Impact of Electronic Commerce Customer Relationship Management on Corporate Success – Results from an Empirical Investigation Nils Madeja, Detlef Schoder WHU, Otto-Beisheim-Graduate School of Management, Chair of Electronic Business {Nils.Madeja|Detlef.Schoder}@whu.edu Abstract In this contribution, we study companies engaging in B2C-E-Commerce and their ability to manage the relationship with their customers. We refer to this as companies’ ECCRM-capability and examine its determinants and the impact that all of these factors have on corporate success in E-Business. We construct a covariance structure- (or LISREL-) model for testing our research hypotheses on a subset of 224 cases from a representative survey of 1,308 general companies. We find that companies’ ECCRM-capability largely relies on data collection and –aggregation as well as managerial competence in planning and implementing these measures. We obtain no evidence that companies’ ECCRM-capability increases their success. While collecting customer data proves to be a strong driver for success, aggregating and condensing this data appears to degrade corporate performance.
1. Introduction Customer Relationship Management (CRM) can be understood as a revolving process during which companies interact with their customers, thereby generate, aggregate, and analyze customer data, and employ the results for service and marketing activities (cf. [19], p. 81 f., [7], [17], or [5]). The motivation for companies to manage their customer relationships is to increase profitability from concentrating on the economically valuable customers, increasing revenue (“share of wallet”) from them, while possibly “demarketing” and discontinuing the business relationship with invaluable customers. Electronic Commerce Customer Relationship Management (ECCRM, cf. [11] or [15]) chiefly relies on Internet- or web-based interaction of companies with their customers. As the term suggests, ECCRM is specifically aimed at supporting electronic commerce, which, in the following, will be understood as the
activities related to initiating, negotiating, and actually executing business transactions online. 1 Since the beginning of the commercial use of the web, ECCRM has been receiving increasing attention from both practitioners (cf. [1], [8], [12] or [14]) and researchers (cf. [16]). As can be seen from a recent overview of existing ECCRM research (cf. [16]), previous studies have frequently focused on technical aspects of corporate web presences (e.g. design, usability, features and the acceptance of web pages) or on marketing issues (e.g. customer behavior, -satisfaction, and -retention as well as trust). In particular, there has been only little work assuming a corporate perspective, examining how, and if at all, ECCRM implementations contribute to corporate success. Also, past ECCRM studies often concentrated on specific industries (e.g. the financial industry) or business models (e.g. web-centric business models). Further, a deficit in broad empirical research with validated constructs and measures has been diagnosed. As a result, the general business value of the ECCRM concept remains difficult to assess, especially for corporate decision makers (cf. [12]). We will address this research gap, presenting the results from an empirical investigation based on a representative survey comprising 1308 cases in the German-speaking market, which constitutes one of the key international E-Business markets. We employ a quantitative, indirect and confirmatory approach, a research methodology that has been almost unused in the study of success factors for electronic business (cf. [16], p. 84 f., and [3], p. 33). The remainder of this article is structured as follows: The research model is developed in the next section, where the methodology and the perspective we assume for our analysis are explained and where the hypotheses are derived from theory. In the subsequent section, the implementation of the model, the survey, the numerical 1
As we employ this broad definition of e-commerce, our study will also include companies who do not actually sell or take orders, but merely initiate and negotiate business transactions online.
model and the results of the statistical analysis are presented. In the fourth section, we interpret the results and discuss the findings, contributions and shortcomings of our research. Finally, we derive implications from our work for further research and for practice.
2. Research Model 2.1. Research Aim and Model Perspective Attempting to fill the existing research gap identified above, the objective of this paper is to contribute to assessing the general business value of ECCRM. Consequently, we must assume an integrated perspective and choose the corporate level as the level of our analysis and the whole company as the object under study. Similarly, we must take a comprehensive view of the dependant, not only concentrating on single facets such as customer loyalty, -retention, brand awareness or the like. As ECCRM constitutes an integrated concept, implementing it in a business comprises, among others, strategic-, organizational- and technological issues and, specifically, moving business processes online that go beyond mere commerce activities. Therefore, ECCRM should be regarded as an E-Business-, rather than an ECommerce concept, if viewed from the perspective of the company using ECCRM (cf.. [19], p. 15). And it also seems appropriate that as the dependant, we employ the broad concept of corporate performance as the result of a company’s E-Business activities, or, in short, corporate success in E-Business. 2 Further, in order to ensure the general validity of our findings, we intend to make no restrictions with respect to industry or size of the companies under study. However, same as in [15], our research focuses on companies primarily operating in the B2C-segment. Finally, corresponding to the purpose of our analysis, we include only directly ECCRM-related concepts in our model in order to obtain a partial-, rather than a total model for corporate success in E-Business.
