Nini

Muhammad Shumail 06682

Literature Review-Advanced Applied Business ResearchMuhammad Shumail 06682

9

Project name: Customer Care Evaluation of Wi-TribeArticle 1: Why Do Customer Relationship Management Applications Affect Customer Satisfaction?Author(s): Sunil Mithas, M. S. Krishnan and Claes FornellSource: Journal of Marketing, Vol. 69, No. 4 (Oct., 2005), pp. 201-209Published by: American Marketing AssociationStable URL: http://www.jstor.org/stable/30166562 . Impact factor: 3.819

This article focuses on the correlation of customer relationship management on customer knowledge and customer satisfaction. Keeping in mind the US firms and how these firms focus on relationship management applications with customer to reap maximum customer satisfaction and retention. Indirectly how firms evaluate whether they should invest in customer relationship programs to enhance customer satisfaction and retention which is highly relevant with our project on customer care centers and whether they are adequate or not.It is done through evaluating qualitative research and focusing on various other empirical studies and research models. They examine and investigate on customer knowledge and customer awareness initially. Marketing is all about customer centered focus therefore research methodology and model is evaluated to reap the maximum understanding of the customer. This is an integral part of an ongoing process to sustain competitive advantage.A primary motivation for a firm to implement CRM applications is to track customer behavior to gain insight into customer tastes and evolving needs. By organizing and using this information, firms can design and develop better products and services. A lot of companies invest in IT infrastructures to cater to customers evolving needs and respond in time.There have proven studies which show that customer satisfaction has significant implications for the economic performance of firms. For example, customer satisfaction has been found to have a negative impact on customer complaints and a positive impact on customer loyalty and usage behavior. Increased customer loyalty may increase usage levels, secure future revenues, and minimize the likelihood of customer defection. Customer satisfaction may also reduce costs related to warranties, complaints, defective goods, and field service costs (Fornell 1992). Finally, in a recent study, Anderson, Fornell, and Mazvancheryl (2004) find a strong relationship between customer satisfaction and Tobin's q (as a measure of shareholder value) after controlling for fixed, random, and unobservable factors.The variable considered for customer satisfaction is the perceived quality. Further hypothesis are set below to understand in detail.H1: The use of CRM applications is associated with an improvement in the customer knowledge that firms gain.H2: Firms with greater supply chain integration are more likely to benefit from their CRM applications and achieve improved customer knowledge.H3: The use of CRM applications is associated with greater customer satisfaction.H4: Customer knowledge mediates the effect of CRM applications on customer satisfaction.Detail of Independent and Dependent variables: Research design and Methodology is derived from customer satisfaction data gathered from American Consumer Satisfaction Index (ACSI), which is considered to be a reliable indicator. CUSTKNOW is another variable is a binary variable for which 1 indicates that the firm has gained significant knowledge about its customer from its customer related IT systems. CRM applications (CRMAPLC). This variable encompasses both the legacy IT applications (i.e., the applications that firms developed before modern CRM applications were introduced) and newer IT applications (i.e., the integrated suite of marketing and sales applications developed by CRM and enterprise resource-planning vendors Supply chain integration (SCMINTGR). This variable refers to the extent to which a firm's suppliers and partners are included in its electronic supply chain and how much access they have to the firm's customer-related data or application. IT intensity (ITINVPC). This variable refers to the level of IT investment as a percentage of the firm's sales revenue. Industry sector (MFG). This indicator variable represents whether the firm's offering is primarily a good or a service (1 = good, 0 = service). Firm size (SIZE). This variable is the natural log of the firm's sales revenue.Model 1Probit model with the specification (1) Probability(CUSTKNOW = 1) = (10 + 11CRMAPLC+ 12ITINVPC + 13MFG + 14SIZE + (15SCMINTGR + 16CRMAPLC x SCMINTGR + )Model 2Linear model ACSI = (20 + 21CRMAPLC+ 22ITINVPC + 23MFG + 24SIZE + (25SCMINTGR + 26CUSTKNOW + ).The sample size for Model 1 is 360, and the sample size for Model 2 is 40. Table 1 shows the results of empirical estimation of the models 1 and 2.Consistent with H1, we find that CRM applications are positively correlated and significant with an improvement in customer knowledge (Column 1 of Table 1; 11 = .280, p < .001). Because the probit model is inherently nonlinear, we interpret the effect of each individual variable, holding all other variables at their mean values. In H2, we posit a moderating effect of supply chain integration on the relationship between CRM applications and customer knowledge. We find support for this hypothesis because the joint hypothesis test for the terms involving CRM applications and the interaction of CRM applications with supply chain integration is statistically significant. This result suggests that CRM applications are likely to have a greater association with customer knowledge when firms are electronically integrated in their supply chain and share customer-related data with their supply chain partners. We also find support for H3, which posits a positive association between CRM applications and customer satisfaction (Column 2 of Table 1; 21 = 1.266, p < .069). In H4, we suggest that the association between CRM applications and customer satisfaction is mediated by the effect of CRM applications on customer knowledge. We used the Sobel test to assess this mediation effect (Baron and Kenny 1986). We find evidence for the indirect effect of CRM applications on customer satisfaction mediated through customer knowledge (26 = 4.307, p < .028). This result implies that, holding other factors constant, firms that report an improvement in customer knowledge due to their customer-related IT systems have ACSI scores 4.307 points greater than firms that report no gains in customer knowledge following investments in CRM applications. Because the coefficient of the CRM applications variable is statistically significant in Column 2 of Table 1, our results suggest partial mediation and imply that CRM applications may also have a direct effect on customer satisfaction. The results showing the effect of control variables on customer satisfaction also provide useful insights. Note that when we control for the presence of CRM applications, the effect of IT investments on customer satisfaction is negative and statistically significant (Column 2 of Table 1; 22 = -.437, p < .011). This result is consistent with the observation that specific IT applications, such as CRM, that are directly involved in business processes affecting the customer experience may be much more effective in improving customer satisfaction than are aggregate IT investments (Mithas, Krishnan, and Fornell 2002). Focusing on CRM applications also avoids aggregation across several IT applications, in which applications may be relevant for customer satisfaction and others may have a negative or zero impact (Banker et al. 2005; Kauffman and Weill 1989; Mukhopad- hyay, Kekre, and Kalathur 1995). Column 2 of Table 1 also shows that, on average, manufacturing firms have greater customer satisfaction than services firms, a finding that is consistent with previous research (Fornell et al. 1996).

Conclusion:First, it builds on previous research that links IT systems and customer satisfaction to contribute to the cumulative knowledge in this stream of literature. Second, the study points to the importance of customer knowledge as one of the mediating mechanisms that explains the association between CRM applications and customer satisfaction. Third, the results of this study suggest that it is important to account for the effect of factors such as supply chain integration in the evaluation of returns from CRM applications. This concludes that CRM applications provides an intangible results in increase in the customer satisfaction and customer knowledge. However this model is well described but it our case it would be limited to the sense if the customer relationship management is done on a micro level and not done on a macro level example being just focusing on customer care center ambiance and training of personell.

Article 2:

Consumer Satisfaction for Internet Service Providers: An Analysis of Underlying ProcessesAuthor(s): SUNIL EREVELLESThe Belk College of Business, University of North Carolina, Charlotte, NC 28223, USASHUBA SRINIVASAN and STEVEN RANGELThe Anderson School of Management, University of California, Riverside, CA 92521,Source: INFORMATION TECHNOLOGY & MANAGEMENTQuarterly ISSN: 1385-951XImpact factor : 0.897URL: http://www.researchgate.net/publication/226527931_Consumer_Satisfaction_for_Internet_Service_Providers_An_Analysis_of_Underlying_Processes

In this paper, we examine the underlying processes involving consumer satisfaction and switching patterns among ISPs using different satisfaction models, including the expectations-disconfirmation model, the attribution model, and an affective model. Our results indicate that the satisfaction levels of ISP consumers are generally relatively low, despite the fact that consumer expectations of ISPs are also low, reflecting mediocrity in the marketplace.Consumer Satisfaction is the ultimate goal of any firm and special emphasis is laid on service providers. We study this aspect to improve the switching pattern through price incentives or free minutes. In this study suggests that there are three strategic dimensions that are determinants in influencing customer choice of ISP are customer service, ease of use and pricing.There are certain models listed to evaluate the decision making method of the ISP selection. the models are listed below:1. The expectation disconfirmation model:It is based on a paradigm where consumers are believed to form expectation prior to purchasing the service. The expectations are derived from the expectation theory based on the consumers belief that the service would have certain attributes. 2. The Attribution modelBased on attribution theory, consumers are viewed as processors of information, actively looking for reasons to explain why a purchase outcome turned out the way it was supposed to.3. Affective ModelMore based on consumers beliefs and emotions post purchase and the sense of satisfaction is derived by the realization of the adequacy of the service.4. Competitive positioning modelPositioning acts as a conformance to the image of the firms offering so the consumer appreciate what the product is and what it offers. It is to target a certain perception in the mindset of the consumers.

