internet strategy, and inter-
national strategy affect web mining adoption.Furthermore, Table 6, which compared the mean
on benefits and competitive pressure, supports the
last twohypotheses (H8,H9).Adopterfirmsperceive
significantlyhigherdegreeof technologybenefits and
competitive pressure than non-adopter firms.
8. Discussions and conclusions
From a sample of 68 IT managers, we examined
differences in adoption, measured by seven factors,
Table 7
Summary of the results of hypotheses testing
No. Hypothesis
HI The size of a company is positively related to the adoption
H2 The age of a company is positively related to the adoption
H3 The complexity of a company is positively related to the a
H4 The IT industry related firms with a higher extent of digita
H5 An information-oriented Internet strategy is more likely to
H6 Firms with the strategy toward internationalized are more
H7 Firms with multiple languages web content are more likely
H8 Firms perceived a higher level of competitive pressure are
H9 Firms perceived a higher level of technological benefits are
to understand the underlying internal and externalcharacteristics. For instance, we found a signifi-
cant relationship between adoption and Internet
entry strategy, but do firms with online transac-
tions result in preference for new technology?
Table 7 summarizes our findings.
Complexity was a critical variable even in B2C
firms. This is consistent with prior studies that have
demonstrated complexity is a critical factor in ITadoption and use. Concerning entry strategy,
transaction-oriented firms are more likely adopt
new technology than information-oriented firms.
Companies with online transaction applications are
more sensitive to technological benefits, and com-
petitive/operational/efficient pressure. TheANOVA
test for external factors, shown in Table 6, indicates
firms do not show significant difference in degree ofpressure for creating leading products/services.
Similarly, neither the pressure of a company to
Empirical evidence
of web mining No
of web mining No
doption of web mining Yes
l are more likely to adopt web mining No
adopt web mining Yes
likely to adopt web mining Yes
to adopt web mining No
more likely to adopt web mining Yes
more likely to adopt web mining Partially supported
Table 8
Summary of organizational innovativeness studies (1995–2001)
Study Subjects Use measure Did internal factors affect
adopt?
Did external factors
affect adopt?
Thong and Yap [47] Small businesses
CEOs
Information technology CEO innovativeness
CEO attitude
CEO IT technology
Size
Brudney et al. [3] Smaller local
governments –
Georgia cities
Computer technology Population
City services
Professionalism
Use of computer
experience
Subramanian and
Nilakanta [44]
Banking industry Banking technology Centralization
Formalization
Size
Slack
Specialization
Chenggalur-Smith and Executives Client–server technology Scope of application Market position
Duchessi [6] Mission-critical application Number of
clients
Budget
Premkumar and
Roberts [38]
Rural small
business
Video
conferencing
Top management
support
Relative
advantage
Satellite
communication
Size Compatibility
Complexity
Internet Competitive
EDI pressure
E-mail External pressure
Fax Vertical linkage
Chau and Tam [5] Senior executives Open systems Satisfaction level with
existing computing
systems
Migration cost
IT human-resource
availability
Tung et al. [49] Singapore 9
companies
Lotus notes Organization type
Kuan and Chau [25] Senior EDI Perceived direct benefits
Perceived financial cost
Perceived
technical
competence
Perceived industry
pressure
Perceived government
pressure
276 E. Huang, T.-C. Chou / Electronic Commerce Research and Applications 3 (2004) 266–279
pursue a market-oriented strategy nor the degree ofa company’s globalization affects the adoption
preference. The pressure for creating comprehensive
website information shows a significant difference in
web mining adoption. Perceived technological ben-
efits are also a significant influence on adoption.
