Abstract
This study proposes an e-service resource bundle (E-SRB) as an antecedent of electronic service channel (e-channel) success in small retail service providers. The E-SRB indicates a collection of three resources: e-market acuity, e-IT competence, and e-service agility. Given the interdependence of these three resources in delivering quality e-services, the authors hypothesize about their complementarity and its positive effect on performance. The results of this structural equation modeling using survey data show support for the proposed hypotheses, demonstrating that the E-SRB positively influences e-channel performance. The performance impact is not limited to perceived financial performance but extends to self-reported dollar-based sales and profits. These results have theoretical implications when it comes to linking e-service quality to financial performance. They also carry managerial implications for small-scale e-retailing, where limited resources can seriously impede the full use of the e-channel. One of these implications concerns what resources are necessary and how to allocate them in order to improve an e-service system. In the end, this study suggests that managers should understand the interrelationships that might exist among resources that collectively influence performance.
Keywords
Introduction
The success of the Internet-based economy has rushed small service providers, including local retailers, into adoption of the electronic service channel (e-channel). After adoption, the principal interest of these retailers lies in allocating key resources to increase performance from the channel. Given their limited resources, however, small retail service providers face tremendous challenges in allocating their resources for e-channel performance (Harland et al. 2007; Pfughoeft et al. 2003). In the past, numerous studies of customers’ online shopping experiences have identified e-service quality as an antecedent of e-channel success (Collier and Bienstock 2006; Parasuraman, Zeithaml, and Malhotra 2005; Rowley 2006), but, from an e-channel provider’s standpoint, this approach has not translated into appropriate resource-allocation decisions. In addition, the limited amount of empirically based survey research on small service providers has yielded little reliable advice for practitioners, except for a few generic statements such as “satisfy online customers at a low cost” (Boyer, Hallowell, and Roth 2002; Rowley 2006).
This article proposes bundling of relevant resources for the successful use of an e-channel since doing so offers a feasible way to handle the numerous quality factors that arise in the presentation and delivery of an e-service. For instance, previous studies have identified multiple e-service quality factors within each dimension of the channel attributes, encompassing website design, navigation characteristics, information integrity, ease and usefulness, security, and operational competence (Collier and Bienstock 2006; Fassnaght and Koese 2006; Heim and Sinha 2001; Koufaris 2002; Novak, Hoffman, and Yung 2000; Parasuraman, Zeithaml, and Malhotra 2005; Rowley 2006). Without bundling multiple functional resources, e-service providers might experience great difficulty in finding an efficient way to address the numerous quality-related factors.
Consistent with the resource-based view (RBV), we include intangible assets, such as organizational competence and capabilities within the resource boundary (Barney and Arikan 2001; Dierickx and Cool 1989). The RBV has been largely dedicated to identifying key resources and highlighting their individual effects on firm or system performance. However, when the resources are interrelated, the bundling effect, so-called complementarity, that is, a collective effect that is higher than the sum of the individual effects, appears to be significant (Cassiman and Veugelers 2006; Milgrom and Roberts 1995; Powell and Dent-Micallef 1997; Segars, Grover, and Teng 1998). Examination of the collective effect requires a holistic view of the system, where the subunits are interrelated and, in turn, will reveal a more efficient way to use certain resources (Checkland 1981). Viewing the entire e-service system, we build on the concept of bundling (Macduffie 1995) and examine the complementary effect of interrelated resources on e-channel performance.
Relevant to this study, our comprehensive literature review found two separate studies that examined the effect of multiple resources within the context of online retailing, although they did not focus on small firms. Zhu’s (2004) study on retailers in Fortune 1000 and the Ward’s Business Directory confirm that, together, direct resources for online business and Information Technology (IT) infrastructure positively influence sales and cost reductions. This study successfully demonstrates that the e-channel significantly contributes to corporate performance, but the findings are hardly applicable to small firms, in which limited resources impede the creation of a solid IT infrastructure (Harland et al. 2007). Zhuang and Lederer (2006) also examined the resources of online retailers of various sizes, with average staff strength of 408. They found a significant relationship to performance from complementary resources, but their structural equation model is, in fact, a direct-effect model that examines individual—rather than complementary—effects (Segars, Grover, and Teng 1998). In addition, using only perceptual performance leads to limited managerial implications.
The current study proposes bundling of the following three resources: e-market acuity, e-IT competence, and e-service agility. We refer to this resource bundle as “e-service resource bundle,” or simply “E-SRB,” and we argue that E-SRB is essential to the success of e-service firms. E-market acuity represents an e-channel provider’s ability to see the competitive environment clearly, thereby gaining an understanding of its customer needs. E-IT competence indicates the ability of an online firm to understand and utilize IT tools and processes within the web-based service process. E-service agility refers to the level of operational competence that facilitates a rapid response to uncertain operational situations in a seamless manner, extending e-service encounters to actual delivery of services (Boyer, Hallowell, and Roth 2002; Raman, DeHoratius, and Ton 2001). This study argues that the three different functional resources, which are useful to deliver quality e-service encounters to online customers, collectively bring enhanced e-channel performance to small retailers.