Consumers listed the following parameters influencing the choice of ISP during the qualitative research Response time Techinical support Price Payment method Responsiveness of the service Banner Ads Ease of software installation User friendliness of software Conclusion and Limitations:It is much more scientific method of evaluating the satisfaction level derived from existing and discontinued ISPs. The expectations-disconfirmation model is a basic framework for determining if the basic expectations of a consumer, based on his/her past experiences as well as marketplace conditions and communications, are met. Affective models go beyond basic expectations, and provides the manager with an understanding of determinant and differentiating factors, that in a crowded marketplace, may be a competitive advantage of the ISP. Attribution theory helps the manager understand consumer processing in case of dissatisfaction, and consequently provides a framework to correct for it.

The study is focused not on actual behavior rather than the actual behavior. The results are based from one geographical location where the consumer awareness is much high and can lead to higher switching behavior due to availability of more substitutes.

Information Technology and Management 4, 6989, 2003 2003 Kluwer Academic Publishers. Manufactured in The Netherlands.

Consumer Satisfaction for Internet Service Providers:An Analysis of Underlying ProcessesSUNIL EREVELLESThe Belk College of Business, University of North Carolina, Charlotte, NC 28223, USA

SHUBA SRINIVASAN and STEVEN RANGELThe Anderson School of Management, University of California, Riverside, CA 92521, USA

Abstract. A key managerial challenge, of interest to academics and practitioners alike, is the assessmentand management of customer satisfaction. In this paper, we examine the underlying processes involvingconsumer satisfaction and switching patterns among ISPs using different satisfaction models, including theexpectations-disconfirmation model, the attribution model, and an affective model. Our results indicate thatthe satisfaction levels of ISP consumers are generally relatively low, despite the fact that consumer expecta-tions of ISPs are also low, reflecting mediocrity in the marketplace. In addition, consumers attribute theirdissatisfaction to ISP indifference and believe that managing dissatisfaction is within the control of the ISP.Moreover, affective factors play an important role in satisfaction processes and switching behavior. Cus-tomer service including technical support and responsiveness of service staff is an important determinantfactor in ISP selection. We suggest that as the ISP market matures, service providers that pay attention toaffective factors and to building relationships with their customers will have a competitive advantage inthe marketplace of the future.

1. Introduction

Consumer satisfaction is the central element of the marketing concept. As Pfaff [35]eloquently stated: There is but little doubt that the maximization of consumer satis-faction is considered by most to be the ultimate goal of the market economy. If con-sumer satisfaction is the fundamental element of the market economy, it is importantto understand satisfaction and dissatisfaction processes among Internet Service Provider(ISP) consumers to better market ISP services to them, develop new products, man-age competitive forces, provide supporting services, and price services. Having under-stood underlying satisfaction processes, it would further be important to develop con-ceptual and empirical guidelines that can be applied in the management and marketingof ISPs.

There has been relatively little research (e.g., [1,30,41,50]) on consumer satisfac-tion in ISP markets. Most of this research has been practitioner-oriented survey researchor conjecture, and none has rigorously studied underlying satisfaction processes. Var-ious theoretical paradigms exist in consumer satisfaction literature that may give us abetter understanding of the satisfaction processes of ISP consumers and the positionof ISPs in consumers minds. In this paper, we examine consumer satisfaction, switch-

70 EREVELLES, SRINIVASAN, AND RANGEL

ing behavior, and consumer perceptions of ISP consumers using four different modelsto obtain a more thorough understanding of the processes involved.

1.1. ISP overview

In the past year, there has been an exponential growth in the use of the Internet, andInternet service providers have rushed into the marketplace to service this demand.Studies conducted by the International Data Corporation (IDC) show that Internet Ser-vice Provider revenue reached $10.7 billion in 1998 and the figure is expected to reach$37.4 billion by 2003, a growth rate of 28% a year. America Online was the leading ISPin 1998, holding a 23% market share. Other leading providers include MSN, Earthlink,MCI and NetZero. The industry continues to undergo consolidation and the number ofInternet service providers is considerably less than it was a few years ago [14]. Estimatesfor the total number of Internet users vary. As of April 1999, the IDC places the figurenear ninety six million with the number in the active universe for one week in April 1999at 41.9 million. These numbers are expected to rise.

Though the number of Internet providers has decreased, competition remains high,especially in the area of new technology. Firms must find ways to attract new customers,perhaps by luring them away from competitors, and retaining those that they alreadyhave. In this effort, many Internet service providers have begun to specialize in provid-ing services that cater to specific market segments. For example, Earthlink now offersa new anti-spam feature and special options for businesses, while NetZero caters to theprice conscious segment. Some firms have taken customer service one step further byoutsourcing the Internet connection and purchasing links. These firms provide customerservice only, and rely on the companies they purchase links from, to provide the connec-tion and the hardware [6].

The industry is undergoing technological changes as well. Most Internet sub-scribers are equipped with a 28.8 kbps modem. Cable Modem access and Digital Sub-scriber Lines (DSL) allow customers to cruise the Internet at speeds up to 7 Mbps [38]providing much faster Internet access and page loading. Recent estimates place the totalnumber of cable access subscribers near one million. Changes such as cable modemaccess and digital subscriber lines have begun to make smaller Internet service providersrealize that the future may be out of their reach unless they can secure portals to connectto high speed lines. The competitive nature of the industry makes this option difficult.

Any Internet company should focus on attracting as many new customers as possi-ble, as well as retaining the customers that they already have. Some firms have focusedon customer service to accomplish this while recent technological innovations may seemto suggest that the focus should be placed on the speed of the connection. The numberof ads per page and their relative complexity adds to the time it takes for a page to load.While this may be a source of discontent for some users, service providers like NetZerooffer free access to those who are willing to tolerate a large number of ads on their webpage.

CONSUMER SATISFACTION OF INTERNET SERVICE PROVIDERS 71

In addition to examining underlying satisfaction processes, this paper also exam-ines those attributes that lead to customer satisfaction among ISP consumers. In addi-tion, we examine the behavior of those individuals that have switched ISPs and thosethat have not. We also analyze service perceptions towards various ISPs, and the posi-tion that competing ISPs occupy in consumers minds. Internet Service Providers haveconcentrated on attracting new subscribers (for example, with price incentives and freeminutes of online access) but have placed relatively little emphasis on strategies to retaincustomers. While attributes like speed and price intuitively appear to play an importantrole in customer satisfaction and retention, fundamental questions remain about the va-lidity of these assumptions and the significance of other attributes as we move acrossmarket segments, as well as into a more competitive marketplace.

Understanding what these attributes are and how they rank in importance to theconsumer can be helpful in determining how to tailor services to accommodate variousmarket segments that exist in the online community. It can also lead to more focusedmarketing campaigns and help ISPs better allocate resources to areas important to theconsumer. With a clear knowledge of customer satisfaction processes, ISPs shouldbetter be able to keep current customers satisfied as well as attract new customers.

1.2. Objectives and organization of paper

A key managerial challenge, of interest to marketing academics and practitioners alike,is the assessment of the extent of customer satisfaction and dissatisfaction among con-sumers. The anecdotal examples in the introductory section illustrate the importance ofcustomer satisfaction in a market that is in a growth phase. Our objective in this paperis to examine consumer satisfaction processes and switching patterns among ISPs usingdifferent models of consumer satisfaction. Specifically, we develop an understanding ofthe processes involved by using the expectations-disconfirmation model, the attributionmodel, affective models and a competitive positioning model.

Previous research on customer satisfaction toward ISP services is sparse and thisstudy fills the gap. Our results indicate that the satisfaction levels of ISP consumers arerelatively low, as are the expectation levels of consumers, suggesting mediocrity in themarketplace. Results of the attribution models indicate that consumers perceive their dis-satisfaction to be due to the negligence or indifference of the ISP. This suggests that ISPsneed to improve customer perceptions and their service quality. The affective model sug-gests that ISPs should pay attention to the affective component, which may be enhancedthrough relationship marketing and personal attention to consumer needs. The resultsof factor analysis suggest that there are three strategic dimensions that are determinant ininfluencing customer choice of an ISP: customer service, ease of use and pricing.The competitive positioning model confirms our findings that the customer satisfactionlevels are low; there is a significant opportunity for a new entrant positioned strongly onthe key determinant factors of customer service and ease of use.

This paper is organized as follows: section 2 discusses the models of consumer be-havior that we use; section 3 outlines the research methodology; and section 4 provides


the results of our study. Section 5 presents the conclusions and limitations of the presentstudy and offers directions for future research.