The overall validity of the research model de-
veloped using Organizational Innovation concepts
was supported in this study. Although not all hy-
potheses were proven, the statistics measuring themodel do fit, showing the research model was sta-
tistically valid. The results showed that organiza-
tional factor has a greater influence on the decision
model than external factors in terms of partial
contribution to the total variance. This indicates
when an organization decides whether to adopt
web mining, business operation needs may be the
most important consideration. Of the three factors
Table 9
The operational definitions and sources of constructs
Dimensions Constructs Operational definitions Sources
The stage of web mining adoption 1. Unknown
2. No adoption, initiation
3. Adoption
4. Implementation
[6,16]
Internal factors Industry types 1. IT firms
2. Non-IT firms
Self-defined
Size 1. The number of employee
2. Total gross sales
[11]
Business complexity 1. High volume transactions
2. Low volume transactions
[18]
Internet strategy 1. Transaction-oriented
2. Information-oriented
[18]
International strategies 1. With international strategies
2. Without international strategies
[23]
Website content 1. Multiple languages
2. Single language
[51]
External factors Perceived competitive pressure 1. How a company perceives the value of Internet
advertising
2. A company’s technical limitation of order processing,
inquiry response, and security mechanism;
3. A desire to achieve a more flexible and dynamic web site
4. Customer requirements of accessing online order, process,
shipping, returning, and maintenance information;
5. Desire to develop a global marketing strategy
[1,18,29,35,36,41,42]
Perceived technological benefits 1. Increase sales
2. Product management and development
3. Support salesperson arrangement
4. Channel relationship management
5. Support promotion program
6. Support marketing research
7. Price management
8. Customer relationship management
[19,40,45,53]
E.Huang,T.-C
.Chou/Electro
nic
Commerce
Resea
rchandApplica
tions3(2004)266–279
277
278 E. Huang, T.-C. Chou / Electronic Commerce Research and Applications 3 (2004) 266–279
found to be significant in affecting adoption, in-
ternational strategy is most similar to organiza-
tional slack and/or size in Roger’s framework. The
Internet entry strategy identifies with complexity in
Roger’s framework. The competitive pressure
in Roger’s framework is still valid in this study.Table 8 gives an overview of the Organizational
Innovations studies. Our findings correspond with
Chengalur-Smith and Duchessi [6], Premkumar
and Roberts [38], and Kuan and Chau [25].
This study was conducted to explore factors
influencing the intention of B2C firms to adopt
web mining technology. As such, there is room for
further investigation. The following are some ad-vices for future studies. First, future studies should
investigate adopter performance and effectiveness.
Second, as web mining technology and customer
relationship management concepts are still rela-
tively new in Taiwan, this study has been unable to
measure the adopting behavior of such technolo-
gies, which has been suggested by Organizational
Innovativeness theory. Future studies should in-corporate this measure once the diffusion has
reached a reasonable base. This way, more com-
prehensive knowledge of web mining and customer
relationship management adoption behavior can
be understood. Third, the study of web mining
adoption can be extended to e-commerce firms.
Comparison can then be made between B2C and
B2B firms in terms of underlying factors that affecttheir adoption decisions.
Appendix A
See Table 9.
References
[1] M.C. Angelides, Implementing the Internet for business: a
global marketing opportunity, Int. J. Inform. Manage. 17
(6) (1997) 405–420.
[2] A. Barua, C. Kriebel, T. Mukhopadhyay, Information
technologies and business value: an analytic and empirical
investigation, Inform. Syst. Res. 6 (1) (1995) 3–23.
[3] J.L. Brudney, S.C. Selden, The adoption of innovation by
smaller local governments, Am. Rev. Public Adm. 25 (1)
(1995) 71–87.
[4] T. Burns, G. Stalker, The Management of Innovation,
London, 1961.
[5] P.Y.K. Chau, K.Y. Tam, Information technologies and
business value: an analytic and empirical investigation,
Inform. Manage. 37 (2000) 229–239.
[6] I.N. Chengalur-Smith, P.J. Duchessi, The initiation and
adoption of client-server technology in organizations,
Inform. Manage. 35 (1999) 77–88.
[7] T.C. Chou, R.G. Dyson, P.L. Powell, Managing strategic
IT investment decisions: form IT investment intensity to
effectiveness, Inform. Resour. Manage. J. 3 (4) (2000) 34–
43.
[8] S. Chowdhury, Databases, data mining and beyond, J.
Am. Acad. Bus. 2 (2) (2003) 576–580.