The need for bundling the three different resources can be demonstrated by their interdependence. For example, understanding customer needs requires IT skills for monitoring online customer behaviors. Data mining applications are commonly used for tracking customers’ interest areas on a particular website (Padmanabhan, Zheng, and Kimbrough 2006). Such consistency between e-market acuity and e-IT competence will increase the level of customer satisfaction. Real estate websites provide another example in that many now include a simple IT application in the form of a video that shows both the interior and the exterior of a house that is for sale. This application satisfies the market’s needs for a realistic online shopping environment (Richtel and Tedeschi 2007) and makes the firm’s operations agile by reducing the number of inquiries from less-serious customers (Cho and Menor 2010). The small local florists who inspired this research emphasized the importance of the complementary relationship of e-service agility, e-market acuity, and e-IT competence since this is the relationship that delivers what the customer wants (floral designs), on time, with quality service encounters suitable for an important event such as valentine’s day, a birthday party, a wedding, or a funeral.
This research has two focused objectives: To empirically validate E-SRB as a bundle of the three resources and to test its effects on e-channel performance. In working toward these objectives, this study makes several important distinctions from the existing e-service studies. First, to the best of our knowledge, this is the first research that uses channel-specific financial performance as a way to examine the exact resources required for design and delivery of e-services, which represents the initial step in understanding the e-service system. Second, we suggest an antecedent of e-channel success from the service provider’s standpoint. We approach online services with resources, not with e-service quality, as many previous studies have done from a customer’s standpoint (Rowley 2006). A particular resource and e-service quality represent two distinct managerial elements, but we believe that they should be interrelated in terms of predicting e-channel performance, and this belief provides the basis of our theoretical development. Third, we find the source of e-channel success from a holistic view of the e-service system, which is useful in examining interrelated system elements (Checkland 1981).
This exploratory study on resource bundling investigates small enterprises only, providing both relevance and rigor. Of all firms in the United States, almost 98% have fewer than 100 employees, and almost 90% have fewer than 20 employees (U.S. Census Bureau 2004). Previous case studies, however, have revealed that limited resources—the major constraint of all small firms—significantly impede development of long-term vision on the e-channel (Chen et al. 2003; Harland et al. 2007; Levy and Powell 2003). Separate attention must be paid to small service providers; otherwise, any analytical tool can distort the results by averaging out the contextual difference.
In the following section, we introduce three resources and propose hypotheses for e-channel success using the concept of bundling. The following Methodology section describes the details of the data collection process, measurement validation, and structural equation modeling (SEM). The results of our research will be presented in a subsequent section. Finally, the Discussion section will include both the theoretical and the managerial implications that we obtained from the research findings, along with suggestions for future research.
Theoretical Development
Use of the Internet forces a change in the focus of management attention from goods to services, efficiency to satisfaction, and tangibles to information (Rust and Lemon 2001). As firms make ever-increasing use of the website as a service delivery channel, e-service quality becomes an increasingly critical competitive dimension (Rowley 2006). Accordingly, the success of network-based businesses depends largely upon resources that are useful in the delivery of quality e-service encounters.
The RBV connects firm resources with competitive advantage by claiming that the unique profile of internal resources—including tangible and intangible assets, such as facilities, processes, competence, capabilities, and knowledge—determines the competitive position of the firm in the market (Barney and Arikan 2001; Wernerfelt 1984). The competitive advantage of producing superior performance is sustained when the firm implements a value-creating strategy using valuable, rare, inimitable, and nonsubstitutable resources (Barney 1991). The RBV raises the insight that, due to the inevitable variations in their resource profiles, small retailers differ in terms of their ability to cost-effectively provide quality e-services.
We selected three resources: e-market acuity, e-IT competence, and e-service agility. The rationale behind the selection process indicates the quality dimensions that we cited earlier, such as website design, navigation characteristics, information integrity, ease and usefulness, security, and operational competence, converge to the functional areas of these three resources, that is, identifying the market needs, using IT tools to address the needs on the website, and performing agile service operations. From a strategic standpoint, these three resources combine to play a key role in coordinating major service system elements, such as target market, service concept, and service delivery system (Heskett 1986; Sasser, Olsen, and Wyckoff 1978; Roth and Menor 2003). If these three resources are indeed complementary, successful retailers will need to acquire the complete set, with the result that the collective effect will be higher than the sum of the individual effects (Macduffie, 1995; Segars, Grover, and Teng 1998).
E-Market Acuity
Multiple service quality factors demonstrate that online customers have diverse expectations from their online shopping. From a service provider’s standpoint, one premise in delivering quality e-services points to the organization’s ability to read the market and customer needs. We suggest e-market acuity as an essential resource in e-service provision since superior market understanding translates into effective service design and delivery practices (Bitran and Pedrosa 1998; Sasser, Olsen, and Wyckoff 1978).