2. Models of customer satisfaction

2.1. The expectations-disconfirmation modelWe first analyze consumer satisfaction and switching behavior of ISP consumers us-ing the expectations-disconfirmation model. The expectations-disconfirmation model isbased on a paradigm that has dominated consumer satisfaction/dissatisfaction researchfor many years [10]. According to this paradigm, consumers are believed to form ex-pectations about a service prior to purchasing the service [31]. The notion of consumersforming expectations is derived from expectancy theory [44], and is generally definedas a consumers beliefs that a service (in this case, the ISP) possesses certain desiredattributes. Subsequent post-purchase usage then reveals to the consumer, the actual per-formance of the service. The consumer then compares this post-purchase evaluation withthe expectations held prior to purchase. If the product performed better than expected(perceived actual performance > expected performance), positive disconfirmation is ex-pected to occur. This leads to consumer satisfaction, strengthens consumers attitudestowards the service, and results in positive word-of-mouth. If however, in the consumersevaluation, the product performs worse than expected (perceived actual performance 44 2%

Current education level Some high school 0%Completed high school 2%Some college 42%Completed college 15%Some graduate school 34%Completed graduate school 7%

support, price, payment method, responsiveness of service, banner ads, ease of softwareinstallation, and user friendliness of software. The six firms that are of interest are AOL,Earthlink, MSN, NetZero, AT&T and Pacbell. Fifty-nine individuals (n = 59) eachrate all firms on the eight attributes (Likert-like scales containing seven rating points areemployed for such measures). The first step is the generation of a correlation matrixfor the j attributes. The correlation matrix would reveal the sets of attributes that are


correlated. Any conclusions about the existence of meaningful factors ultimately mustbe supported by theoretical considerations. Having identified the factors, the brandsare then evaluated on these factors by computing an overall factor score. The factorscore is simply the average of all the attribute evaluations contributing to that factor,which may be averaged over n individuals for the brands overall position on the map.For more details on factor analysis, see Rummel [39] and Harman [15]. Firms are thenpositioned by factor scores.

4. Results and discussion

4.1. Expectations-disconfirmation modelConsumers overall satisfaction with their current ISP was significantly higher than theirsatisfaction with their previous ISP (Ms = 5.36 vs. 3.52, t (1, 24) = 3.85, p = 0.001).This is not surprising, as the relatively low satisfaction of consumers with their previousISP is what prompted them to switch ISPs.

There was no significant difference between consumer expectations of their currentISP and disconfirmation of those expectations (t = 0.535). This lack of either positiveor negative disconfirmation is quite interesting, as it indicates only simple confirma-tion among consumers (see table 2). In other words, consumers perceptions of theircurrent service provider can best be thought of as neutral, and indicates a lack of sig-nificant satisfaction, much less delight among ISP consumers. In other words, thereis considerable potential for the improvement of the services provided by ISPs, and theconsequent consumer delight.

A significant difference between participants expectations of their previousISP and the consequent disconfirmation was, however revealed (Ms = 5.32 vs. 4.08,t (1, 24) = 2.509, p = 0.019). This disconfirmation is negative, and indicates dissat-isfaction among consumers. This is consistent with earlier results, and explains whyconsumers switched ISPs.

Table 2Expectations-disconfirmation model.

Variable M t df P -level(2-tailed)

Current ISP satisfaction 5.36 3.846 24 0.001Previous ISP satisfaction 3.52

Current ISP expectations 5.28 0.535 24 0.598Current ISP disconfirmation 5.08

Previous ISP expectations 5.32 2.509 24 0.019Previous ISP disconfirmation 4.08

Current ISP expectations 5.28 0.118 24 0.907Previous ISP expectations 5.32


There was no difference between participants expectations of their current ISPand their previous ISP (t = 0.118). This again is interesting, as it may indicate thatconsumers, while dissatisfied with their previous ISP, did not really expect significantlybetter service from their new ISP. Rather, the reason for their switching may be explainedby their motivation to avoid the bad experiences with their previous ISP, rather than theirmotivation to gain a better experience with another ISP. Again this indicates mediocrityin the marketplace, and leaves open the possibility of a competitor entering the marketwith a superior service, and consequently gaining considerable competitive advantage.

4.2. Attribution model

Dissatisfaction with their previous ISP was rated by consumers (see table 3) as be-ing overwhelmingly due to the company (external), rather than to themselves (internal)(Ms = 4.78 vs. 2.96, t (1, 22) = 2.939, p = 0.008). This highly significant differ-ence in means is interesting, as it may suggest that the consumers perceive the ISP to bealmost negligent in providing the service, as opposed to perceiving a general weaknessor deficiency in the industry. Moreover, further support for this perception is obtainedfrom consumers responses to the control issue. Dissatisfaction with the previousISP was attributed to being under the control of the company rather than the consumer(Ms = 5.21 vs. 3.04, t (1, 23) = 4.138, p = 0.0004). In other words, subjects feltthat the ISP in question could have provided better service (i.e., it was under their controlto do so), but simply did not strive to do so. Two implications can be drawn from thisfinding. The first is that, at the very least, the management of consumer perceptions ofmost ISPs can be improved. If ISPs are able to convince consumers that some of theirdissatisfaction towards the ISP is not under the control of the company, consumers maybe more understanding, and less likely to switch service providers. Second, there ap-pears to be tremendous potential opportunities for an ISP that is able to minimize someof the causes for this dissatisfaction.

4.3. Affective modelAffective (emotional) ratings were significantly more positive (see table 4) for currentvs. previous ISPs (Ms = 0.735 vs. 0.145, t (1, 24) = 2.275, p = 0.032). In ad-dition, cognition ratings were significantly more positive for current vs. previous ISPs

Table 3Attribution model.


Previous ISP dissatisfaction due to self (internal) 2.96 2.939 22 0.008Previous ISP dissatisfaction due to ISP (external) 4.78Previous ISP dissatisfaction controllable by company 5.21 4.138 23 0.00003Previous ISP dissatisfaction controllable by consumer 3.04


Table 4Affective model.


Previous ISP affect 0.145 2.275 24 0.032Current ISP affect 0.735

Previous ISP cognition 0.147 3.014 24 0.006Current ISP cognition 1.18

Previous ISP attitude 0.22 2.881 24 0.008Current ISP attitude 1.1

Previous ISP affect 0.15 2.626 24 0.015Previous ISP cognition 0.15

Current ISP affect 0.735 6.073 24 0.00003Current ISP cognition 1.18

(Ms = 1.18 vs. 0.147, t (1, 24) = 3.014, p = 0.006). These results are interesting,because they indicate that both emotional and cognitive factors drive, or are affected byconsumer-switching behavior. The emotional component may be influenced by inter-personal dealings with the ISP, reputation and image of the ISP, etc., while the cognitivecomponent may be influenced by technical superiority, speed of ISP, etc. Not surpris-ingly, overall attitude was significantly more positive for the current ISP vs. the previousISP (Ms = 1.10 vs. 0.220, t (1, 24) = 2.881, p = 0.008).

We also conducted pair-wise comparisons of the affective and cognitive compo-nents both before and after the switching behavior. If there is no significant differencebefore, and a significant difference after switching (or vice-versa), one could concludethat the switch had either been driven by, or had resulted in an improvement of, one orthe other component. This would have been an interesting result with potential impli-cations for ISP management. However the components were found to be significantlydifferent in magnitude both before and after switching behavior (at the 0.05 and 0.0005levels, respectively).

Affect and cognition ratings for both previous and current ISPs were highly cor-related (r = 0.959). On ratings of their current ISP, participants rated affect as lesspositive than cognition ratings (Ms = 0.735 vs. 1.18, t (1, 24) = 6.073, p = 0.0003).A similar pattern was revealed in ratings for previous ISP; affect was less positive thancognition ratings (Ms = 0.145 vs. 0.147, t (1, 24) = 2.626, p = 0.015). Thismay indicate that there is potential for all ISP providers to build stronger emotional linksto their customers, and thus minimize switching behavior in the event of a cognitivelybased reason for dissatisfaction.

4.4. Exploratory factor analysisA series of tests was conducted to determine if the data were suitable for factor analy-sis (see [42]). Bartletts sphericity test indicates that the distribution is ellipsoidal and


Table 5Pair-wise correlations among ISP attribute variablesa.

RT TECS PRICE PAY RESP BADS EASE FRIEND

RT 1.00 0.509 0.074 0.069 0.504 0.272 0.309 0.282TECS 0.509 1.00 0.109 0.094 0.647 0.220 0.247PRICE 0.074 0.109 1.00 0.407 0.207 0.027PAY 0.069 0.094 0.407 1.00 0.355RESP 0.504 0.647 0.207 0.355BADS 0.272 0.220 0.027EASE 0.309 0.247FRIEND 0.282aRT = response time, TECS = technical support, PAY = payment method, RESP = responsiveness ofstaff, BADS = banner ads, EASE = ease of software installation and FRIEND = user friendliness ofsoftware.

Table 6Factor matrix after varimax rotation.