2.2. The Concept of Corporate Success in E-Business We limit our view on the concept of corporate success in E-Business to the shareholders’ perspective. The concept is viewed and implemented as a complex construct comprising several theoretical subdimensions, such as to accommodate for the major theories on 2 In the following, the terms “corporate success”, “coporate success in electronic business” and “corporate performance” are all used interchangeably.
competitive advantage, value creation and firm performance (cf. [2]): • “Hard factors”: This subdimension reflects economic performance indicated by economic quantities or coefficients, e.g. revenue- and profit growth. • “Soft factors”: This subdimension accounts for a company’s achievement(s) in the relation to or perception by its customers, e.g. increased loyalty, improvement in the corporate image, or increased customer satisfaction. • “Cost reduction”: This subdimension indicates a company’s improvement(s) in process efficiency as well as procurement conditions, e.g. reduced purchasing- or marketing costs, therefore covering firm performance from a transaction-cost economical perspective. • “Innovation”: This subdimension measures to what extent a company has strengthened its competitive position from the perspective of Schumpeterian theory, i.e. by being innovative, e.g. by offering new services and by developing new markets. • “Increased corporate value”: A company’s valuation does not only depend on its economicor overall performance, yet (mostly) on the way it is perceived by third parties such as investors. Therefore, this final subdimension is the broadest and most susceptible to external influential factors (micro- and macro-economic, psychological, etc.).
2.3. Model Structure In order to fulfill our research goal, we construct a model consisting of two interdependent parts: 1. The first part is centered around the concept of companies’ ECCRM-capability, which we define as the ability to classify its new- and existing customers and to deliver individual service efficiently and economically. We examine how companies’ ECCRM-capability depends on a number of influential factors: the use of ECCRMinstruments, e.g. data collection or –aggregation, as well as competence in implementing ECCRM. 2. In the second part, we analyze how companies’ ECCRM-capability, now viewed as an independent construct, and the other influential factors impact corporate success. With this second part of our model, we allow for a direct influence that companies’ ECCRM-related activities and competence may have on corporate success in order to accommodate for different mechanisms
than the indirect effect which is modeled indirectly via companies’ ECCRM-capability. E.g. it is reasonable to assume that a company that collects customer data also utilizes that data to make strategic mid-to-long-term decisions – which may have a great impact on firm performance, rather than just for classifying customers and providing individual service (which is, relatively speaking, a short-term goal). H 2a
companies’ ECCRMcapability
H 3a aggregating and condensing customer data
H 4a
H 5a
H1
H 2b interactively collecting customer data
corporate success in E-Business
H 3b
H 4b H 5b
informed decision against implementing ECCRM competence in planning and implementation
Figure 1. Overview of the research model
2.4. Derivation of Model Hypotheses Figure 1 illustrates the structure of our research model, the independent factors, the concept of companies’ ECCRM-capability and the dependent construct of corporate success in E-Business. It also shows how the hypotheses in our research model are organized: H 1 is the most important hypothesis in our model, postulating a relationship between the two central constructs. Further, corresponding to the model structure motivated above, a pair of hypotheses is motivated on the basis of each influential factor in the following. Our primary hypothesis follows immediately from the introductory literature and from numerous case studies (e.g. cf. [19], p. 79. ff., or [10]). Companies who are able to manage their customer relationships and, thus, (mainly) focus on profitable customers are more successful than companies who, in an economic sense, waste their resources serving unprofitable customers: H 1: Companies’ ECCRM-capability positively contributes to corporate success.3 3 Terms like “A impacts B”, “A contributes to B” or “A influences B” are to be understood as expressing a synchronous dependence of the type “the more A, the more B” or “A corresponds to B”.