The concept of acuity was suggested by Stalk, Evans, and Schulman (1992) as a leading capability within the era of “capabilities-based competition.” These authors forecasted a more dynamic business environment in the coming decade in terms of short product life cycle, technology advances, and globalization, which would demand a higher level of market acuity. In view of that, Roth and Jackson (1995) understood market acuity as a firm resource and first operationalized the concept within the banking industry. Their empirical study uncovered positive relationships between market acuity and firm performance, as mediated by service quality. Menor and Roth (2007) also empirically verified that market acuity is a critical dimension of new service development. In a similar concept, market orientation has been used to indicate an organization-wide culture or climate that enriches market intelligence (Kohli and Jaworski 1990; Narver and Slater 1990), but the concept of market acuity represents a capability that actually enables a firm to obtain market intelligence (Stalk, Evans, and Schulman 1992).
We define market acuity in an e-service system, so-called e-market acuity, as the ability to see the competitive environment clearly in order to design e-service offerings that fit customers’ evolving needs and preferences. Possessing e-market acuity is particularly important for an e-service firm because online customer needs and preferences vary dramatically alongside changes in the economical and technological environments (Collier and Bienstock 2006; Iqbal, Verma, and Baran 2003). In addition, e-service providers should consider customer-driven variance from demographic and psychological characteristics that point to different e-service quality factors (Venkatesan, Kumar, and Ravishanker 2007). For example, Parasuraman (2000) classified customers into various groups, according to their different levels of technology readiness, such as explorers, pioneers, skeptics, paranoids, and laggards. Importantly, these variables should fit the given target market segment right from the earliest stages of new service design and development (Bitran and Pedrosa 1998). There is another reason that it is important to accurately understand the online market and its customer: Customers can easily defect from a certain online store without face-to-face interaction. The volatility of customer loyalty within the virtual environment places a constant pressure on e-service providers to introduce “stickiness” (Boyer, Hallowell, and Roth 2002; Reichheld and Schefter 2000) to their e-services, which requires a high level of e-market acuity.
E-IT Competence
IT represents an essential resource for a firm that makes network-based transactions. Previous empirical studies on the performance implications of this particular resource reveal some interesting results by pointing to a higher association between organizational competence or capabilities in the use of IT—rather than simply the existence of an IT system—and firm performance (Powell and Dent-Micallef 1997; Teo and Ranganathan 2003). This increased degree of association derives from the fact that supporting elements in IT competence or capability, such as IT skills, knowledge, experience, and infrastructure add unique properties to the easily imitable IT system itself (Bharadwaj 2000; Tippins and Sohi 2003).
Among the many elements that collectively constitute a firm’s IT resource, our concept of e-IT competence includes IT skills and knowledge in using IT tools and processes. We define e-IT competence as the ability to understand and utilize the IT tools and processes that are necessary to maintain online operations and functionality in the delivery of service offerings. Previous studies demonstrate that IT skills and knowledge represent a significant asset to small firms. Harland et al. (2007) found that many small firms do not have IT skills and knowledge and are therefore likely to have nothing more than a simple web presence. Further, Levy and Powell (2003) explain that while a simple web presence does provide the most elementary stage of e-channel use among small firms, retailers can advance their e-channel to a full-grown e-business solution by adopting and applying a variety of IT skills and knowledge. Outsourcing can be a good strategy for small firms that do not have enough e-IT competence or that simply want to focus on their own area of expertise. In such cases, the e-IT competence of the IT partner should be carefully evaluated before the partnership begins, since this factor ultimately influences the speed and quality of cooperation on the technical issues, such as adding new web functions or updating the company website (Chen et al. 2003).
While conventional IT competence or IT capabilities emphasize the coverage and efficiency of an information network (Bharadwaj 2000), e-IT competence focuses on IT skills and knowledge that design and deliver online services for use on the website where the actual customer interaction occurs (Froehle and Roth 2004; Pfughoeft et al. 2003). The quality of this interaction directly influences a customer’s online shopping experience, which makes this resource a critical part of the resource bundle. For instance, a service provider’s e-IT competence can deeply influence the online customer’s psychological state during the interaction by manipulating the visual and navigational characteristics of a website. The concept of “flow” dictates that a smart website captures the web-user’s “holistic sensation” (Novak, Hoffman, and Yung 2000) during e-service encounters. Recently, web 2.0 technologies that use blogs and wikis have facilitated more interactive e-service encounters, and many retail websites now welcome their customers to social networking sites such as Facebook, Twitter, or YouTube (Cho and Menor 2010).
E-Service Agility
Within the operations management area, definitions of agility have converged largely upon two different meanings. The first view of agility focuses on rapid response or adaptation to environmental changes, emphasizing the speed of the response to both anticipated and unexpected changes (Dove 2001; Zhang and Sharifi 2000). The second view of agility focuses on the strategic flexibility of an operations function that allows simultaneous pursuit of various competitive priorities such as cost, quality, delivery, and flexibility (Fliedner and Vokurka 1997; Narasimhan and Das 1999).
Online retailers encounter a variety of expected and unexpected operational situations that challenge a manager’s goal to deliver the order on time, to rapidly respond to customer inquiries, and to quickly recover any service failures (Raman, DeHoratius, and Ton 2001). Multiple studies on e-service quality demonstrate that time-based diligence along the first stream of agility research is critical in delivering customer value in Internet retailing (Collier and Bienstock 2006; Heim and Sinha 2001; Rowley 2006). This article refers to the operational agility as e-service agility and defines it as the ability to rapidly respond to dynamic operational situations in a seamless manner when delivering services to online customers. The term “dynamic operational situations” refers to both anticipated and unanticipated situations within e-service delivery operations that arise due to operational uncertainties and service failures.