Variable Factor 1 Factor 2 Factor 3Customer Service Ease of Use Pricing

Response time 0.722 0.357 0.154Technical support 0.806 0.178 0.046Price 0.039 0.137 0.774Pay 0.145 0.195 0.822Responsiveness of service staff 0.848 0.035 0.281Banner ads 0.568 0.070 0.205Ease of installation 0.165 0.850 0.198User friendliness of software 0.055 0.892 0.176Variance explained by each factor 2.266 1.739 1.492

therefore amenable to data reduction [7]. The KaiserMeyerOlkin measure of samplingadequacy is 0.85, indicating a high-shared variance and a relatively low uniqueness invariance [18].1 Next, we examine the pair-wise correlations between variables. An ex-amination of the correlation matrix in table 5 reveals that the attributes response timeto inquiries, technical support and representative friendliness are correlated. Like-wise, price and payment methods are correlated. Another set that is correlated isthe variable pair consisting of ease of software installation and user friendliness ofsoftware. Since the correlations among variables is not small, we can conclude thatfactor analysis is an appropriate method for analyzing this data.

We performed exploratory factor analysis on the data set using the maximum likeli-hood method. The maximum likelihood method of estimation is most efficient resultingin smaller standard errors. Using a stopping rule of eigenvalues of greater than one, wefind that the number of factors should be three. Table 6 shows the factor matrix after1 We also plotted the latent roots obtained from matrix decomposition. It indicated one sharp break, sug-

gesting the appropriateness of factor analysis [42].


Figure 3. Competitive positioning map.

rotation by the varimax method. Factor 1 consists of three variables: response-timeto inquiries, technical support and responsiveness of service staff, and pertains toCustomer Service. Factor 2, comprises two variables: ease of software installationand user friendliness of software and represents the Ease of Use of the software. Fac-tor 3, consisting of price and payment methods represents the Pricing Policy of theInternet service provider.

In sum, factor analysis reveals that there are three strategic dimensions that de-scribe Internet Service Providers. These are Ease of Use, Pricing and Customer Service.We are interested in developing a product-positioning map to help with respect to ser-vice design. We consider for analysis of the product positioning, four major firms inSouthern California, namely AOL, Earthlink, MSN and NetZero. The remaining firmsare classified as others. As seen in figure 3, Earthlink is perceived to be the best on


Customer Service while NetZero is perceived to be quite poor relative to competition onthis dimension. These results are consistent with past survey results [30], where Earth-link was preferred over competition by customers valuing service and user-friendliness.Consistent with these results, the Telechoice performance survey also rated AOL as be-ing poor on the Customer Service. With respect to Pricing, NetZero is perceived to bethe best. This is as expected since NetZero does not charge an access fee. There is not asignificant differentiation between MSN, AOL and Earthlink on the Pricing dimension.On the Ease of Use dimension, AOL and MSN are well positioned while NetZero isrelatively weak on the Ease of Use dimension. It is notable that the others fair wellon the Customer Service dimension. These results are consistent with the anecdotalperceptions of the ISPs gleaned from discussions with several users, and also supportthe contention [19] that smaller firms are focusing on customer service and on buildingpersonalized relationships with their customers.

A clear implication from a marketing strategy perspective is that a firm that dif-ferentiates itself from competition by offering superior customer support and reliableservice will have a sustainable strategic advantage in the market place. In a growth en-vironment such as the ISP market, only the ISPs that differentiate themselves on the keystrategic dimensions are likely to survive and be profitable in the long run.

5. Conclusions and limitations

The results of this study contribute to the recent, albeit relatively sparse body of knowl-edge about consumption behavior towards ISP services. It also clarifies a number of pastfindings on satisfaction towards ISP services that have been based on simplistic, oftenunscientific satisfaction constructs. No theoretically founded, academic research has sofar examined satisfaction processes in the ISP market. The few articles (e.g., [20,36])that have addressed issues on the topic have been based mostly on anecdotal evidenceand lack a theoretical premise. For example, Kavanaghs [20] conclusions are based on aconvenience focus group study, and on the opinions of unscientifically selected con-sultants. Radosevichs [36] conclusions, while based on a more rigorously designedstudy conducted over 2 weeks by a consulting firm, has no theoretical framework. Con-sequently, the findings have limited generalizability, and their reliability and validity areunclear. In this paper, we have examined consumer satisfaction of ISP services, usingvarious validated theoretical frameworks, in order to get a better understanding of theunderlying processes involved. We believe that an understanding of the processes in-volved should provide better inputs for the management and marketing of ISPs in thefuture.

The results obtained indicate among other things, that satisfaction levels of ISPconsumers are relatively low, as evidenced only by the simple confirmation of theirexpectations, even for the market leaders. This is an anomaly in todays highly com-petitive marketplace, and leaves current providers vulnerable to competitors, who mayfocus on better service. Moreover, consumer expectations of ISPs are relatively low,which makes confirmation of those expectations rather easy. This makes the low sat-


isfaction results especially disappointing, and suggests mediocrity in the marketplace.In addition, attribution model results indicate that consumers perceive their dissatis-faction to be due to the negligence or indifference of the ISP, and also perceive thatthis is within the control of the ISPs. This may imply that ISP providers may needto change consumer perceptions, their own service quality, or both. The results alsosuggest that both affective (emotional) and cognitive factors affect switching behavior.ISPs should pay attention to this affective component, that may be enhanced throughrelationship marketing and personalized responsiveness to consumer problems. Over-all, it appears that ISPs have neglected the affective component, preferring instead tomarket their services like commodities. We suggest that an ISP that does not thinkof their service as a commodity will have a tremendous advantage in the market-place.

It is likely that the reason for the mediocre satisfaction scenario described abovemay be due to the fact that the ISP market is in the early stages of its product life cycle.Consequently, service providers are more interested in staking out and gaining marketshare, rather than consolidating their position, through building strong relationships withcustomers. This is a shortsighted strategy, because, as the market matures, the providerswho have nurtured these relationships are likely to be the market leaders. There is nodearth of ISPs entering the market, and we suggest that existing ISPs understand un-derlying satisfaction processes and act upon them, if they hope to survive in the highlycompetitive marketplace of the future. This study suggests that there is considerablepotential for the improvement of services provided by most ISPs.

This study also seriously questions the erroneous assumption by ISPs that con-sumers base their selection of their ISP solely on specification attributes and price. Theresults indicate that this is not always the case. Results suggest that there are strategicdimensions that are determinant in influencing customer choice of an Internet serviceprovider: customer service, ease of use and pricing. The results underscore the im-portance of customer service consisting of attributes such as response time to inquiries,technical support and responsiveness of service staff.

The results of the competitive positioning model indicate Earthlink is perceivedto be offering the highest levels of customer service while America Online is perceivedto be good on the ease of use dimension. NetZero has an advantage with respect tothe pricing dimension. The results suggest that there is a significant opportunity fora new entrant positioned strongly on the key determinant factors of customer serviceand ease of use, dimensions current major players are neglecting as indicated by theresults.

It may be that specification attributes and price are salient factors, but other fac-tors such as customer service and ease of use are determinant. These results supportrecent findings (e.g., [1,19]) on the importance of service quality and relationships onpurchase decisions, and also provide an explanation for seemingly unexplainable switch-ing behavior. From a marketing strategy perspective, our results suggest that an ISP thatdifferentiates itself from competition by offering superior customer service and reliableservice will have a sustainable strategic advantage in the market place.


In a growth environment such as the ISP market, only the ISPs that differentiatethemselves on the key strategic dimensions are likely to survive and be profitable in thelong run. Smaller ISPs that provide personalized services and build strong relationshipswith consumers are likely to carve lucrative niches for themselves, at the expense of thelarger providers. On a similar vein, analysts believe that ISPs will need to offer a morecomplete package of services to satisfy their customers [19]. This will include hightechnology and personnel costs, and is likely to result in an increase in mergers amongISPs.

This paper contributes theoretically to the literature by providing a rigorous un-derstanding of the underlying processes involved in consumer satisfaction with ISPs.The paper also provides a better understanding of the underlying processes involved inswitching behavior in case of dissatisfaction with an ISP. No such rigorous analysis ex-ists in MIS literature so far. This is especially so in the study of satisfaction processesfor ISPs. Managerially, this paper makes a contribution by helping an ISP managerdetermine where the strengths and weaknesses of their service and marketing strategylie. In the past, the manager may just have been aware of the customer satisfaction rat-ings of his/her ISP service, but not have known exactly where the problem lies, or whathe/she should do to correct for it. The findings of this paper help a manager by pro-viding a framework for analyzing satisfaction problems and strategy. The expectations-disconfirmation model is a basic framework for determining if the basic expectationsof a consumer, based on his/her past experiences as well as marketplace conditions andcommunications, are met. Affective models go beyond basic expectations, and providesthe manager with an understanding of determinant and differentiating factors, that ina crowded marketplace, may be a competitive advantage of the ISP. Attribution the-ory helps the manager understand consumer processing in case of dissatisfaction, andconsequently provides a framework to correct for it.