As an intermediate step between interacting with its customers and analyzing customer data on the level of individual profiles, a company should gain a unified view on the customer (cf. [7]). Therefore, the aggregation and condensation of customer data is a prerequisite for ECCRM (cf. [21]): H 2a: Aggregating and condensing customer data from different sources increases companies’ ECCRM-capability. H 2b: Aggregating and condensing customer data from different sources positively impacts corporate success. The necessity of collecting customer data for the concept of ECCRM is evident. As known from studies of web-site usability and user interaction, it should be performed online, interactively, and in a feature-rich environment, in order to meet customers’ expectations of a company’s website and therefore increase their convenience and loyalty (cf. [4]). Adams (cf. [1]) even goes as far as recommending subsidized electronic commerce functionalities (online trading in case of the banking industry) as a measure for retaining customers. We can thus postulate: H 3a: Interactively collecting customer data increases companies’ ECCRM-capability. H 3b: Interactively collecting customer data positively impacts corporate success. Same as with every (E-)business concept, ECCRM may not be suitable for every company and therefore, there may be many companies who resist the implementation of ECCRM. Especially in the B2Csegment, a major reason may be the fear of unresolved legal- or privacy issues or even consumer protests. Typically, consumers are very sensitive to infringements upon their privacy and are concerned that profiles based on their personal data, their surfing habits and shopping behavior might be abused by third parties. If consumers feel that a company violates their personal rights, they are likely to turn away from that company. On the one hand, it is obvious that companies who opt against the implementation of ECCRM will have a lower ability to individually address and serve their customers than companies who decide in favor of implementing ECCRM. However, we propose that if companies have made an active and informed decision against the implementation of ECCRM, the expertise gained in the underlying process of information gathering and analysis should be a driver for corporate success:
H 4a: Companies’ active and informed decision not to implement ECCRM decreases their ECCRM-capability. H 4b: Companies’ active and informed decision not to implement ECCRM positively correlates with corporate success. The issues and difficulties in planning and implementing ECCRM have become subject of numerous articles in scientific research (e.g., cf. [20]) and practitioner’s literature (e.g. cf. [14], [8], or [12]). An ECCRM-system must be selected (or built), configured and integrated with existing systems. Moreover, it must be implemented into the business processes of the respective company, i.e. the employees need to be able to use the system and be motivated to actually do so. Therefore, decision makers competent in managing these issues and difficulties are a key factor for the success of their companies success in ECCRM. As these executives should also be able to successfully manage other EBusiness initiatives, they can be regarded as a key factor for their company’s success in E-Business in general. This leads us to our final pair of hypotheses: H 5a: Managerial competence in planning and implementing ECCRM promotes companies’ ECCRM-capability. H 5b: Managerial competence in planning and implementing ECCRM positively correlates with corporate success.
3. Method The research model as presented above is implemented as a structural equation model (SEM) or, more specifically, as a covariance structure model (also known as a LISREL model). As described in the following, this model is to be tested with numerical data obtained from a large-scale survey. After construct reliability and global model fit have been assessed, the numerical results are evaluated as to if they give support for the research hypotheses.
3.1. The Survey The numerical data used in the statistical analysis of this model has been collected in a large survey that was conducted from May to June 2000 and which has been published as the “e-reality 2000 study” in September 2000 (cf. [18]). This survey was targeted at decision makers of companies in the German-speaking area (Germany, Austria, and Switzerland). To gather data,
market research professionals conducted personal interviews with upper- to top-level executives from 1308 companies. The companies for conducting the interviews were explicitly selected according to a superset of company data, such as to render the survey representative with respect to general company size and industry in the German-speaking market. In case that an interview could not be conducted as planned, a replacement was determined from the same superset in order to maintain the representativity of the sample.
3.2. Aggregation of Survey Data Prior to the statistical analysis, the gathered raw data was reduced and condensed to an essential subset as follows: At first, we concentrated only on the cases of those companies in the survey who had identified consumers as their main customer segment, reducing the original data set of 1308 cases to 685.Then, the cases of companies who did not have a web page online at the time of the survey were eliminated, yielding a subset of 378 cases. (Another 100 companies were still planning to launch their site within the next 12 months.) In a third step, we focused only on those respondents who had specified that they had yet gained sufficient online experience such as to provide information on the success of their company’ s electronic business activities, leaving a total of 224 valid cases for the numerical analysis.