There is much evidence that e-service agility greatly influences customer evaluation of the online store. An examination of the customer complaints posted on evaluation websites such aswww.bizrate.com (where customers judge and express their shopping experience with different online stores) reveals that delayed delivery of ordered items negatively influences customer intention to use the store again, which is consistent with the findings by Heim and Sinha (2001). Cho, Im, and Hiltz (2003)’s study on customer complaints about online shopping experiences also found that the most common issues include delivery after the due date and nonresponsiveness to the operational failures, both of which will negatively influence customer loyalty (Collier and Bienstock 2006). In addition, previous studies regard firm activities in the service recovery stage as independent services because the quality of the recovery activities is still measurable by the gap between expectation and perception, and their quality outcomes have implications for customer satisfaction and loyalty (Parasuraman, Zeithaml, and Malhotra 2005). Importantly, e-service agility is a common requirement for service delivery and service recovery processes (Holloway and Beatty 2003).
E-SRB
The three e-service resources have been individually discussed in the previous sections, but we propose their complementary effect on performance as a bundle. Resources in a bundle are internally consistent and thus are correlated (Macduffie 1995). Under the circumstances, the presence of all three resources becomes important since any change in one of the three can influence the entire bundle. Consequently, a lack of any one of these resources can easily serve to dissatisfy an online customer, which will negatively influence the performance.
An e-service encounter in online retailing includes the entire interaction period from “initial landing on the website until the final product has been delivered and is fit for use” (Boyer, Hallowell, and Roth 2002, p. 178). This description indicates that customer perception of e-service quality is influenced by multiple factors along the course of the interaction. As a result, a service failure in any moment of the e-service encounter can negatively influence customer evaluation of the entire online experience. Given the need for multiple resources, the three interrelated resources in E-SRB are proposed to work together to deliver quality e-services to online customers, covering both the order procurement and the fulfillment stages of an e-service encounter.
Using a holistic view, existing service strategy frameworks explain a service system from its major elements, such as target market, service concept, and service delivery system (Heskett 1986; Roth and Menor 2003; Sasser, Olsen, and Wyckoff 1978). These frameworks claim that the success of the service process is directly linked to the degree of coordination between the system elements. E-SRB, an internally consistent collection of the three resources, provides an efficient way for online retailers to coordinate the major service system elements, which has implications to e-service quality (Roth and Menor 2003). E-market acuity enables a service provider to understand the target market, which is the key to developing effective service concepts (Menor and Roth 2007; Rosenzweig and Roth 2007). E-IT competence enables a company to translate its service concepts into technological tools and practices, thereby projecting the concepts to the e-service delivery system, that is, the website (Harland et al., 2007; Zhuang and Lederer 2006). Finally, e-service agility represents the operational competence that seals the gap between the delivery system and the target market during the actual delivery of services. Together, these three resources constitute a complete set of tools with which a service firm can align the three major elements in an e-service system.
This interdependence among the three resources explains why all three resources are required. Resource interdependence generates a situation where the system optimum is superior to the sum of the local optimums (Checkland 1981, 2000). The systemic approach provides an underpinning of the bundling of the interdependent components. By defining a system as a conglomerate of interrelated and interdependent subunits, the systemic approach pinpoints the best way to benefit the system in terms of understanding and managing the interdependence (Cleland and King 1972; Wilson 1984). As it applies to this study, the coexistence of all three resources is more likely to result in an optimum performance of the entire e-service system, which is an outcome that would not be possible if any one of those resources was missing.
This requirement of the complete set of interdependent elements lines up well with the concept of complementarity, which occurs when the value of one element increases under the presence of others (Menor, Kristal, and Rosenzweig 2007; Milgrom and Roberts 1995; Powell and Dent-Micallef 1997). Previous studies present bundling or complementarity as being manifested by covariation of the subject elements since a change in one element triggers a change in the others (Macduffie 1995; Menor, Kristal, and Rosenzweig 2007). Figure 1 illustrates our research model, depicting the relationships among the relevant constructs. Testing the hypotheses requires the second-order factor model because covariation of three resources, as first-order factors, creates a higher-level latent variable, which is the E-SRB. Segars, Grover, and Teng (1998) refer to the second-order model as a coaligment model—as opposed to a direct-effect model—because the second-order construct is required in order to fully understand the performance effect of the first-order constructs.

Proposed research model.
Our first three hypotheses posit E-SRB as a bundle of the three resources. We hypothesize the resource complementarity using one statement, which includes three different relationships to test.
Hypotheses 1-3 (H1-3). E-SRB is reflected by and is positively associated with e-market acuity, e-IT competence, and e-service agility as complementary dimensions.
Hypotheses 4-6 (H4-6). E-SRB has a positive effect on e-channel performance.