[9] R.L. Daft, A dual core model of organizational innova-
tion, Acad. Manage. J. 21 (1978) 193–210.
[10] S. Devaraj, M. Fan, R. Kohli, Antecedents of B2C channel
satisfaction and preference: validating e-commerce metrics,
Inform. Syst. Res. 13 (3) (2002) 316–333.
[11] R.D. Dewar, J.E. Dutton, The adoption of radical and
incremental innovations: an empirical analysis, Manage.
Sci. 32 (11) (1986) 1422–1433.
[12] S. Drew, Strategic use of e-commerce by SMEs in the East
of England, Eur. Manage. J. 21 (1) (2003) 79–88.
[13] B. Farbey, F. Land, D. Targett, How to Assess Your IT
Investment: A study of Methods and Practice, Butter-
worth-Heinemann, Oxford, 1993.
[14] R.G. Fichman, in: Proceedings of the Thirteenth Interna-
tional Conference on Information Systems (ICIS), Dallas,
1992, pp. 195–206.
[15] A. Ghobadian, D. Gallear, Total quality management in
SMEs’, Omega 24 (1) (1996) 83–106.
[16] Varun Grover, Martin D. Goslar, The initiation, adoption,
and implementation of telecommunications technologies in
US organizations, J. Manage. Inform. Syst. 10 (1) (1993)
141–163.
[17] J.F. Hair, Multivariate Data Analysis with Readings, third
ed., Macmillian Publishing Co., New York, 1992.
[18] J. Hamill, The Internet and international marketing, Int.
Market. Rev. 14 (5) (1997) 300–323.
[19] K. Hoffman, M. Parejo, J. Bessant, L. Perren, Small firms,
R&D, technology and innovation in the UK: a literature
review, Technovation 18 (1) (1998) 39–55, 72–73.
[20] S.C. Hui, G. Jha, Data mining for customer service
support, Inform. Manage. 38 (2000) 1–13.
[21] O. Jones, T. Edwards, M. Beckinsale, Technology man-
agement in a mature firm: structuration theory and the
innovation process, Technol. Anal. Strateg. Manage. 12 (2)
(2000) 445–464.
[22] D.R. Kalakota, M. Robinson, e-Business 2.0: Roadmap
for Success, Addison-Wesley, Boston, 2001.
[23] L.R. Klein, J.A. Quelch, Business-to-business market
making on the Internet, Int. Market. Rev. 14 (5) (1997)
345.
[24] J.V. Koch, R.J. Cebula, Price, quality, and service on the
Internet: sense and nonsense, Contemporary Economic
Policy Huntington Beach 20 (1) (2002) p–p25.
[25] K.K.Y. Kuan, P.Y.K. Chau, A perception-based model for
EDI adoption in small business using a technology-
E. Huang, T.-C. Chou / Electronic Commerce Research and Applications 3 (2004) 266–279 279
organization-environment framework, Inform. Manage. 44
(2001) 507–521.
[26] K.K.Y. Kuan, P.Y.K. Chau, A perception-based model for
EDI adoption in small businesses using a technology-
organization-environment framework, Inform. Manage. 38
(2000) 507–521.
[27] T.H. Kwon, R.W. Zmud, Unifying the fragment models
of information systems implementations, in: Critical Issues
in Information Systems Research, 1987, pp. 227–252.
[28] J. Larsen, L.K. Hansen, A.S. Have, T. Christiansen, T.
Kolenda, Webmining: learning from the world wide web,
Comput. Stat. Data Anal. 38 (2002) 517–532.
[29] D.E. Leidner, S.A. Carlsson, J. Elam, M. Corrales,
Mexican and Swedish managers’ perceptions of the impact
of EIS on organizational intelligence, decision making, and
structure, Dec. Sci. 30 (3) (1999) 633–659.
[30] S. Li, B.J. Davies, GloStra – a hybrid system for
developing global strategy and associated Internet strategy,
Ind. Manage. Data Syst. 101 (3) (2001) 132–140.
[31] J.C. Nunnally, Psychometric Theory, McGraw-Hill, New
York, 1978.