3.3. Descriptive Analysis Same as in the original survey, the remaining cases constitute a heterogeneous selection of companies from all industry backgrounds, company sizes, and business models, even if the original claim to be a representative selection for the German-speaking market must be relaxed. The vast majority (95.7%) of companies are traditional “brick-and-mortar-” enterprises, only 2.5% (corresponding to 6 cases) are spin-offs and only 0.6% (corresponding to one single case) are E-Commerce startups. A fraction of 34.0% of the represented companies generate 50% or more of their revenue from selling services, i.e. can be viewed as belonging to the service industry. Another fraction of 17.6% of the businesses generate 10% or more of their revenue online and can thus be regarded as “true E-Businesses” (cf. [2]). Moreover, 60.1% of the remaining 224 companies offer consumers the opportunity to place orders online. Finally, 16.1% (36) of these companies also actively collect and 42.8% (96) aggregate customer data, while 39.9% (89) feel that they can economically provide individual service on the basis of their customer data.
addition of the values of two or three underlying indicator variables for a total of 13 indicator variables. Only the subdimension “increased corporate value” is measured by a single indicator variable. The indicator variables have been derived immediately from the subdimensions as outlined above. With this procedure, we achieve that the subdimensions exhibit three important characteristics at the same time:
3.4. Encoding of Variables and Operationalization of Independent Constructs In the survey, indicator variables have usually been operationalized on an equidistant interval (or Likert-like) scale ranging from “1” (representing strong dissent) to “5” (representing strong agreement). The indicator variable “number of CRM features” in the construct “data collection” has been obtained for every case as the number of checked items from a list of 9 instruments for customer retention, which the respective company employs on its web site. With the exception of the construct for “aggregating and condensing customer data” (H 2a/b), all constructs are implemented as multiitem measures in the model, i.e. as indirectly measured via several indicator variables.
1. With the exception stated above, every subdimension is operationalized as a multi-item measure, accommodating for the theoretical broadness of each topic. 2. Nevertheless, one single scalar value is obtained for every subdimension. This is a prerequisite for descriptive statistics, e.g. comparisons or classification of select cases or groups of cases, and improves clarity. 3. The subdimensions are highly intercorrelated and can therefore be employed as depending upon the same superordinate construct of corporate success in a covariance structure model.
3.5. Operationalization of the Concept of Corporate Success in E-Business Every subdimension of the construct for corporate success consists of a score obtained from the unweighted
Figure 2. Covariance Structure Model and Results of the Estimation zeta CRM abil.
Global Measures of Fit: .88***
.58 .56*** (H 2a) aggregate diff. sources
.43 e21
active data collection
.56 e22
# of loyalty features
.80 e31
due to privacy issues
.75 e32
due to exp. cons. prot.
.85*** SW diff. to choose
.72 e41
1.00#
.66# .75***
data aggregation
ECCRM capability
.61***
.07
prov. effic. # indiv. service
.84*** (H 3b)
.19***
-.09 (H 5b)
.51 e42
high costs f. own dev.
.57*** distrib. data no unif. view
.32
.58 e43
e44
.63***
IT implem. req. reorg.
integr. w. IT is costly
.40
.32 e45
e46
.66*** hi expenses f. int. comm.
.43 e47
ee1
.79# .72 .85*** "soft factors"
ee2
.61*** .38 cost red.
.77# high training expenses
.59 e48
ee3
.58
innovation
zeta success
.57***
"hard factors"
.76***
.71
ECCRM competence
.34***
.76***
corp. success in E-Biz.
.27*** (H 4b)
ECCRM rejection
insufficient functionality
.62
-.26** (H 2b) .18*
.983 .978 .971 .179
e12
.23* (H 1)
.14** (H 5a)
.89#
.72***
: : : :
-.13** (H 4a)
interact. data coll.
.00
.87***
GFI AGFI NFI RMR
e11
.81 .90
.23** (H 3a)
.77 categorize imp. cust.
ee4
.70*** .49 increased comp. value
ee5
Key for Significance Measures: * : α < 0.10 ** : α < 0.05 *** : α < 0.01 # : for model identifiability, this path coefficient was set to 1 in the unstandardized case
construct / influential factor data aggregation and condensation interactive data collection rejection of ECCRM as an informed decision
competence in planning and implementation
ECCRMcapability
corporate success in E-Business
abbreviated wording of the indicator variable
standardized indicator regres- reliability sion (>= 0.4?) weights
“we try to aggregate and condense different sources of information about 1 our customers” “we actively collect data by 0.657 questioning our customers” number of online instruments for 0.746 customer retention/-loyalty
n. a.
n. a.
n. a.