While the service strategy literature and the systemic approach show support for the positive relationship between the resource complementarity and e-channel performance, the RBV provides theoretical justification. It explains that a group of complementary resources increases the likelihood of presenting the characteristics of a key resource—valuable, rare, inimitable, and nonsubstitutable—albeit its individual components have fewer of those characteristics (Barney 1991; Powell 1992). Thus, a positive relationship between E-SRB and performance is expected.
We use three channel-specific performance measures, which are perceived financial performance (Hypothesis 4), e-sales (Hypothesis 5), and e-profit (Hypothesis 6). Perceived financial performance may not necessarily match with dollar-based performance, but it is useful in understanding how much of the expected performance is actually achieved (Wade and Nevo 2005-2006). E-sales and e-profit represent self-reported dollar-based sales and profit from the e-channel, respectively.
Research Method
Sample Selection
To empirically test our hypotheses, a survey method that would collect data from small service providers was required. We found that flower retailers who used a website as an e-channel—referred to here as “e-florists”—represented an ideal sample for this study for several reasons. First, florists operated a small business, typically with fewer than 20 employees. An industry report showed that florists constituted one of the best examples of a small business in terms of the fragmented market structure, which was characterized by the presence of fewer giant players compared to other industry sectors (First Research 2006). Another attractive feature of studying this sector came from the fact that many of the players had undergone, or were undergoing, e-channel adoption. By studying the antecedent of e-channel success in the sector, we expected to provide the industry with a useful prescription for e-channel implementation after the initial adoption of e-channel. While the selection of only one industry sector can reduce generalizability, in this case, it was a useful choice because it allowed us to control for externalities from the product category and scalability of the product, which can significantly influence customer behaviors with regard to channel selection (Konuş, Verhoef, and Neslin 2008) and largely determines process options (Boyer, Hallowell, and Roth 2002; Hallowell 2001), respectively. For these reasons, the survey was conducted targeting e-florists, all of whom were located in North America.
Data Collection
A reliable sampling frame was not available, so we developed our own, based on the information in the online Yellow Pages (http://www.yellowpages.com/ and http://www.yellowpages.ca/). These websites provided comprehensive lists of florists and their contact information more than any other source. Our researcher visited each e-florist’s website and included the site in the sampling frame only when it had an online transaction function. After developing the sampling frame, we adopted a web-based survey using the dedicated software, Ultimate Survey Enterprise.NET v3.0.10, which let us efficiently manage the survey process, including sending invitations and receiving and saving responses. Two invitation-sending channels were available: e-mail and an inquiry window. As a webpage that is dedicated to receiving customer inquiries, the inquiry window was the only channel by which to reach those e-florists who did not reveal their e-mail addresses. Following Dillman’s (2007) total design method, we sent an initial prenotice to potential respondents, followed a few days later by an invitation that contained a hyperlink to the survey website. A total of three invitations were sent in 2- or 3-week intervals.
We collected contact information for 1,264 North American e-florists from all over North America, all of whom were listed in the online Yellow Pages and had effective electronic channels. The Canadian province of Ontario had the most e-florists (195), followed by the state of California (188) and Texas (84) in the United States. The number of e-florists largely differed from the total number of florists because e-florists made up a very small portion of the entire florist sector. Our recent revisit to the Yellow Pages, however, confirmed that Ontario and California had the most florists in Canada and United States, respectively, consistent with the numbers of e-florists above. In addition, we found that the portion of e-florists among total florists had significantly increased since we conducted the survey in 2007.
Among the total 1,264 North American e-florists in the sampling frame, 149 responded. One of these replies came from a pure online player and it was removed from the sample due to the potential bias from this contextual difference. With 148 responses, the total response rate was 11.7%, with response rates of 6.89%, 3.01%, and 1.82% from the first, second, and third rounds of invitations, respectively. Given the relatively low response rate, we tested nonresponse bias using the extrapolation method (Armstrong and Overton 1977), which compares early and late responses based on the Armstrong and Overton’s finding that subjects who answer later are more like nonrespondents. The multivariate analysis of variance (MANOVA) found no significant difference between the respondents and the potential nonrespondents in the levels of three resources, firm performance, and control variables such as organizational size and online experience (Wilks’s λ = .602, p = .642).
As summarized in Table 1, among the 148 responses, 93 came from American florists (63%) and 55 from Canadian florists (37%). Our further analysis showed that the 93 responses from the United States came from all over the country. Using the four conventional blocks, we traced the location of the respondents—Western (22 responses), Central (32 responses), Northeastern (17 responses), and Southeastern (22 responses). In the case of Canadian responses, most of them came from the province of Ontario (44 responses). Given the significantly higher response rate from the Canadian province of Ontario—probably because the participants recognized the name of the research institute in the survey instrument, which was located in their home province—we compared between American and Canadian responses on the three resources, firm performance, and the control variables, using MANOVA and confirmed no significant difference (Wilks’s λ = .862, p = .317). As expected, most respondents were owners (86.5%). The average number of full-time employees was 4.1 (SD = 4.2), with a maximum of 30, only 2 among the 148 responded e-florists had more than 20 full-time employees. The average length of a company’s online experience was 4.7 years (SD = 2.9), with a maximum of 14 years.