[32] E. Oren, The world wide web: quagmire or gold mine?,
Commun. ACM 39 (11) (1996) 65–68.
[33] G. Paliouras, C. Papatheodorou, V. Karkaletsis, C.D.
Spyropoulos, Discovering user communities on the Inter-
net using unsupervised machine learning techniques, In-
teract. Comput. 14 (2002) 761–791.
[34] Ismail Parsa, Web-Mining: New Data Tools to Manage
Web Strategy, 1999.
[35] P.T. Peterson, Doing business in a web-based world,
Mortgage Bank. 58 (2) (1997) 58–65.
[36] S. Poon, P.M.C. Swatman, Small business use of the
Internet findings from Australian case studies, Int. Market.
Rev. 14 (5) (1997) 385.
[37] P. Powell, Causality in the alignment of information
technology and business strategy, J. Strateg. Inform. Syst.
2 (4) (1993) 320–334.
[38] G. Premkumar, M. Roberts, Adoption of new information
technologies in rural small businesses, Omega (27) (1999)
467–484.
[39] E.M. Rogers, Diffusion of Innovations, The Free Press,
New York, 1983.
[40] M.L. Roberts, Expanding the role of the direct marketing
database, J. Interact. Market. 11 (4) (1997) 26–36.
[41] E. Schwartz, Internet apps hosting draws industry giants,
InfoWorld 21 (40) (1999) 8.
[42] P. Seybold, How to succeed in E-business, Computerworld
32 (45) (1998) 85–87.
[43] David Streitfeld, Amazon.com’s Data-mining Technology
Stirs Internet Privacy Controversy, The Washington Post,
August 28, 1999.
[44] A. Subramanian, S. Nilakanta, Organizational innovative-
ness: exploring the relationship between organizational
determinants of innovation, types of innovations, and
measures of organizational performance, Omega 24 (6)
(1996) 631–647.
[45] Tae Kyung Sung, Namsik Chang, Gunhee Lee, Dynamics
of modeling in data mining: interpretive approach to
bankruptcy prediction, J. Manage. Inform. Syst. 16 (1)
(1999) 63–86.
[46] E.B. Swanson, Information systems innovation among
organizations, Manage. Sci. 40 (9) (1994) 1069–1092.
[47] J. Thong, C.S. Yap, CEO characteristics, organizational
characterisitcs and information technology adoption in
small businesses, Omega, Int. J. Manage. Sci. 23 (4) (1995)
429–442.
[48] L.G. Tornatzky, K.J. Klein, Innovation characteristics and
innovation adoption-implementation: a meta-analysis of
findings, IEEE Trans. Eng. Manage. 0 (1982) 28–45.
[49] L.L. Tung, J.H. Tan, J.P.L. Er, K. Lian, E. Turban,
Adoption implementation and use of lotus notes in
Singapore, Int. J. Inform. Manage. 20 (2000) 369–382.
[50] J. Vos, J. Keizer, J. Halman, Diagnosing constraints in
knowledge of SMEs, Technol. Forecast. Soc. Change 58 (3)
(1998) 227–239.
[51] A. Walker, Out of site, Small Business News 3 (5) (1997)
25.
[52] J. Woiceshyn, Technology adoption: organizational
learning in oil firms, Organ. Stud. 21 (1) (2000) 1095–
1118.
[53] D. Worhach, Mining the data in the mail, Public Utilities
Fortnightly (1998) 48–54.
[54] Heng-Li Yang, Adoption and implementation of CASE
tools in Taiwan, Inform. Manage. 35 (1999) 89–112.
[55] O.R. Za€ıane, Building virtual web views, Data Knowl.
Eng. 39 (2001) 143–163.
[56] R.E. Zeller, D.D. Achabal, L.A. Brown, Market penetra-
tion and locational conflict in franchise systems, Decision
Sci. Atlanta 11 (1) (1980) 58.
[57] S. Zou, S.T. Cavusgil, Global strategy: a review and an
integrated conceptual framework, Eur. J. Market. 30 (1)
(1996) 52–69.