The generalizability of these results should not, however, be overstated. One rea-son may be that this study is based on reported rather than actual behavior. Anotherlimitation may be the fact that participants in this study come from a single geographicalregion and country (Southern California, USA), where a fair assumption can be madethat ISP consumers are relatively more knowledgeable and experienced with ISPs thanconsumers in other geographical regions or countries. It is possible that consumer satis-faction may be higher and switching behavior lower in other geographical regions. Nodoubt, further empirical research needs to be carried out in order to obtain a more com-prehensive picture of the ISP market. A possibility would be to examine the moderatingeffects of product class knowledge on satisfaction and switching behavior. It is also pos-sible that different results may be obtained for ISPs selected for personal or professionaluse. It is therefore not clear how generalizable the findings of this study are to varioussituations. It seems to be a reasonable conclusion, however, to say that the results of thisstudy serve to provide a considerably more thorough understanding of the current ISPmarket, the underlying satisfaction processes and the constructs involved. Additionally,the findings provide a better picture of consumer service perceptions and the positionsof various providers in consumer minds.


Acknowledgements

The authors acknowledge the invaluable research assistance of Lorie Silva in the prepa-ration of this manuscript.

References

[1] L. Armstrong, The mac of Internet providers, Business Week 3557 (December 1997) 138.[2] W.O. Bearden and J.E. Teel, Selected determinants of customer satisfaction and complaint reports,

Journal of Marketing Research 20(2) (1983) 128.[3] R.N. Bolton and J.H. Drew, A multi-stage model of customers assessments of service quality and

value, Journal of Consumer Research 17 (March 1991) 375384.[4] J.B. Cohen and C.S. Areni, Affect and consumer behavior, in: Handbook of Consumer Behavior, eds.

A. Robertson and H. Kassarjian (Prentice-Hall, Englewood Cliffs, NJ, 1991) pp. 183240.[5] S.L. Crites, L.R. Fabrigar and R.E. Petty, Measuring the affective and cognitive properties of attitudes:

Conceptual and methodological issues, Personality and Social Psychology Bulletin 20(6) (1994) 619634.

[6] R.M. Dewan, M. Freimar and A. Seidman, Internet service providers, proprietary content and thebattle for users dollar, Communications of the ACM 41(8) (1998) 4348.

[7] W. Dillon and M. Goldstein, Multivariate Analysis: Methods and Applications (Wiley, 1984).[8] J.F. Engel, R.D. Blackwell and P.W. Miniard, Consumer Behavior (Holt, Rinehart and Winston, Inc.,

New York, 1990).[9] S. Erevelles, The role of affect in marketing, Journal of Business Research 42(3) (1998) 199216.

[10] S. Erevelles and C. Leavitt, A comparison of current models of consumer satisfaction/dissatisfaction,Journal of Consumer Satisfaction, Dissatisfaction and Complaining Behavior 5 (1992) 104114.

[11] V.S. Folkes, Consumer reactions to product failure: An attributional approach, Journal of ConsumerResearch 10 (1984) 398409.

[12] V.S. Folkes, Recent attribution research in consumer behavior: A review and new directions, Journalof Consumer Research 14 (March 1989) 548565.

[13] T. Green, Qwest and IXC join big boys with DSL offering, Network World 16(32) (August 1999) 8,61.

[14] M. Hamblen, Telcos threaten Internet service providers, Computer World 31(34) (August 1997) 32.[15] H.H. Harman, Modern Factor Analysis, 3rd edn. (University of Chicago Press, Chicago, IL, 1976).[16] J.R. Hauser and F.S. Koppelman, Alternative perceptual mapping techniques: Relative accuracy and

usefulness, Journal of Marketing Research 16 (November 1979) 495506.[17] C.E. Izard, Human Emotions (Plenum Press, New York, 1977).[18] H.F. Kaiser and B.A. Cerny, Factor analysis of the image correlation matrix, Educational and Psycho-

logical Measurement 39(4) (Winter 1979) 711714.[19] K. Kaplan, ISPs face a choice: Evolution or extinction, Los Angeles Times (6 September 1999) C1,

C6.[20] M. Kavanagh, ISP churn caused by poor service, not pricing, Marketing Week 21(30) (September

1998) 43.[21] H.H. Kelly, Causal schemata and the attribution process, in: Attribution: Perceiving the Causes of

Behavior, eds. E.E. Jones et al. (General Learning Press, Morristown, NJ, 1972) pp. 151174.[22] C.S. Kohli and L. Leuthesser, Product positioning: A comparison of perceptual mapping techniques,

Journal of Product and Brand Management 2(4) (1993) 1019.[23] P. Kotler, Future speak, Sales and Marketing Management 151(5) (1999) 14.[24] S. Krishnan and V.A. Valle, Dissatisfaction attribution and consumer complaint behavior, Advances

in Consumer Research, 6th edn., ed. W.L. Wilkie (Association for Consumer Research, Ann Arbor,MI, 1979) pp. 445449.


[25] P.A. LaBarbera and D. Mazursky, A longitudinal assessment of consumer satisfaction/dissatisfaction:The dynamic aspect of the cognitive process, Journal of Marketing Research 20 (November 1983)393404.

[26] R. Lal and R. Rao, Supermarket competition: The case of everyday low pricing, Marketing Science16(1) (1997) 6080.

[27] H. Mano and R.L. Oliver, Assessing the dimensionality and structure of the consumption experience:Evaluation, feeling and satisfaction, Journal of Consumer Research 20(3) (1993) 451466.

[28] C.F. Mela, S. Gupta and K. Jedidi, Assessing long-term promotional influences on market structure,International Journal of Research in Marketing 15(2) (1998) 89107.

[29] J.A. Miller, Studying satisfaction, modifying models, eliciting expectations, posing problems andmaking meaningful measurements, Conceptualization and Measurement of Consumer Satisfactionand Dissatisfaction, ed. H.K. Hunt (Marketing Science Institute, Cambridge, MA, 1977) pp. 7291.

[30] A. Morri, Make it click, Telephony 234(9) (March 1998) 2228.[31] R.L. Oliver, A cognitive model of the antecedents and consequences of satisfaction decisions, Journal

of Marketing Research 17 (November 1980) 460469.[32] R.L. Oliver, Measurement and evaluation satisfaction processes in retail settings, Journal of Retailing

57 (Fall 1981) 2548.[33] R.L. Oliver, Cognitive, affective, and attribute bases of the satisfaction response, Journal of Consumer

Research 20(3) (December 1993) 418430.[34] R.L. Oliver and W.S. DeSarbo, Response determinants in satisfaction judgements, Journal of Con-

sumer Research 14 (March 1988) 495507.[35] M. Pfaff, The index of consumer satisfaction: Measurement problems and opportunities, in: Concep-

tualization and Measurement of Consumer Satisfaction and Dissatisfaction, ed. H.K. Hunt (MarketingScience Institute, Cambridge, MA, 1976).

[36] L. Radosevich, ISP performance remains steady, InfoWorld 19(43) (1997) 79, 84.[37] M.L. Richins, Negative word-of-mouth by dissatisfied consumers: A pilot study, Journal of Marketing

47 (Winter 1983) 6878.[38] D. Rohde, DSL explosion ready to rip, Network World 16(29) (July 1999) 1, 14.[39] R.J. Rummel, Applied Factor Analysis (Northwestern University Press, Evanston, IL, 1970).[40] R.A. Spreng, S.B. MacKenzie and R.W. Olshavsky, Reexamination of the determinants of consumer

satisfaction, Journal of Marketing 60(3) (July 1996) 1532.[41] T. Stein, Success scope, Success 43(3) (April 1996) 22.[42] D. Stewart, The application and misapplication of factor analysis in marketing research, Journal of

Marketing Research 35 (1981) 5162.[43] J.E. Swan and L.F. Trawick, Disconfirmation of expectations and satisfaction with a retail service,

Journey of Retailing 57 (Fall 1981) 4967.[44] E.C. Tolman, Purposive Behavior in Animals and Men (Appleton-Century, New York, 1932).[45] G. Urban and J. Hauser, Design and Marketing of New Products (Prentice-Hall, Englewood Cliffs,

NJ, 1993).[46] B. Weiner, Human Motivation (Holt, Rinehart and Winston, Inc., New York, 1980).[47] R.A. Westbrook, Intrapersonal affective influences upon consumer satisfaction with products, Journal

of Consumer Research 7 (June 1980) 4954.[48] R.A. Westbrook, Product/consumption-based affective responses and postpurchase processes, Journal

of Marketing Research 24 (August 1987) 258270.[49] R.A. Westbrook and R.L. Oliver, The dimensionality of consumption emotion patterns and consumer

satisfaction, Journal of Consumer Research 18 (June 1991) 8491.[50] J. Wexler, Nobodys perfect, Network World 13(18) (April 1996) 10.[51] P.T. Wong and B. Weiner, When people ask why questions and the heuristics of attributional search,

Journal of Personality and Social Psychology 40 (April 1981) 650663.