Sample Characteristics
Construct Measurement
Before the survey, measurements of constructs were developed under rigorous reliability and validity assessment. While previous research had used the concepts of market acuity, IT competence and agility in non–e-service contexts, we needed to develop new items that would be suitable for an e-service context since newly defining and measuring a construct is recommended when the object, attribute, or rater changes (Rossiter 2002).
In developing the measurement scales, we followed Menor and Roth’s (2007) two-stage approach, which uses validity and reliability assessment in both an item level and a scale level. As suggested, the item-level assessment began with a comprehensive literature review and expert consultation in order to concretize adequate content validity (Malhotra and Grover 1998). In the process, we interviewed the owners of three local florists who were highly familiar with the business. The interviews focused on the three resources as well as the general business environments of each florist. The interviews revealed that outsourcing was a popular practice in the IT area, mainly due to the fact that many of the florists did not have sufficient IT skills to develop and maintain a website. For this reason, we developed the measurement items of e-IT competence to include that of the IT partners. Based on these interviews and our literature review, we developed an initial pool of measurement items. We conducted two rounds of Q-sort over the initial item pool by 12 academicians, all of whom were versed in empirical scale development. After the Q-sort, we visited seven other local florists as a convenient subsample of this study in order to confirm the items, an action that resulted in only a few small changes.
Objective performance data for e-florists were not available because the e-florists in our sample group were not publicly held companies, a factor that constitutes a common challenge in studying small firms (Wade and Nevo 2005-2006). Given this constraint, we followed a recommendation from the local florists and added questions on channel performance into the survey. This addition was useful mainly because, in each case, the owner or the manager was the sole source of information. Given the confidential nature of the information, we asked questions within the categories of total sales, percentage of total sales contributed by the website, and percentage of the net profit margin of the online portion of the business. The midpoint was used to calculate the actual dollar-based performance, such as e-sales and e-profit. Except for the self-reported objective performance, each item was measured by a 7-point Likert-type scale, where 1 was the minimum and 7 was the maximum value for a given measurement scale.
Measurement Validation
From the two-round Q-sort, we assessed the interrater reliability of each item according to conventional methods (Perreault and Leigh 1989). To make sure of content validity at the item level, we assessed the face validity and substantive validity by the methods developed by Hardesty and Bearden (2004) and Anderson and Gerbing (1991), respectively. Based on the results of the assessment, necessary changes of item descriptions were made between the two rounds of Q-sort, and at the end, we confirmed that there was no serious measurement concern in individual items.
After the survey, the collected data were used to assess the scale-level reliability and validity. For evaluation of internal consistency among the items, Cronbach’s α was analyzed. Scale reliability was assessed using a composite reliability measure and the average variance extracted (AVE). Convergent and discriminant validity was assessed as a part of construct validity since the multiple items used to measure the same construct should be in agreement, while items between different constructs should be distinct (Campbell and Fiske 1959). Using confirmatory factor analysis (CFA), we established convergent validity by the magnitude and significance of the factor loadings (Segars 1997). For assessment of discriminant validity, we compared chi-squares between the measurement model and the nested models in which the correlation of a pair of constructs was constrained to equal one (Bagozzi, Yi, and Phillips 1991). There were four constructs directly measured from observable multiple indicators, including the performance variable in each measurement model, which produced six nested models. The chi-square differences between the original and nested models ranged from a minimum of 219.1 to a maximum of 378.3, all statistically significant, which demonstrated good discriminant validity. Before running CFA, we checked the normality of each observed variable by skewness and kurtosis using Amos 16.0, which showed no serious concern in using the maximum likelihood estimation (Kline 1998).
The results of the scale-level reliability and validity assessment are summarized in Table 2. All factor loadings were highly significant, indicating good quality of the measurement items. Cronbach’s α and composite reliability were all above the conventional cutoff of .7, and AVE was more than .5.
The Results of Scale-Level Reliability and Validity Assessment
Note.
*Item loadings were all statistically significant at p = .01.
**includes an information technology (IT) partner’s e-IT competence.
SEM
After validating the measurement scales, this study used the SEM to test the hypothesized relationships. Our research model included a second-order construct, E-SRB, which has the ability to parsimoniously capture the covariation pattern among first-order factors (Rindskopf and Rose 1988; Segars, Grover, and Teng 1998). While many different interpretations of the hierarchical model do exist, the model specification is purely theory-driven (Gerbing and Anderson 1984). The model was identified and fit statistics were examined before any structural relationship was tested. Our analysis includes the comparison between the second-order proposed model and the first-order direct-effect model, which will clearly show whether E-SRB as the second-order latent factor really exists (Segars, Grover, and Teng 1998).
In order to rigorously examine the performance effects of the resource bundle, we controlled for two variables: organizational size, which was conventionally measured by the number of full-time employees; and online experience, which was measured by the number of years of use of the online channel.
Results
We report the results of SEM and its fit statistics in Table 3, which summarizes the analyses of three structural equation models with different performance variables. The chi-squares are all significant, but the ratios, chi-square/degree of freedom, are all less than 3. In all three models, standardized root mean square residual (SRMR) and root mean square error of approximation (RMSEA) are less than .1, and comparative fit index (CFI) and Tucker-Lewis index (TLI) or ρ2 are more than .9, which indicates a reasonably good fit between the data and the posited relationships (Kline 1998).
