You are on page 1of 7

Proceedings of the 42nd Hawaii International Conference on System Sciences - 2009

The Continuous Service Usage Intention in the Web Analytics Services

Jaesung Park Kyungho Jung Yunhee Lee


Chonnam national univ. Chonnam national univ. Chonnam national univ.
pamto@hanmail.net haveagoodday@nate.com dydduwk@lycos.co.kr

Geon Cho JaeJon Kim Joon Koh


Chonnam national univ. Chonnam national univ. Chonnam national univ.
gcho@chonnam.ac.kr jaejon@chonnam.ac.kr kjoon@chonnam.ac.kr

Abstract with this growth. We expect that the growth of the E-


market will continue. The Gartner Group (2001) re-
The World Wide Web (WWW) has continued to ports that the worldwide B2B e-commerce market size
grow at very rapid speed in both the sheer volume of would increase from $433 billion in 2000 to $8,500
traffic and size, complexity of web sites. Web analyt- billion in 2005. As the B2B e-market has been rapidly
ics industry also has been growing rapidly. Web ana- expanding, many researchers have conducted their
lytics is to analyze web log files to discover accessing researches on B2B e-market related issues. Sculley &
patterns of web pages. In our work described in this Woods (2000) state that the unique feature of a B2B
paper, we identify factors which can affect the conti- exchange is that it brings many buyers and sellers to-
nuous usage intention of a firm using services in Web gether in one central virtual market space and enables
Analytics service and empirically validate the relation- them to buy and sell from each other at a dynamic
ships between the identified factors. For this purpose, price that is determined by the exchange rules. Kaplan
we analyze 174 Korea firms. The analysis results show & Sawhney (2000) define the landscape for B2B mar-
that the satisfaction is significantly associated with ketplaces along two key characteristics: product use
service quality and switching cost and the service and procurement method. Lucking-Reiley & Spulber
usage period is not significantly associated with conti- (2001) identify three types of B2B exchanges. Reuters
nuous service usage intention. We measure service (2003) estimates that the market for Web analytics is
quality using SERVQUAL. It turn out that two dimen- worth $520million in 2001 but will quickly grow to
sions of SERVQUAL, reliability and empathy are sig- over $4billion by 2005. A Web analytics services is
nificantly associated with satisfaction, but another service of data mining techniques to large web-data
dimension of SERVQUAL, responsibility, is not. Final- repositories in order to produce results that can be used
ly, satisfaction is significantly associated with conti- in management, web-design, and so on. A web analyt-
nuous service usage intention. ics service is about to enter on growth period and
many service providers are to face some stiff competi-
tion from its competitor. So, many service providers
have trouble in holding their customers. because there
is no barrier in changing partner A ratio of changing a
1. Introduction service provider is high.
The purpose of this study is to identify factors
The World Wide Web (WWW) has continued to grow which may affect the continuous usage intention of the
at very rapid speed in both the sheer volume of traffic firms using web analytics services and examine the
and size and complexity of Web sites. Also, E- relationship among the identified factors.
business model (B2B, B2C) based on World Wide
Web has been changed and grown. As the technology
2. Theoretical Background
of WWW has improved, the complexities of tasks such
as Web site design, Web server design, and of simply
navigating through a Web site have increased along 2.1 Web analytics

978-0-7695-3450-3/09 $25.00 © 2009 IEEE 1


Authorized licensed use limited to: K J Somaiya College of Engineering. Downloaded on July 29,2010 at 07:16:36 UTC from IEEE Xplore. Restrictions apply.
Proceedings of the 42nd Hawaii International Conference on System Sciences - 2009