Why Do Customer Relationship Management Applications Affect Customer Satisfaction?Author(s): Sunil Mithas, M. S. Krishnan and Claes FornellSource: Journal of Marketing, Vol. 69, No. 4 (Oct., 2005), pp. 201-209Published by: American Marketing AssociationStable URL: http://www.jstor.org/stable/30166562 .Accessed: 21/09/2014 08:41

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .http://www.jstor.org/page/info/about/policies/terms.jsp

.

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range ofcontent in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new formsof scholarship. For more information about JSTOR, please contact [email protected].

.

American Marketing Association is collaborating with JSTOR to digitize, preserve and extend access toJournal of Marketing.

http://www.jstor.org

This content downloaded from 111.68.111.155 on Sun, 21 Sep 2014 08:41:42 AMAll use subject to JSTOR Terms and Conditions

Sunil Mithas, M.S. Krishnan, & Claes Fornell

Why Do Customer Relationship Management Applications Affect

Customer Satisfaction? This research evaluates the effect of customer relationship management (CRM) on customer knowledge and customer satisfaction. An analysis of archival data for a cross-section of U.S. firms shows that the use of CRM applications is positively associated with improved customer knowledge and improved customer satisfaction. This article also shows that gains in customer knowledge are enhanced when firms share their customer-related information with their supply chain partners.

Despite substantial investments in customer relationship management (CRM) applications, there is a lack of research demonstrating the benefits of such

investments. In particular, there has been limited research on the role and contribution of CRM applications in managing customer encounters (Bitner, Brown, and Meuter 2000; Meuter et al. 2000). Although marketing and information systems researchers have developed theories about the effect of CRM applications, with some progress toward empirical validation (Jayachandran et al. 2005; Reinartz, Krafft, and Hoyer 2004; Romano and Fjermestad 2003; Srinivasan and Moorman 2005), there is limited knowledge about the effect of CRM applications on a firm's customer knowledge and customer satisfaction. Furthermore, prior research does not shed light on why CRM applications affect customer satisfaction or the role of complementary investments in supply chain management systems.

Against the backdrop of significant investment in CRM applications and limited empirical work on the effect of CRM applications on customer relationships, this article

Sunil Mithas is an assistant professor, Robert H. Smith School of Busi- ness, University of Maryland (e-mail: [email protected]). M.S. Krishnan is a Professor of Business Information Technology and Area Chair and Michael R. and Mary Kay Hallman e-Business Fellow (e-mail: [email protected]), and Claes Fornell is Donald C. Cook Professor of Business Administration and Director of the National Quality Research Center (e-mail: [email protected]), Ross School of Business, University of Michigan. The authors thank William Boulding, Richard Staelin, and the three anonymous JM reviewers for their guidance and helpful comments in improving this manuscript. They thank InformationWeek and the National Quality Research Center at the University of Michigan for providing the data for this research. They thank Rusty Weston, Bob Evans, Brian Gillooly, Stephanie Stahl, Lisa Smith, and Helen D'Antoni for their help in data and for sharing their insights. They also acknowledge helpful comments from Jonathan Whitaker and participants at the 2005 AMA Winter Educators' Conference in San Antonio. They thank Eli Dragolov for her excellent research assistance. Financial support for this study was provided in part by a research grant from A.T. Kearney and the Michael R and Mary Kay Hallman fellowship at the Ross School of Business.

poses the research question, What is the effect of CRM applications on customer knowledge and customer satisfaction? We performed an empirical study using a cross- section of large U.S. firms. This study includes the develop- ment of a theoretical model and the collection of archival data from the National Quality Research Center at the Uni- versity of Michigan and an InformationWeek survey of senior information technology (IT) managers. We examine the role of customer knowledge as a mediating mechanism to explain the effect of CRM applications on customer satisfaction. We also study the moderating effect of supply chain integration in leveraging CRM applications.

We structure the rest of the article as follows: In the next section, we review the literature and develop the hypotheses. Then, we discuss the methodology and present the results. We conclude with a discussion of the implications of the study.

Research Model and Theory The customer equity literature provides the basic rationale for investing in customer relationships. There is increasing recognition of the importance of managing customer relationships and customer assets. Marketing has moved from a brand-centered focus to a customer-centered approach. Hogan, Lemon, and Rust (2002) argue that the ability to acquire, manage, and model customer information is key to sustaining a competitive advantage. Berger and colleagues (2002) develop a framework to assess how customer database creation, market segmentation, customer purchase forecasting, and marketing resource allocations affect customers' lifetime value to the firm. Hogan and colleagues (2002) extend this work and provide conceptual support for linking customer assets (in terms of customer lifetime value) and financial performance.

Next, we develop the hypotheses for the effect of CRM applications on customer knowledge and customer satisfaction. We also discuss the moderating role of supply chain integration to understand the effect of CRM applications on customer knowledge.

(c) 2005, American Marketing Association Journal of Marketing ISSN: 0022-2429 (print), 1547-7185 (electronic) 201 Vol. 69 (October 2005), 201-209


The Effect of CRM Applications on Customer Knowledge A primary motivation for a firm to implement CRM applications is to track customer behavior to gain insight into customer tastes and evolving needs. By organizing and using this information, firms can design and develop better products and services (Davenport, Harris, and Kohli 2001; Nambisan 2002). Davenport and Klahr (1998) argue that customer knowledge has certain attributes that make it one of the most complex types of knowledge. For example, customer knowledge may be derived from multiple sources and media and may have many contextual meanings. Customer knowledge is also dynamic, and it changes rapidly.

Customer relationship management applications facilitate organizational learning about customers by enabling firms to analyze purchase behavior across transactions through different channels and customer touchpoints. Glazer (1991) provides examples of how FedEx and Ameri- can Airlines used their investments in IT systems at the customer interface to gain valuable customer knowledge. More recently, firms have invested in an integrated set of tools and functionalities offered by leading software vendors to gather and store customer knowledge. Firms with greater deployment of CRM applications are in a better position to leverage their stock of accumulated knowledge and experience into customer support processes. In addition, firms with a greater deployment of CRM applications are likely to be more familiar with the data management issues involved in initiating, maintaining, and terminating a customer relationship. This familiarity gives firms a competitive advantage in leveraging their collection of customer data to customize offerings and respond to customer needs.

Customer relationship management applications help firms gather and use customer knowledge through two mechanisms. First, CRM applications enable customer contact employees to record relevant information about each customer transaction. After this information is captured, it can be processed and converted into customer knowledge on the basis of information-processing rules and organizational policies. Customer knowledge captured across service encounters can then be made available for all future transactions, enabling employees to respond to any customer need in a contextual manner. Firms can also use customer knowledge to profile customers and identify their latent needs on the basis of similarities between their purchase behaviors and those of other customers. Second, firms can share their accumulated customer knowledge with customers to enable those customers to serve themselves by defining the service and its delivery to suit their needs (Pra- halad, Ramaswamy, and Krishnan 2000). The process of customer self-selection of service features provides additional opportunities for firms to learn about their customers' evolving needs and to deepen their customer knowledge.

H1: The use of CRM applications is associated with an improvement in the customer knowledge that firms gain.

The Moderating Role of Supply Chain Integration Supply chain integration refers to the extent to which a firm shares relevant information about its customers with its sup-

ply chain partners. Supply chain integration ensures that products and services offered by various organizational units and suppliers are coordinated to provide a better customer experience. Previous research suggests that integration of IT systems in a firm's value chain is essential to the realization of the full benefits of seamless information sharing and data completeness (Brohman et al. 2003; Gosain, Malhotra, and El Sawy 2005; Rai, Patnayakuni, and Pat- nayakuni 2005). For example, Fisher, Raman, and McClel- land (2000) note that IT-enabled data accuracy is critical for efficient forecasting and to design agile supply chain management processes. Anderson, Banker, and Ravindran (2003, p. 94) argue that "interweaving of IT links through- out the supply chain create[s] value by enabling each mem- ber of the supply chain to identify and respond to dynamic customer needs." Creating an integrated IT infrastructure enables organizational units to leverage their resources effectively to address customers' evolving needs (Samba- murthy, Bharadwaj, and Grover 2003). For example, superior customer ratings and the success of customer relationship programs at Saturn, Dell, and Southwest have been attributed to their excellent supply chain management integration (Harvard Business Review 2003). Conversely, industry observers have noted that the failure of many CRM efforts is due to "the propensity of firms to avoid the important 'data transformation and convergence' processes including all transactions, interactions, and networked touch points" (Swift 2002, p. 95). Thus, we expect that firms with greater supply chain integration benefit more from their CRM applications in terms of improved customer knowledge.

H2: Firms with greater supply chain integration are more likely to benefit from their CRM applications and achieve improved customer knowledge.