0.660
0.494
0.580
0.873
0.775
0.843
0.880
0.483
0.891
0.885
0.794
0.892
0.861
0.557
0.834
0.556 0.797
“we will not implement. CRM as we 0.867 fear negative consumer reactions”
0.752
“choosing adequate CRM-software is a very complex issue”
0.849
0.720
“standard CRM-software provides only 0.717 insufficient functionality”
0.515
“high costs arise from further develo0.760 ping standard CRM-software inhouse”
0.578
“customer data is usually distributed across several systems and does not 0.565 provide a consolidated view”
0.320
“IT-integration reorganization”
0.566
0.320
“integration of CRM with existing IT0.631 infrastructure is extremely costly”
0.398
“for the implementation, intensive communication is needed to overcome 0.656 internal resistance”
0.430
“CRM implementation requires high expenses for training”
0.765
0.586
“we can categorize important customers with the data obtained”
0.879
0.772
”we can provide efficient individual service with the data obtained”
0.903
0.815
"hard factors"
0.787
0.620
"soft factors"
0.849
0.721
cost reduction
0.613
0.376
innovation
0.759
0.576
increased company value
0.700
0.490
additional
n. a.
0.431
“we will not implement CRM as we 0.893 fear legal issues concerning privacy”
requires
avg. fracfactor Cronbach’s tion of rec. reliability Alpha variance (>= 0.6?) (>= 0.7?) (>= 0.5?)
Table 1. Constructs, indicator variables and reliability measures.
3.6. Statistical Analysis of the Numerical Model On the basis of the aggregated survey data, the coefficients in the covariance structure model are estimated. As the observed variables in the data set lack multivariate normal distribution, the unweighted least squares (ULS) method is employed. Significance values have been obtained from repeated bootstrap analyses (1000 samples). Figure 2 shows the complete covariance structure model and the numerical results of the estimation.
3.7.
Construct Measures
Reliability
and
Local
Fit
Table 1 shows the constructs employed in the structural model and the respective indicator (or manifest-) variables used for their operationalization. Additionally, path coefficients (standardized regression weights) between each latent variable and its manifest variables as well as reliability measures are given. For the latter, minimum values as they are commonly used in literature (e.g. cf. [9], p. 13) are included. Note that no reliability measures can be given for the construct for “data aggregation and condensation” (H 2a/b), because it is measured directly with only one indicator variable and because the measurement error of this indicator variable has been set to zero. All estimated standardized regression weights in Table 1 are highly significant at the 1%-level. Further, a comparison of the reliability measures and recommended minimum values shows that in general and with only a few exceptions, all factors meet and largely exceed the recommended minimum values: While the deviation of the average fraction of recorded variance from the recommended minimum value of 0.5 is negligible for the factor “interactive data collection”, the value of Cronbach’s Alpha for this factor remains somewhat below the commonly demanded minimum value of 0.7 (cf. [13], p. 245). However, due to the importance of the theoretical dimension of this construct for our model and as this construct essentially fulfills all other reliability measures, it is not excluded from our model. Further, as the two middle indicator variables of the factor “competence in planning and implementation” fall somewhat short of the recommended indicator reliability of 0.4, the average fraction of recorded variance stays slightly below the recommended minimum value of 0.5. But again, this discrepancy is not dramatic, and the indicator variables are kept for theoretical breadth of the construct. For the same reason, the component “cost reduction” is kept in the construct for
“corporate success in E-Business”, although it does not meet the recommended indicator reliability of 0.4. With these few concessions, we conclude that our factors constitute reliable constructs within the structural model.
3.8. Measures for Global Model Fit The measures for global model fit included in Diagram 1 mostly suggest that our covariance structure model fits the underlying data quite well. However, while the values for GFI, AGFI and NFI clearly exceed the recommended minimum value of 0.9, a more differentiated discussion of the numerical base of our model is necessary for evaluating the deviation of the RMR value (0.179) from the commonly desired maximum value of 0.1. Two reasons greatly contribute to the height of the RMR value: 1. heterogeneity of the numerical base and 2. missing values in the data set (between 5% and 20%, depending on the construct).4 Therefore, we conclude that our model exhibits sufficient overall numerical fit and need not be rejected. Note that, as the ULS method was used for estimating the model coefficients and as the values of several indicator variables are not distributed normally, certain indices and quality measures for global model fit (e.g. CFI or the chisquare value) are not applicable and have therefore not been calculated. As our model also fulfills the Fornell/Larcker-criterion (cf. [6], p. 46), our factors can be considered to have sufficient discriminant validity (i.e., there are no competing independent constructs), although we must concede that the fulfillment of this criterion is self-evident for the directly measured constructs.