According to the Wikipedia, Web analytics is the ment from marketing (Parasuraman et al., 1985; 1988)
study of online behavior in order to improve it. There to operationalize the IS service quality construct. They
are two categories; off-site and on-site web analytics. propose modification of the wording of the instrument
Off-site web analytics refers to web measurement and to better accommodate its use in the context. They
analysis irrespective of whether you own or maintain a conclude that SERVQUAL is proper to assess service
website. On-site web analytics measure a visitor's jour- quality.
ney once on your website. Web analytics is to reveal Kettinger & Lee (1997), for IS services, suggest
the knowledge hidden in the log files on one or more four dimensions of SERVQUAL, excluding tangibles
Websites. The goal is to capture, model, and analyze
out of the original five dimensions proposed by Pitt et
the behavioral patterns and profiles of users interacting
al.(1995) and Parasuraman et al. (1985; 1988) since
with a Website. Web analytics use on the secondary
web data such as Web server access logs, proxy server most IS services are requested, not by a visit to the
logs, browser logs, user profiles registration data, user department of IS, but rather by sending mail, phoning,
sessions or transactions, cookies, user queries, book- or by visiting a Web site. As a result, the tangibles of
mark data, mouse clicks and scrolls and any other gen- an IS department is not important to the user of infor-
erated by the interaction between users and the web. mation. Service quality of transactions, likewise, in in
The discovered patterns are usually represented as col- the web analytics services can be measured by four
lections of pages, objects, or resources that are fre- dimensions of SERVQUAL: (1) reliability, (2) respon-
quently accessed by groups of users with common sibility, (3) assurance, and (4) empathy.
needs or interest. Although service quality and satisfaction are closely
Up to now, many researchers have taken an interest correlated, their concepts may be clearly separated
in about web analytics. Cooley et al. (1997; 2000) do (Bitner et al., 1994). Cronin & Taylor (1992) validate
in-depth research to all the procedure of web usage that service quality is a preceding variable of satisfac-
mining. They discuss methods to pre-process the user tion from their research. As a consequence, it is possi-
log data and to separate web page references into those ble that service quality is measured by SERVQUAL,
made for navigational purposes and those made for
which has an affirmative effect on satisfaction.
content purposes. Ramli (2005) explores the use of
Web usage mining techniques to analyze web log
H-1: Service quality has an effect on service client
records collected from e-learning portal using apriori
algorithm. Drott (1998) explains the various Web serv- firm˅s satisfaction.
er logs mining methods that could be used to improve
site design. Sarukkai(2000) has discussed about link H-1a: Reliability has an effect on a service client
prediction and path analysis for better user navigations. firm˅s satisfaction.
He propose a Markov chain model to predict the user H-1b: Responsibility has an effect on service
access pattern based on the user access logs previously client firm˅s satisfaction.
collected. H-1c: Empathy has an effect on service client
firm˅s satisfaction.
2.2 Service Quality and Satisfaction H-1d: Assurance has an effect on service client
firm˅s satisfaction.
Based on an extensive series of focus group inter-
views, Parasuraman et al. (1985) argue that service
quality is founded on a comparison between what the 2.3 Satisfaction and Continuous Service Usage
customer feels should be offered and what is provided. Intention
Other marketing researchers also support the notion
that service quality is the discrepancy between custom- With respect to the cost and benefit by exit, the de-
ers' perceptions and expectations. Parasuraman and his pendency of inter-firms is related to the expected value
colleagues (Parasuraman et al., 1985; 1988) assert that which is served by the transaction partner (Lim et al.,
service quality could be assessed by measuring cus- 1995). The expected value is determined by competen-
tomers' expectations and perceptions of performance cy advantages like service quality or product quality
levels for a range of service attributes. They suggest through a comparison of existing and new channels.
SERVQUAL with five dimensions (tangibles, reliabili- Competency is defined as the particular skills and re-
ty, responsiveness, assurance, and empathy) that are sources a firm possesses, and the superior way in
which they are used (Reed & Defilippi, 1990).
used by customers when evaluating service quality,
Rusbult et al. (1982; 1988) suggest that high satis-
regardless of the type of service. In addition, Pitt et al.
faction and investment not only encourage voice and
(1995) proposed the use of the SERVQUAL instru-