The Effect of CRM Applications on Customer Satisfaction Customer satisfaction has significant implications for the economic performance of firms (Bolton, Lemon, and Ver- hoef 2004). For example, customer satisfaction has been found to have a negative impact on customer complaints and a positive impact on customer loyalty and usage behavior (Bolton 1998; Fornell 1992). Increased customer loyalty may increase usage levels (Bolton, Kannan, and Bramlett 2000), secure future revenues (Rust, Moorman, and Dick- son 2002), and minimize the likelihood of customer defection (Anderson and Sullivan 1993; Mithas, Jones, and Mitchell 2002). Customer satisfaction may also reduce costs related to warranties, complaints, defective goods, and field service costs (Fornell 1992). Finally, in a recent study, Anderson, Fornell, and Mazvancheryl (2004) find a strong relationship between customer satisfaction and Tobin's q (as a measure of shareholder value) after controlling for fixed, random, and unobservable factors.

Customer relationship management applications are likely to have an effect on customer satisfaction for at least three reasons. First, CRM applications enable firms to customize their offerings for each customer. By accumulating information across customer interactions and processing this information to discover hidden patterns, CRM applica-

202 I Journal of Marketing, October 2005


tions help firms customize their offerings to suit the individual tastes of their customers. Customized offerings enhance the perceived quality of products and services from a customer's viewpoint. Because perceived quality is a determinant of customer satisfaction, it follows that CRM applications indirectly affect customer satisfaction through their effect on perceived quality. Second, in addition to enhanc- ing the perceived quality of the offering, CRM applications also enable firms to improve the reliability of consumption experiences by facilitating the timely, accurate processing of customer orders and requests and the ongoing management of customer accounts. For example, Piccoli and Applegate (2003) discuss how Wyndham uses IT tools to deliver a consistent service experience across its various properties to a customer. Both an improved ability to customize and a reduced variability of the consumption experience enhance perceived quality, which in turn positively affects customer satisfaction. Third, CRM applications also help firms manage customer relationships more effectively across the stages of relationship initiation, maintenance, and termination (Reinartz, Krafft, and Hoyer 2004). In turn, effective management of the customer relationship is the key to managing customer satisfaction and customer loyalty.

H3: The use of CRM applications is associated with greater customer satisfaction.

The Mediating Role of Customer Knowledge Although customer knowledge and customer satisfaction by themselves are important metrics for tracking the success of CRM applications, from a theoretical perspective, it is important to consider whether the association of CRM applications with improvement in customer satisfaction is mediated by an improvement in customer knowledge. From a managerial perspective, an understanding of causal mechanisms will shed light on the conditions that facilitate CRM success in terms of customer satisfaction. We posit that the real value of CRM applications lies in the collection and dissemination of customer knowledge gained through repeated interactions. This customer knowledge subse- quently drives customer satisfaction because firms can tailor their offerings to suit their customers' requirements. Pre- vious research provides support for this view. For example, Bharadwaj (2000) notes the advantages of gathering customer knowledge from customer encounters and dissemi- nating this knowledge to employees for cross-selling and forecasting product demand. Bolton, Kannan, and Bramlett (2000) provide empirical evidence that IT-enabled loyalty programs enable firms to gain valuable customer knowledge about customers' purchase behavior. Jayachandran, Hewett, and Kaufman (2004) show that customer knowledge processes enhance the speed and effectiveness of a firm's customer response. Better knowledge of customer behavior enables firms to manage and target customers on the basis of evolving service experiences rather than stable demographic criteria, which increases the perceived value of the firm's offering and decreases the chance of loyal customers defecting to the competition. Firms also derive a competitive advantage by making cumulative customer

knowledge available to their customers to help those customers manage their internal operations using information from the firm (Glazer 1991). As the preceding discussion suggests, better customer knowledge facilitated by CRM should enable a firm to improve its customer satisfaction. Therefore, we posit that the effect of CRM applications on customer satisfaction is mediated through customer knowledge.

H4: Customer knowledge mediates the effect of CRM applications on customer satisfaction.

Because this research studies the effect of CRM applications on customer satisfaction, we control for other relevant variables to account for alternative and complementary explanations. We control for firms' aggregate IT investments because such investments influence perceived quality, perceived value, and customer satisfaction (Prahalad, Krishnan, and Mithas 2002). We control for firm size, which may influence a firm's ability to benefit from CRM investments as a result of organizational inertia and a greater potential in large organizations for leveraging slack resources. Finally, consistent with previous research, we control for sector differences (manufacturing versus services), which may affect gains in customer knowledge and customer satisfaction.

Research Design and Methodology A major strength of this study is its use of data on key independent and dependent variables from separate sources to avoid common method bias. We obtained the CRM and IT- related data from InformationWeek, a leading, widely circu- lated IT publication in the United States. InformationWeek collected this data by surveying the top IT managers at more than 300 large U.S. firms during the 2001-2002 period. InformationWeek is considered a reliable source of information, and previous academic studies have used data from InformationWeek surveys (Santhanam and Hartono 2003). We collected customer satisfaction data (American Customer Satisfaction Index [ACSI]) that was tracked by the National Quality Research Center (NQRC) at the Uni- versity of Michigan to obtain an archival measure of customer satisfaction for the firms common in the Information- Week data and the NQRC database.

Variable Definition ACSI. The ACSI is considered a reliable indicator of a

firm's customer satisfaction, and the data have been used in several academic studies in the accounting and marketing literature (e.g., Anderson, Fornell, and Mazvancheryl 2004; Fornell et al. 1996).

Customer knowledge (CUSTKNOW). Customer knowledge is a binary variable for which 1 indicates that a firm has gained significant knowledge about its customers from its customer-related IT systems, and 0 indicates that a firm does not perceive any gains in customer knowledge from its customer-related systems.

CRM applications (CRMAPLC). This variable encompasses both the legacy IT applications (i.e., the applications that firms developed before modern CRM applications were

Customer Relationship Management Applications 1203


introduced) and newer IT applications (i.e., the integrated suite of marketing and sales applications developed by CRM and enterprise resource-planning vendors). We measured the first component of CRM applications (legacy customer-related IT applications) using a 12-item summative index that indicates the deployment of IT systems to support business processes associated with customer acqui- sition and disposal of a firm's products and services. The specific IT systems covered by this scale are related to product marketing information, multilingual communica- tion, personalized marketing offerings, dealer locator, product configuration, price negotiation, personalization, transaction system, online distribution and fulfillment system, customer service, and customer satisfaction tracking. We measured the second component of CRM applications (modern CRM systems) using a binary variable (1 = the firm has deployed modern CRM systems, 0 = the firm has not deployed modern CRM systems). We added these two components of CRM systems after standardizing the legacy CRM component (mean = 0, standard deviation = 1). Thus, the variable (i.e., CRM applications) provides greater weight to modern CRM systems but also captures the deployment of legacy customer-related IT applications. Overall, this variable measures a firm's sophistication in managing customer-related information.

Supply chain integration (SCMINTGR). This variable refers to the extent to which a firm's suppliers and partners are included in its electronic supply chain and how much access they have to the firm's customer-related data or applications. It consists of a five-item summative index that describes whether a firm provides its suppliers with electronic access to the following types of application or data: sales forecasts, marketing plans, sales or campaign results, customer demographics, customer loyalty, and satisfaction metrics. We used the standardized (after standardization, mean = 0, standard deviation = 1) value of this variable in our estimation for easier interpretation of the results, partic- ularly because we also investigate the interaction of this variable with CRM applications.

IT intensity (ITINVPC). This variable refers to the level of IT investment as a percentage of the firm's sales revenue.

Industry sector (MFG). This indicator variable represents whether the firm's offering is primarily a good or a service (1 = good, 0 = service).

Firm size (SIZE). This variable is the natural log of the firm's sales revenue.

Empirical Models Because the dependent variable (i.e., customer knowledge) appears as a binary choice, the ordinary least squares (OLS) approach for modeling the binary dependent variable is not appropriate because of heteroskedastic error distribution. A linear model may result in predicted probabilities less zero or greater than one. In addition, a linear model does not allow us to consider the nonlinear nature of the effect of independent variables on the binary dependent variable. To overcome these estimation problems inherent in the OLS approach, we conducted our analysis for this model using the probit approach with the following specification:

204 / Journal of Marketing, October 2005

(1) Probability(CUSTKNOW = 1) = (1)(P10 + 1312ITINVPC + f3i3MFG + 1314SIZE + (315SCMINTGR + 1316CRMAPLC x SCMINTGR + c),

where his are the parameters for the respective variables, and 1 denotes the normal cumulative distribution function (the area under the normal curve).

We used the OLS approach to estimate the customer satisfaction model because the ACSI is a continuous measure of customer satisfaction.

(2) ACSI = (1-20 + 1322ITINVPC + 1323MFG + 1324SIZE + (325SCMINTGR + [326CUSTKNOW + E).

The sample size for Equation 1 is 360, and the sample size for Equation 2 is 40. Table 1 shows the results of empirical estimation of the models in Equations 1 and 2.

Results Consistent with H1, we find that CRM applications are positively associated with an improvement in customer knowledge (Column 1 of Table 1; 1311 = .280, p < .001). Because the probit model is inherently nonlinear, we interpret the effect of each individual variable, holding all other variables at their mean values.