Results of Three Proposed Structural Equation Modeling (SEM) Models
Note. E-MA = e-market acuity; E-IT = e-IT competence; E-SA = e-service agility; org. size = organizational size; e-exp = online experience. Org. size and e-exp. are used as control variables. SRMR = standardized root mean square residual; RMSEA = root mean square error of approximation; CFI = comparative fit index; TLI = Tucker-Lewis index.
*p < .05.
**p < .01.
Resource Complementarity
The relationship coefficients between E-SRB and the three resources are slightly different for each performance variable, but they are all significant at the .01 level, as shown in Table 3. Among the three resources, the path to e-market acuity has the highest coefficient in all models, but the links between E-SRB and the other two resources are still strong at more than .60. These results support Hypotheses 1-3, thereby validating our conceptualization of E-SRB as a collection of the three complementary resources.
Performance Implications
In contrast to the consistently strong relationships between E-SRB and the three resources in all three proposed models, the relationship between E-SRB and performance depends highly upon the performance variable. A comparison of the results between perceptual and other dollar-based performance models indicates that the model with perceived financial performance has a much higher relationship coefficient. When perceived financial performance is used, the relationship coefficient reaches .722, which strongly supports Hypothesis 4. The relationship coefficients between E-SRB and the self-reported objective performance variables become substantially lower, .148 for e-sales and .171 for e-profit, but they are still significant at the .05 level, which supports our Hypotheses 5 and 6.
In terms of the control variables, the effect of organizational size is statistically significant on e-sales and e-profit at the .01 level. The result was highly expected, given that they were estimated from total firm sales which highly depend upon the organizational size. However, the effect on perceived financial performance is not statistically significant. The result indicates that, regardless of the size of firms, owners or managers of the firms with higher level of E-SRB perceive that their firms perform better than their competitors do. In contrast to organizational size, online experience is not a significant control variable in all three models, which indicates that the small retailers who have more years of e-channel use do not show a significantly higher e-channel performance.
In summary, the results of SEM using the empirical data show support for the proposed hypotheses. The three resources of e-market acuity, e-IT competence, and e-service agility are found to collectively constitute E-SRB, which positively affects e-channel performance. The results also show patterns in the relationship coefficients within the models with different e-channel performance variables, as explained.
Alternative Direct-Effect Model
While the significant relationship between E-SRB and e-channel performance is validated, it does not exclude the possibility that, individually, these three functional resources influence performance. The direct-effect model offers an alternative explanation of the performance impacts of the three resources (Segars, Grover, and Teng 1998). For comparison purposes, we next test the direct effects of the three resources on performance, as illustrated in Figure 2.

Alternative direct-effect model.
In running the direct-effect models, we allowed correlations among the three resources; otherwise, the model fit statistics went far below the acceptable range. Under the poor fit statistics, comparison between the proposed and alternative models was impossible, which illustrates that the empirical data did not support the purely independent effects. Conversely, the data did support our theoretical position. Only after allowing the correlations among the three resources did the model fit statistics become comparable with the proposed models. The results of direct-effect models that allow correlations are summarized in Table 4.
Results of Three Alternative Direct-Effect Structural Equation Modeling (SEM) Models
Note: E-MA = e-market acuity; E-IT = e-IT competence; E-SA = e-service agility; org. size = organizational size; e-exp = online experience. SRMR = standardized root mean square residual; RMSEA = root mean square error of approximation; CFI = comparative fit index; TLI = Tucker-Lewis index.
*p < .05.
**p < .01.
The results in Table 4 tell a somewhat different story compared to those from the proposed models in Table 3. While the path coefficients from the control variables to the performance variables remain almost the same, the direct effects of the three resources on performance are not statistically significant, with the exception of one path (out of nine), which lies between e-market acuity and perceived financial performance and has a relationship coefficient of .464 (p < .01). These results explain that the individual effects of the three resources are not critical in influencing e-channel performance, particularly in e-sales and e-profit. The comparison study demonstrated that performance variance cannot be fully understood without E-SRB, which supports the position of the proposed models.
Discussion
Theoretical Implications
Previous studies have proposed the positive relationship between service quality and financial performance. However, it is difficult to empirically test the relationship because of the measurement dilemma: Service quality is measured from customer feedback, while financial performance is measured from feedback by the service provider. To overcome the dilemma, many studies used customer loyalty as a proxy of the financial performance (Collier and Bienstock 2006; Heim and Sinha 2001; Koufaris 2002; Reichheld and Schefter 2000), but the two are fundamentally different constructs. We believe focusing on the resources necessary to deliver quality e-services—and the impact of these resources on performance—provides another angle from which to view the relationship between e-service quality and performance of an e-service system.