3.9. Results of the Statistical Analysis The numerical results for our research model can be obtained directly from the path coefficients and significance measures in the covariance structure model displayed in Diagram 1: H 1, our primary hypothesis, that companies’ ECCRM-capability positively correlates to corporate success cannot be supported. Considering the large sample size of 224 cases employed for our numerical
4
In case of the indicator variables employed for composing the scores in the endogenous construct for corporate success (originally typically at the level of 10%), missing values were first substituted by mean values before the score values were calculated in order to prevent propagation of missing values into the score values. In case of all the other indicator variables the pairwise deletion method was employed for calculating the matrix of correlations serving as the input for our covariance structure model.
analysis, significance of the path coefficient at the mere 10%-level cannot be considered sufficient. Further, our hypotheses about companies’ ECCRMcapability depending on different related measures and judgments are fully backed by our numerical model: We have obtained strong statistical support (i.e. path coefficient significant at 1%-level) for our hypotheses H 2a and H 5a as well as support (i.e. path coefficient significant at 5%-level) for hypotheses H 3a and H 4a. Further, diverse statistical evidence was obtained for our hypotheses concerning the impact of companies’ ECCRM-related activities and judgments on overall corporate success: While strong numerical support could be found for hypotheses H 3b and H 4b, none could be found for H 5b. In case of hypothesis H 2b, the numerical results even suggest the opposite of the proposed relationship (with a path coefficient significant at the 5%-level), i.e. that aggregating customer data from different sources actually impedes corporate success. Finally, in our covariance structure model 58% of the variance of the construct representing corporate CRM capability are accounted for by the exogenous constructs in the model. Analogously, 71% of the variance of the endogenous construct representing corporate success are explained by the other constructs in the model.
• •
•
• •
sufficient explanatory power and contains relevant factors. No significant contribution to corporate success can be attributed to companies’ ECCRMcapability. Aggregating and condensing customer data, although the key driver for companies’ ECCRMcapability, seems to have a negative direct impact on overall corporate success. Direct collection of customer data combined with the use of online features to increase customer loyalty appear to be core drivers for corporate success in B2C-E-Commerce. The deliberate and informed decision not to implement ECCRM also has a significant and positive impact on corporate success. Considering the fraction of 71% of explained variance in the dependent construct of corporate success in E-Business, it becomes apparent that ECCRM-related factors alone can be interpreted as central determinants of corporate success in B2C-commerce. Such a high fraction of explained variance would even be sufficient for a comprehensive model covering all relevant theoretical dimensions (cf. [3], p. 207 ff.).
4. Critical Discussion
4.2. Interpretation of Selected Results
4.1. Summary and Explanation of the Findings
Not all of our research hypotheses could be backed by the results of our numerical analysis. Especially two issues need further clarification: First, our analysis suggests that there is no significant impact of companies’ ECCRM-capability on overall corporate success is. Possible reasons for this might be: 1. The construct for ECCRM-capability employed in our model exhibits a somewhat low resolution. We have not recorded which special consequences companies draw from their ECCRM-capability, e.g. for actively (de-)marketing customers 2. Similarly, also according to objective criteria, the ECCRM ability alone might not always suffice to ensure corporate success. Possibly, it is rather an enabling business technology that should be combined with other E-Business concepts such as one-to-one marketing or mass customization in order to increase corporate performance. 3. Another reason may lie in the business segment itself: Widely available low- to medium-value commodities (such as the often-cited books, CDs, DVDs or similar consumer products) are typical products to be traded in B2C-E-Commerce. Profit margins and potential savings are generally low.