In H2, we posit a moderating effect of supply chain integration on the relationship between CRM applications and customer knowledge. We find support for this hypothesis because the joint hypothesis test for the terms involving CRM applications and the interaction of CRM applications with supply chain integration is statistically significant. This result suggests that CRM applications are likely to have a greater association with customer knowledge when firms are electronically integrated in their supply chain and share customer-related data with their supply chain partners.

We also find support for H3, which posits a positive association between CRM applications and customer satisfaction (Column 2 of Table 1; 1321 = 1.266, p < .069). In H4, we suggest that the association between CRM applications and customer satisfaction is mediated by the effect of CRM applications on customer knowledge. We used the Sobel test to assess this mediation effect (Baron and Kenny 1986). We find evidence for the indirect effect of CRM applications on customer satisfaction mediated through customer knowledge (1326 = 4.307, p < .028). This result implies that, holding other factors constant, firms that report an improvement in customer knowledge due to their customer-related IT systems have ACSI scores 4.307 points greater than firms that report no gains in customer knowledge following investments in CRM applications. Because the coefficient of the CRM applications variable is statistically significant in Column 2 of Table 1, our results suggest partial mediation and imply that CRM applications may also have a direct effect on customer satisfaction.

The results showing the effect of control variables on customer satisfaction also provide useful insights. Note that


TABLE 1 Customer Knowledge and Customer Satisfaction Models

Model 1: Improvement in Customer Knowledge Model 2:

Dependent Variable (Probit) ACSI (OLS)a

.280*** CRM applications 1311 . P21 1.266" (.001) (.069)

IT investments as percentage of revenue P12 .042* 022 -.437"" (.056) (.011)

Manufacturingb 1313 .011 1323 7.420*** (.474) (.000)

Firm size 1314 .254*** 1324 -.987"" (.009) (.046)

Supply chain integration 1315 .011 1325 -.555 (.453) (.166)

Interaction term (CRM x supply chain integration) 1316 .120* (.084)

Improvement in customer knowledge 1326 4.307** (.028)

Constant R10 .488** 1320 73.333m (.018) (.000)

Observations 360 40 Overall fit x2 28.06 R2 .661 *p < .10 (one-tailed test). **p < .05 (one-tailed test). "*"p < .01 (one-tailed test). awe also conducted an additional analysis that controlled for the ACSI score before CRM implementation, and we obtained broadly similar results.

bWe also estimated models with more detailed industry classification, and our primary results remain unchanged. Notes: p values are shown in parentheses.

when we control for the presence of CRM applications, the effect of IT investments on customer satisfaction is negative and statistically significant (Column 2 of Table 1; R22 = -.437, p < .011). This result is consistent with the observation that specific IT applications, such as CRM, that are directly involved in business processes affecting the customer experience may be much more effective in improving customer satisfaction than are aggregate IT investments (Mithas, Krishnan, and Fornell 2002). Focusing on CRM applications also avoids aggregation across several IT applications, in which applications may be relevant for customer satisfaction and others may have a negative or zero impact (Banker et al. 2005; Kauffman and Weill 1989; Mukhopad- hyay, Kekre, and Kalathur 1995). Column 2 of Table 1 also shows that, on average, manufacturing firms have greater customer satisfaction than services firms, a finding that is consistent with previous research (Fornell et al. 1996).

Additional Analyses We conducted additional sensitivity analyses to check the robustness of our results. Because Equation 1 uses data from InformationWeek sources on dependent and independent variables, we tested for common method bias using Harman's one-factor test. Because no single factor emerged as a dominant factor accounting for most of the variance, common method bias is unlikely to be a serious problem in the data.

As we previously noted, the variable (i.e., CRM applications) represents a combination of legacy CRM systems and modern CRM applications. We considered whether the use of modern CRM applications (captured by a binary variable in our data set) "causes" customer knowledge. Because a treatment such as CRM is not exogenously assigned to firms, we investigated the sensitivity of our results due to the potential correlation of CRM with unobservable variables that may have affected our findings (Boulding and Staelin 1995; Wierenga, Van Bruggen, and Staelin 1999). We used a matching estimator based on propensity scores to calculate the treatment effect of CRM implementation on improvement in customer knowledge (Heckman, Ichimura, and Todd 1997; Rubin 2003). Using a procedure that Rosenbaum (2002) suggests, we bound the matching estimator to evaluate the uncertainty of the estimated treatment effect due to selection on unobservables.

After matching the propensity scores and thus adjusting for the observed characteristics, we find that the average CRM effect for improvement in customer knowledge is positive and statistically significant. This calculation is based on the assumption that treatment and control groups are different because they differ on the observed variables in the data set. However, if treatment and control groups differ on unobserved measures, a positive association between treatment status and performance outcome would not neces- sarily represent a causal effect (Boulding and Staelin 1995). Given that we already accounted for selection bias due to

Customer Relationship Management Applications / 205


observed characteristics, sensitivity analysis provides an assessment of the robustness of treatment effects due to factors not observed in the data. Because it is not possible to estimate the magnitude of selection bias due to unobservables with nonexperimental (i.e., observational) data, we calculated the upper and lower bounds on the test statistics used to test the null hypothesis of the no-treatment effect for different values of unobserved selection bias.

For firm i, assume that ui is an unobserved variable and that y is the effect of ui on the probability of participating in a treatment. Under the assumption that the unobserved variable u is a binary variable, the following expression can be derived (Rosenbaum 2002):

1/1"- where F = exp(y), i and j are two different firms within a stratum, and TC is the conditional probability (propensity score) that a firm with given observed characteristics will be in the treatment group. If unobserved variables have no effect on the probability of getting into the treatment group (i.e., 7 = 0), or if there are no differences in unobserved variables (i.e., [ui there is no unobserved selection bias, and the odds ratio will be exp(0) = 1. In the sensitivity analysis, we evaluate how inferences about the treatment effect will be altered by changing the values of y and (ui - ui). If changes in the neighborhood of exp(y) = 1 change the inference about the treatment effect, the estimated treatment effects are posited to be sensitive to unobserved selection bias.

We find that improvement in customer knowledge is not sensitive to unobserved selection bias even if the binary unobserved variable makes it twice as likely for a firm to be in the treatment group than in the control group (after we control for several observed characteristics). Overall, these results provide evidence for the robustness of our findings, showing the positive effect of CRM applications on customer knowledge and, in turn, on customer satisfaction.

Discussion and Conclusion Our goal in this research was to study the effect of CRM applications on customer knowledge and customer satisfaction. We developed a theoretical model that posits a mediating role of customer knowledge and a moderating role of supply chain integration in explaining the effect of CRM applications on customer satisfaction. We used archival data on CRM applications and a perceptual measure of customer knowledge on a cross-section of large U.S. firms during the 2001-2002 period. The study's time frame encompasses a period when firms made significant investments in IT, par- ticularly Internet-based and integrated suites of CRM systems. We matched part of this data set with the sample of firms common to the ACSI to study the effect of CRM applications on customer satisfaction.

Consistent with our expectations, we find that CRM applications are associated with a greater improvement in customer knowledge when firms are willing to share more information with their supply chain partners. Our results suggest that the effect of CRM applications on customer satisfaction is mediated by an improvement in firms' cus-

tomer knowledge. These results lend support to our previously developed theory and conceptual framework.

Contributions and Research Implications This study makes three contributions: First, it builds on previous research that links IT systems and customer satisfaction to contribute to the cumulative knowledge in this stream of literature (Balasubramaniam, Konana, and Menon 2003; Brynjolfsson and Hitt 1998; Chabrow 2002; Devaraj and Kohli 2000; Susarala, Barua, and Whinston 2003). More specifically, given the paucity of research on the benefits gained from CRM technology investments, this study augments the understanding of the beneficial effects of CRM applications by relating them to customer knowledge and customer satisfaction at the firm level. We extend previous research on the effect of CRM processes at the customer-facing level in European countries (Reinartz, Krafft, and Hoyer 2004) and at the strategic business unit level in the United States (Jayachandran et al. 2005) by providing answers to strategic questions about the effect of CRM technology investments on customer knowledge and customer satisfaction. By emphasizing the strategic benefits of CRM applications in terms of customer knowledge and customer satisfaction, we provide a complementary view of judging returns from CRM applications by considering non- tangible aspects that are critical for the creation of shareholder wealth (Anderson, Fornell, and Mazvancheryl 2004).

Second, our study points to the importance of customer knowledge as one of the mediating mechanisms that explains the association between CRM applications and customer satisfaction. Although several studies provide conceptual support for the effect of CRM applications on customer knowledge, our study empirically establishes this association. More broadly, our study provides support for the emerging view that IT applications affect firm performance by enabling other business processes and capabili- ties, which in turn may affect firm performance (Mithas et al. 2005; Pavlou et al. 2004).

Third, the results of this study suggest that it is important to account for the effect of factors such as supply chain integration in the evaluation of returns from CRM applications. We find that though CRM applications are associated with customer knowledge and customer satisfaction, they are even more beneficial if firms share their customer- related information with supply chain partners. This result provides empirical support for the impo