2
Authorized licensed use limited to: K J Somaiya College of Engineering. Downloaded on July 29,2010 at 07:16:36 UTC from IEEE Xplore. Restrictions apply.
Proceedings of the 42nd Hawaii International Conference on System Sciences - 2009

loyalty but also discourage exit and neglect. Withey & H-3: Service client firm's service usage period has an
Cooper (1989) propose that prior satisfaction is posi- effect on service client firm's continuous
tively related to voice but it is inversely related to exit. usage intention.
Ping (1993) argues that the more satisfied the more
reduced the exit. He shows that overall satisfaction is 2.5 Switching Cost and Continuous Service Usage
negatively related to exit. Thus, we can hypothesize Intention
that overall satisfaction will be positively associated
with continuous usage intention. Switching cost which occurs by changing an existing
transaction firm is one of the important factors that can
H-2: Service client firm's satisfaction has an effect on explain dependency (Weiss & Anderson, 1992).
service client firm's continuous usage intention. Switching cost includes cost of retraining personnel,
capital requirements for changeover, and costs of ac-
2.4 The Service Usage Period and Continuous quiring new ancillary equipment (Porter, 1985). If the
Service Usage Intention switching cost is, the service client firm's dependency
on its service provider will be increased.
With regard to transaction cost analysis, the switching Also, the switching cost is changed by product attri-
cost for the changing of an existing partner may be in bution, the consumer's characteristic, and the firm's
proportion to the scale of transaction-specific assets in strategy. Fornell (1992) claims that the switching cost
the view of service clients. If service client firms, in- consists of search costs, transaction costs, learning
vest their transaction-specific assets on the firm for a costs, loyal customer discounts, customer habits, the
transaction to change an existing service firm, this will emotional cost, and the cognitive effort. These are
result in losing transaction-specific assets. Thus, de- coupled with financial, social, and psychological risks
pendency on an existing service firm is related to ser- on the part of the buyer. Jones (1998) note that the
vice client firm's assets which are specialized to the switching cost has its sub-dimensions of continuality
transaction. If changing an existing transaction partner costs, contractual costs, learning costs, search costs,
increases switching cost, the service client firm's de- setup costs, and sunk costs. His further research with
pendency on the partner (or immersion of relation- his colleagues (2002), also propose the following six
ships) will be heightened. dimensions of switching costs: lost performance costs,
Transaction-specific assets, mainly, are determined by uncertainty costs, pre-switching search and evaluation
education for salespersons or know-how which is spe- costs, post-switching behavioral and cognitive costs,
cialized to sale. Also, if the service usage period is setup costs, and sunk costs.
long, the knowledge and experience of the firm will be
H-4: The service client firm's switching cost has an
accumulated. Thus, the service usage period in the
effect on service client firm's continuous usage
Web Analytics Services will increase transaction-
intention.
specific assets.

[Figure 1] Research Model

3
Authorized licensed use limited to: K J Somaiya College of Engineering. Downloaded on July 29,2010 at 07:16:36 UTC from IEEE Xplore. Restrictions apply.
Proceedings of the 42nd Hawaii International Conference on System Sciences - 2009

[Table 1] Operational Definitions and Measures of the Variables


Variable Operational definition Source

Caring, individualized attention the service provider gives its custom-


Empathy
ers
Parasuraman et al.
(1995,1998)
Service Reliability Ability to perform the promised service dependably and accurately
Pitt et al.
Quality
Knowledge and courtesy of employees and their ability to inspire trust (1995,1998)
Assurance
and confidence Kettinger & Lee
Respon- (1997)
Willingness to help customers and provide prompt service
siveness

Swit- Jones(1998)
Switching cost which occurs by changing of an service provider
ching cost Jones et al(2002)

transform into natural


Service usage period Period of using service
logarithm

Satisfaction Degree of sufficiency for using service compared to expectation Oliver (1997)