Our research, based on an indirect, quantitative and confirmatory approach, provides empirical results about the use and business value of ECCRM from a large-scale representative survey. Apart from a number of reliable complex constructs, especially for overall corporate success, we have obtained the following evidence: • B2C-companies gain ECCRM-capability (namely the ability to categorize important customers and to efficiently provide individual service) from interactively collecting, aggregating and condensing customer data and from managerial competence in planning and implementing ECCRM, while a decision not to implement ECCRM (for legal reasons or because of privacy issues) – as one would expect – decreases that ability. • Further, since all of our hypotheses concerning the influential factors of companies’ ECCRMcapability have found numerical support and as 58% of the variance of this construct has been accounted for, we may conclude that our partial model for companies’ ECCRM-capability has
Therefore, in this customer segment there may be no business benefit from selectively concentrating on “economically attractive” customers only. Secondly, although there is strong statistical evidence that aggregating and condensing customer data is a key driver for companies’ ECCRM-capability, our data suggests that it may be even detrimental to overall corporate performance at the same time (cf. path coefficient). Maybe the implementation of the necessary IT infrastructure and business processes requires a considerable investment and organizational effort, which causes a negative direct impact on corporate performance at first (also cf. the discussion of lag problems), while it has a positive indirect impact as it enables companies’ ECCRM-capability (cf. the indirect path via the construct for ECCRM-capability).
4.3. Limitations and Weaknesses of the Research As our survey featured ECCRM as only one of many other E-Business-related issues and concepts, it must be conceded that the resolution and granularity with which the tested constructs have been implemented is partially somewhat low. As mentioned in the discussion of the result for our primary hypothesis H1, ideally, it would have been desirable to implement a more refined and objective measure for assessing how companies’ ECCRM-capability translates into concrete action. Similarly, it would have been helpful to also refine the construct “aggregating and condensing customer data“ in order to record how companies analyze and further process customer data after aggregation. Lag problems might be another limitation of our research. As the implementation of integrated E-Business concepts such as ECCRM usually requires a considerable investment and the accompanying structural changes take time, it also takes time for a net positive effect on corporate success to become measurable. The fact that ECCRM was still in the emerging phase at the time of the survey (especially in the German-speaking market) implies that a considerable fraction of companies in our survey could have been affected by lag problems. Further, the above findings should be interpreted as reflecting an early stage of development of the ECCRM concept. The high fraction of explained variance of the construct for “corporate success in E-Business” in our model must also be reviewed critically. It is important to note that it is not exclusively accounted for by the exogenous constructs in our model, but that (at least large parts of) the same variance can also be explained by other influential factors, i.e. competing constructs. We did not control for these, since our model focuses on the effectiveness of ECCRM as a singular concept.
5. Conclusion 5.1. Suggestions for further Research Weighing the contributions of our study against its limitations and shortcomings, it is clear that our contribution must be viewed as a first-level analysis – as a “snapshot” and not as final empirical evidence. Although it produces some useful insights, it leaves a number of issues open for future empirical research. Some suggestions are: • The survey should be repeated in a similar manner in order to eliminate lag problems and assess how the identified interrelations change with time, as the integrated business concept ECCRM and the market environment matures. • Similarly, it remains to be investigated how the findings and vary in different markets or market segments (e.g. for B2B-enterprises?) or with industry or company size. • Future empirical research should investigate data aggregation and –analysis in ECCRM-implementations in more detail and attempt to provide a more objective, comparable and quantitative measure for companies’ ECCRM-capability. • Finally, it seems important to study the role of ECCRM as an enabler for other integrated EBusiness concepts such as one-to-one-marketing and mass customization.
5.2. Managerial Implications ECCRM is an integrated E-Business concept aiming at increasing revenues from profitable customers, which is supposed to lead to increased profits. Therefore, when making the decision whether to implement ECCRM or not, executives should first of all estimate the potential increase in revenues and profits and then weigh that potential business benefit in against the potentially high implementation costs. Further, in order to deliver maximum benefit, ECCRM probably has to be implemented in combination with other integrated EBusiness concepts such as one-to-one-marketing and mass customization. In any case, managers should ensure that they actually make use of their companies’ ECCRM capabilities to improve everyday business processes, especially online selling. Also, decision makers should concentrate their company’s ECCRM activities on interacting with customers, and not on technological issues. They should focus on directly collecting customer data in combination with deploying online instruments for retaining
customers, which has proven to be the core driver for corporate success. However, when collecting and evaluating customer data from internal sources, they must pay attention to the quality and reliability of these internal sources as well as to the criteria employed for the data evaluation and -analysis. Finally, in some cases, it seems advisable to opt against the implementation of ECCRM, e.g. when managers fear that legal conflicts or consumer protests could arise because online customers’ might feel that their privacy has been violated.
[10] Kannan, V. R.; Tan, K. C.; Handfield, R. B.; and Ghosh, S. (1998): Managing Competition, Quality, Customer Relations, and the Supply Base, and Its Impact on Firm Performance, in K. B. P. (Coordinator), Proceedings of 29th Annual Meeting of the Decision Sciences Institute, Las Vegas, Nevada, 1998, pp. 1259-1261.