Oliver (1997)
Continuous service usage
Intention through the service satisfaction McDougall & Levesque
intention
(2000)

to send mail, to phone, or to visit Web site, we meas-


3. Research Model ured the four dimensions of SERVQUAL (Kettinger &
Lee, 1997): (1) reliability, (2) responsibility, (3) assur-
ance, and (4) empathy. Satisfaction items were devel-
3.1 Operationalization of the Research Variables
oped based on Rusbult et al. (1982, 1988). The service
usage period was transformed to measure by natural
Operational definitions and measurement of the re-
logarithm. The measurement of continuous usage in-
search variables are provided in Table 1. We developed
tention was developed based on Oliver's (1997) as well
outcome variables based on relevant literature. The
as McDougall & Levesque's (2000).
variables were measured by a seven-point Likert scale
(1=strongly disagree, 7=strongly agree). Switching
cost items were developed based on marketing research
4 Research Method
such as Jones (1998). Measurement of service quality
4.1 Data Collection
came from studies such as Pitt et al. (1995; 1997;
1998) and Parasuraman et al. (1985; 1988). Since IS The instruments of the study were developed based
service involved not to visit IS department but instead on the relevant literature and the results of prior inter-

[Table 2] Sample Characteristics


Measure Frequency Percentage
under $1million 29 16.67
from $1million to $5million 54 31.03
Sales
from 5million to $10million 72 41.38
over $10million 19 10.92
CEO 31 17.82
CIO 60 34.48
Respondent's Position
Department Manager 59 33.91
Department Chief 24 13.79
Total 174 100%

4
Authorized licensed use limited to: K J Somaiya College of Engineering. Downloaded on July 29,2010 at 07:16:36 UTC from IEEE Xplore. Restrictions apply.
Proceedings of the 42nd Hawaii International Conference on System Sciences - 2009

[Table 3] Factor Analysis Results for Independent Variables


Component Cronbach's Į
1 2 3 4 5
REL2 .774 .237 .157 .296 .120 .890
REL1 .697 .312 .106 .082 .135
REL3 .691 .269 .180 .248 .048
ASU2 .666 .328 .115 .352 .274
ASU1 .638 .431 .103 .266 .256
RES2 .216 .877 .094 .157 .100 .900
RES1 .278 .787 .039 .186 .147
RES3 .329 .759 .135 .099 .241
RES4 .309 .734 .146 .176 .123
SWC2 .263 .074 .778 .115 .065 .846
SWC3 .091 .058 .772 .030 .117
SWC4 -.092 .140 .772 .070 -.005
SWC1 .265 .034 .763 .042 .133
SWC5 .052 .074 .740 .135 .137
SAT2 .134 .100 .101 .819 .067 .851
SAT3 .334 .210 .114 .788 .083
SAT1 .465 .234 .149 .692 .170
EMP4 .067 .106 .146 .022 .836 .795
EMP5 .432 .230 .192 .153 .694
EMP3 .202 .406 .182 .415 .547
EMP2 .158 .325 .087 .431 .521
Eigenvalue 9.031 2.481 1.361 1.131 .922
Percentage of Variance Ex-
43.005 11.815 6.480 5.385 4.389
plained
Cumulative Percentage 43.005 54.820 61.300 66.685 71.075
views with the leaders of the Web Analytics Company considered satisfactory for measures (Nunally, 1978).
in Korea. A pilot test was conducted with 25 firms Reliability and assurance which were expected to have
used by the Web Analytics service provider, “Ace- each factor of service quality were grouped into a sin-
counter” in Korea. We installed a Web-based ques- gular factor by a factor analysis. We, then, renamed the
tionnaire system within the Acecounter.com site in factor as Reliability.
December 2006. In total, 203 firms participated in the Each variable was measured by multiple items, so a
Web-survey during a three weeks’ period. We, how- factor analysis was conducted to check their uninten
ever, discarded 29 cases with a lack of consistency in tionality. Most of the factor loadings for the items
the answers of the companies, and 174 cases were used appeared above 0.6. The items were collected well
for the analysis. corresponding to each singular factor, demonstrating a
high convergent validity. As the factor loadings for a
4.2 Analysis and Testing variable (factor) were bigger than the factor loadings
for the other variables, it supported the instrument's
4.2.1 Reliability and Validity discriminant validity (Chin, 1998).
The eigenvalues of the each factor were all greater
The internal consistency reliability of the variables than 1.000 and the five derived factors explained were
was assessed by computing Cronbach's alphas. The above 70 percent for the total variance. From this re-
Cronbach's alpha values of all the variables, ranging sult, the reliability and validity of research variables
from 0.795 to 0.900, were well over 0.700, which was were all acceptable.

5
Authorized licensed use limited to: K J Somaiya College of Engineering. Downloaded on July 29,2010 at 07:16:36 UTC from IEEE Xplore. Restrictions apply.
Proceedings of the 42nd Hawaii International Conference on System Sciences - 2009

nuous service usage intention. When there are multiple


4.2.2 Hypothesis Testing service providing firms or when the service of the ex-
isting partner is unsatisfied, as it were, service client
Table 4 shows the results of the multiple regression firms may actually consider “exit.”
analysis, testing the six hypotheses. The results indi- Second, the switching cost was not related to conti-
cate that the two regression models were significant at nuous service intention. The reason for this is not hard
the p<0.01 level (F-value=49.420, 60.781). Also, the to see: it is because switching cost which may cause
predictors of each model explained 45% and 50% of sunk cost or continuality cost is not important for ser-
the total variance, respectively. In Hypothesis H-1, vice clients, implying that in Korea, Web Analytics
reliability and empathy were significantly related to service markets are not mature yet. However, we might
satisfaction (ȕ=0.525, p<0.01; ȕ=0.197, p<0.01). find a clue for interpretation in that, generally, there is
However, responsibility was not significantly re- little investment on the transaction-specific asset be-
lated to satisfaction. Hypothesis H-1b was not one of tween two firms in the case of Web Analytics Services.
the supported hypotheses, while both H-1a and H-1c In addition, the service usage period was not signifi-
were supported. In Hypothesis H-2, satisfaction was cantly related to continuous service intention. The
significantly related to continuous usage intention main reason may be in the basis that the service usage
(ȕ=0.677, p<0.05), but the service usage period and period of the firms which participated in this survey
switching cost was not significant. Therefore, neither was not more than two years. A maximum of two years
Hypothesis H-3 nor H-4 was supported.  may be not enough to affect continuous service inten-
tion.
5. Discussion and Conclusion Third, the two dimensions of service quality, relia-
bility and empathy, were significantly associated with
5.1 Discussion and Implications satisfaction while responsibility was not. Responsibili-
ty may be a mandatory or a prior condition in Web
 Analytics Services rather than service quality itself.
The purpose of this study is to understand the conti-
nuous service usage intention in Web Analytics Ser- 5.2 Study Limitations and Future Research Direc-
vices. To predict continuous usage intention of the tions
client firms, we adopted the three variables, satisfac- In spite of the research implications, there are several
tion and Service usage period, switching costs which limitations in this study.
have been considered as important in the areas of or- First, since the survey participant firms’ service
ganization, strategy, and marketing. Results and impli- usage period is relatively short (average is two years),
cations of the findings are: the service usage period may not be related to depen-
First, a continuous usage intention in web analytics dency. Thus, overcoming the limitation, future research
services was significantly and positively associated should include firms with long periods of using servic-
with satisfaction. This implies that web analytics ser- es.
vices, satisfaction would be an important factor. Im- Third, the cross-sectional data collection is limited
proving a service quality of web analytics helps service to our findings: For example, satisfaction is generally
providers to holding their customers. created and accumulated over a long period of time.
Satisfaction is significantly associated with conti- Therefore, future research is needed in the form of a
[Table 4] Results of Hypotheses Tests
Standardized
Model R2 adj. R2 F Result
coefficient (ȕ)
(1) Satisfaction (SAT)
SAT = REL+RES+EMP+errors 0.466 0.456 49.420***
REL 0.525*** O (H1a)
RES 0.021 X (H1b)
EMP 0.197** O (H1c)

(2) Continuous Usage Intention (CUI)