6. Acknowledgements
[12] Malis, E. (2000): The CRM Buzz, Manufacturing Systems, May 2000, p. 54.
We would like to thank both the anonymous HICSS reviewers and, especially, the minitrack chairs for their constructive and helpful comments, which have greatly helped to improve an earlier version of this paper.
[13] Nunnally, J. (1978): Psychometric Theory, 2nd edition, New York.
7. References [1] Adams, J. (2000): The hidden opportunity of E-Commerce: e-CRM, Future Banker, Nov. 2000, Vol. 4 Issue 11, p.41. [2] Amit, R.; Zott, C. (2001): Value Creation in E-Business, Strategic Management Journal, Vol. 22, pp. 493-520. [3] Böing, C. (2001): Erfolgsfaktoren im Business-toConsumer-E-Commerce, Gabler Verlag, Wiesbaden. [4] Breitenbach, C. S.; Van Doren, D. C., Value-added marketing in the digital domain: enhancing the utility of the Internet, Journal of Consumer Marketing, Vol. 15, No. 6, pp. 558, 1998. [5] Eggert, A. (2001): Konzeptionelle Grundlagen des elektronischen Kundenbeziehungsmanagements, in: Eggert, A.; Fassott, G. (editors): eCRM – Electronic Customer Relationship Management, Stuttgart 2001, p. 87-106. [6] Fornell, C.; Larcker, D. (1981): Evaluating Structural Equation Models with Unobservable Variables and Measurement Error, Journal of Marketing Research, 18, Feb. 1981, pp. 39-50. [7] Hippner, H.; Martin, S.; and Wilde, K. D. (2001): Customer Relationship Management, WiSt Heft 8, August 2001, pp. 417422. [8] Holden, J. (2001): CRM shouldn't hurt, Sales & Marketing Management, Vol. 153, No. 2, pp. 29-30, 2001. [9] Homburg, C.; Giering, A. (1996): Konzeptualisierung und Operationalisierung komplexer Konstrukte – Ein Leitfaden fuer die Marketingforschung, ZFP Heft 1, 1. Quartal 1996.
[11] Kundisch, D.; Wolfersberger, P.; Calaminus, D.; and Kloepfer, E. (2001): Enabling eCCRM: Content Model and Management for Financial eServices, Proceedings of the 34th Hawaii International Conference on System Sciences, Hawaii, 2001.
[14] J. Orr (2001): Strategic options for CRM: Which way off the roundabout?, Vital Speeches of the Day, Vol. 67, No. 20, pp. 615-618, 2001. [15] Romano, N. C., Jr.; Fjermestad, J. (2001a): Introduction to the Special Section: Electronic Commerce Customer Relationship Management (ECCRM), International Journal of Electronic Commerce, Winter 2001-2002, Vol. 6, No.2, pp. 7 f. [16] Romano, N. C., Jr.; Fjermestad, J. (2001b): Electronic Commerce Customer Relationship Management: An Assessment of Research, International Journal of Electronic Commerce, Winter 2001-2002, Vol. 6, No.2, pp. 61-113. [17] Seybold, P. B. (2001): The Customer Revolution, New York 2001. [18] Strauß, R.; Schoder, D. (2000): e-Reality 2000 – Electronic Commerce von der Vision zur Realität: Status, Entwicklung, Problemstellungen, Erfolgsfaktoren sowie Management-Implikationen des Electronic Commerce, ISBN: 3-00-006870-8, Consulting Partner Group, Frankfurt a.M. [subsequent survey to the “Electronic Commerce Enquête 1997/98”]. [19] Strauß, R.; Schoder, D. (2002): eReality – Das E-Business Bausteinkonzept. Strategien und Erfolgsfaktoren für das EBusiness-Management. F.A.Z.-Institut für Management-, Markt- und Medieninformationen, Frankfurt a. M. 2002. [20] Wilson, H.; Daniel, E.; McDonald, M. (2002): Factors for Success in Customer Relationship Management (CRM) Systems, Journal of Marketing Management, Feb. 2002, Vol. 18, Issue 1/2, pp. 193-219. [21] Yu, L. (2001): Successful Customer-Relationship Management, MIT Sloan Management Review, Summer 2001, p. 18f.