CUI = SAT+SWC+SUP+errors
0.518 0.509 60.784***
SAT
0.677*** O(H2)
SWC
0.105 X(H3)
SUP
-0.014 X(H4)
** p < 0.05, ***p < 0.01

6
Authorized licensed use limited to: K J Somaiya College of Engineering. Downloaded on July 29,2010 at 07:16:36 UTC from IEEE Xplore. Restrictions apply.
Proceedings of the 42nd Hawaii International Conference on System Sciences - 2009

longitude study. Oliver, R.L. (1993). A conceptual Model of Service Quality


and Service Satisfaction: Compatible Goals, Different Con-
cepts", in Swartz, T.A., Bowen, D.E., Brown, S.W. (Eds),
Advances in Services Marketing and Management: Research
and Practice, 2nd ed., JAI Press, Greenwich, CT.
Oliver, R.L. (1993). "Cognitive, Affective, and Attribute
10. References Bases of the Satisfaction Response," The Journal of Consum-
Chin, W.W., (1998). The Partial Least Squares Approach for er Research, Vol. 20, No. 3, pp. 418-430.
Structural Equation Modeling, in George A. Marcoulides Oliver, R.L. (1997). Satisfaction: A Behavioral Perspective
(ed.), Modern Methods for Business Research, New Jersey: on the Consumer, New York: The McGraw-Hill International
Lawrence Erlbaum Associates. Editions.
Cooley, R., B. Mobasher, et al. (1997). "Web mining: infor- Parasuraman, A., Zeithaml, V.A., Berry, L.L. (1985). "A
mation and pattern discovery on the World Wide Web." Conceptual Model of Service Quality and Its Implications for
Tools with Artificial Intelligence, 1997. Proceedings., Ninth Future Research," Journal of Marketing, Vol. 49, No. 4, pp.
IEEE International Conference on: 558-567. 41-50.
Cooley, R., B. Mobasher, et al. (1999). "Data Preparation for Parasuraman, A., Zeithaml, V.A., Berry, L.L. (1988).
Mining World Wide Web Browsing Patterns." Knowledge "SERVQUAL: A Multiple-Item Scale for Measuring Con-
and Information Systems 1(1): 5-32. sumer Perceptions of Service Quality," Journal of Retailing,
Drott, M. (1998). "Using Web server logs to improve site Vol. 64, pp. 12-40.
design." Proceedings of the 16th annual international confe- Ping, R.A. Jr. (1993), "The Effects of Satisfaction and Struc-
rence on Computer documentation: 43-50. tural Constraints of Retailer Exiting, Voice, Loyalty, Oppor-
Fornell, C. (1992). "A National Customer Satisfaction Baro- tunism, and Neglect," Journal of Retailing, Vol. 69, No. 3, pp.
meter: The Swedish. Experience," Journal of Marketing, Vol. 320-52.
56, No. 1, pp. 6-22. Pitt, L.F., Watson, R.T, Kavan, C.B. (1995). "Service Quali-
Gartner (2001). E-Business Service Market Size and Forecast ty: A Measure of Information Systems Effectiveness," MIS
2001-2006, Gartner. Quarterly, Vol. 19, Issue. 2, pp. 173-187.
Jones, M.A. (1998). "Satisfaction and Repurchase Intentions Pitt, L.F., Watson, R.T, Kavan, C.B. (1997). "Measuring
in the Service Industry: The Moderating Influence of Switch- Information Systems Service Quality: Concerns for a Com-
ing Barriers," Unpublished Dissertation University of Ala- plete Canvas," MIS Quarterly, Vol. 21, Issue. 2, pp. 209-221.
bama. Pitt, L.F., Watson, R.T, Kavan, C.B. (1998). "Measuring
Jones, M.A., Mothersbaugh, D.L., Beatty, S.E. (2002). "Why Information Systems Service Quality: Lessons From Two
Customers Stay: Measuring the Underlying Dimensions of Longitudinal Case Studies," MIS Quarterly, Vol. 22, Issue. 1,
Services Switching Costs and Managing Their Differential pp. 61-80.
Strategic Outcomes," Journal of Business Research, Vol. 55, Porter, M.E. (1985). Competitive Advantage: Creating and
pp. 441-450. Sustaining Superior Performance, New York: Free Press.
Kang, H. and Bradeley G. (2002), “Measuring the Perfor- Reed, R, DeFillippi, R.J. (1990). "Causal Ambiguity, Bar-
mance of IT Service: An Assessment of SERVQUAL,” In- riers to Imitation, and Sustainable Competitive Advantage,"
ternational Journal of Accounting Information System, Vol. 3, The Academy of Management Review, Vol. 15, No. 1, pp. 88-
pp. 151-164. 102.
Kaplan, S., Sawhney, M. (2000). "E-Hubs: The New B2B Ramli, A. (2005). "Web usage mining using apriori algo-
Marketplaces," Harvard Business Review, Vol. 78. No. 3, pp. rithm: UUM learning care portal case." International confe-
97-103. rence on knowledge management, Malaysia: 1-19.
Kettinger, W.J., Lee, C.C. (1994). "Perceived Service Quality REUTERS (2003). The Web Analytics Outlook: Increasing
and User Satisfaction with the Information Services Func- Online Profit. REUTERS.
tion," Decision Sciences, Vol. 25, pp. 737-766. Rusbult, C.E., Zembrodt, I.M., Gunn, L.K. (1982). "Exit,
Kettinger, W.J., Lee, C.C. (1997). "Pragmatic Perspective on Voice, Loyalty and Neglect: Responses to Dissatisfaction in
the Measurement of Information Systems Service Quality," Romantic Involvements," Journal of Personality and Social
MIS Quarterly, Vol. 21, Issue. 2, pp. 223-240. Psychology, Vol. 43, pp. 1230-1242.
Laver, M. (1976). "Exit, Voice, and Loyalty Revisited: The Sarukkai, R. (2000). "Link prediction and path analysis using
Strategic Production and Consumption of Public and Private Markov chains." Computer Networks 33(1-6): 377-386.
Goods," British Journal of Political Science, Vol. 6, No. 4, Sculley, A.B., Woods, W.W.A. (1999). B2B Exchanges-The
pp .463-482. Killer Application in the Business to Business Internet Revo-
Lim, Y.K., Lee, L., Park, T.H. (1995). "Effects of Perceived lution. ISI Publications, Hong Kong, Bermuda.
Transactional Characteristics on Channel Members With- Sharma, S. (1996). Applied Multivariate Techniques, John
Drawal Intentions," Korean Management Review, Vol. 24, Wiley & Son, Inc. New York.
No. 3, pp. 113-144. Srivastava, J., R. Cooley, et al. (2000). "Web usage mining:
Lucking-Reiley, D. and Spulber, D.F. (2001). "Business-to- discovery and applications of usage patterns from Web data."
Business Electronic Commerce," Journal of Economic Pers- ACM SIGKDD Explorations Newsletter 1(2): 12-23.
pectives, Vol. 15, No 1, pp. 55-68. Weiss, A.M., Anderson, E. (1992). "Converting from Inde-
McDougall, G.H.G., Levesque, T. (2000). "Customer Satis- pendent to Employee Salesforce: The Role of Perceived
faction with Services: Putting Perceived Value into the Equa- Switching Costs," Journal of Marketing Research, Vol. 29,
tion," Journal of Services Marketing, Vol. 14, Issue. 5, pp. No. 1, pp. 101-115.
392-410. Withey, M.J., Cooper, W.H. (1989). "Predicting Exit, Voice,
Nunnally, J. (1978). Psychometric Theory, 2nd ed., McGraw- Loyalty and Neglect," Administrative Science Quarterly, Vol.
Hill, New York. 34, Issue 4, pp. 521-539.

7
Authorized licensed use limited to: K J Somaiya College of Engineering. Downloaded on July 29,2010 at 07:16:36 UTC from IEEE Xplore. Restrictions apply.