The varying effects of resources on the different performance variables provide scholars with an opportunity to accurately evaluate different performance variables. The results in Table 3 show that the performance impacts of E-SRB are all statistically significant, but the sizes of their coefficients are quite varied, ranging from .722 for perceived financial performance to .148 for e-sales. These results confirm the general pattern of strong relationships between the perceptual predictor and perceptual performance measures, and weaker relationships between perceptual predictor and objective performance measures (Zhu 2004; Zhuang and Lederer 2006). However, the results also send an important message regarding the selection of performance measures, that is, the potential risk of using perceptual performance only. Of course, perceptual performance differs from objective performance in that service providers evaluate performance based on individually different expectations (Wade and Nevo 2005-2006). Nevertheless, Ketokivi and Schroeder (2004) demonstrate by the confirmatory factor analysis-based multitrait-multimethod analysis (CFA-MTMM) that the use of perceptual performance alone can introduce common method bias, particularly when a single informant is used. This finding presents a measurement issue when studying small firms, where objective data about financial performance are largely not available. Using both perceptual and self-reported objective performance measures, this article achieves a better overall understanding of e-channel performance.
Use of a concept often hinges on availability and validity of its operational instrument, particularly when the concept represents a set of resources like E-SRB. In this study, the second-order factor model was used to operationalize E-SRB, and furthermore, the comparison analysis between the proposed and the direct-effect models provides clear evidence that E-SRB does exist as a set of complementary resources. The individual effects of the resources were mostly nonsignificant, but the bundling effect was significant. In fact, Macduffie (1995) introduces several methods to validate the bundling effect without comparing with the direct-effect model, even before connecting to performance, which can increase the robustness of our research results. Among the three analytical methods he proposed, such as reliability tests, factor analysis, and cluster analysis, he considers factor analysis as being best suited to understanding the interrelationships among the bundling subjects, particularly when they are measured using a scale rather than an index. The three resources in this study are measured by scale, which offers an ideal condition for using the factor analysis.Table 5 summarizes the results of bundle validation tests using both the CFA and the reliability tests that analyze correlation. The significant correlation coefficients among the three resources and the significant and high factor loadings in the CFA results confirm that the bundle does exist, even before positing the performance implication.
Bundle Validation Results
aFit statistics: χ2 = 167.2 (p < .01), χ2/df = 2.26, SRMR =.056, RMSEA = .093, CFI = .933, TLI ρ2 = .917. SRMR = standardized root mean square residual; RMSEA = root mean square error of approximation; CFA = confirmatory factor analysis; CFI = comparative fit index; TLI = Tucker-Lewis index.
*p < .05.
**p < .01.
Managerial Implications
The findings from this research have implications for managers in small firms that operate within the unique environment of e-service provision. A high degree of hype and the competitive pressure of adopting the e-channel have preoccupied the small firm (Pfughoeft et al. 2003) due to the rapid growth of the online retail market, that is, from $28 billion in 2000 to $135 billion in 2009 in the United States (U.S. Census Bureau 2009). This pressure has pushed the firms to take action—such as e-channel adoption—even when they were not properly prepared to do so. After e-channel adoption, however, managers need to constantly ask themselves which resources are necessary and how to allocate them for e-service success. E-SRB exactly responds to those questions. Our empirical findings about the positive relationships between E-SRB and e-channel performance provide managers with useful insight regarding optimum resource allocation beyond the generic advice of, “satisfy online customers at low costs.” Such insights offer a distinct advantage to firms. Instead of spending valuable time on a hit-and-miss approach toward identifying and allocating the appropriate resources, managers can deal with the challenging situation that follows the channel adoption by drawing on their in-depth understanding of the three subject resources and their bundling effects.
The findings from the comparison between the proposed and the direct-effect models offers managers a new insight into the value of a resource. If they believe in independent effects of individual resources on performance, it is sufficient for improved performance for them to change the contents of any resource, for example, either by hiring a competent e-marketer in the e-market acuity domain, purchasing new IT hardware to improve e-IT competence, or redesigning their delivery processes for enhanced e-service agility. However, our research results support an alternative view on resource value that depends upon correlation of coexisting resources. The significant bundling effect suggests that managers should seek to understand the interrelationships that might exist among those resources that collectively influence performance. It is important for managers to consider the bundling effect from a holistic perspective because, in the above example, the competent e-marketer might ask the firm to use advanced web design skills and tools and to implement efficient delivery processes in order to maximize marketing effects. In the end, the changed insight on the value of a resource will significantly change the content and outcome of the resource-allocation decision process.
The positive influence of E-SRB on the size of both sales and profit from the e-channel is quite encouraging. Delivering high-quality e-services can certainly serve to attract customers and increase sales, but a firm will be profitable only when the overall costs remain less than the sales. Normally, cost and quality are the two major competitive priorities that introduce trade-offs against each other, which represents a traditional view of operations strategies. There is, however, much evidence that a selection force favors those firms that target both (Ferdows and De Meyer 1990; Hill 1994). In the case of small retailers’ e-channel use, e-sales increase is critical for e-profit generation because Internet server membership and update of the website as well as the service delivery incur costs surrounding the e-channel. Beyond the mechanism of profit generation, the concurrent effect of different competitive priorities can be aptly applied to the provision of e-service. The Internet lifted information asymmetry for the most part, and online customers can easily find the e-service provider that promises the best deal in both cost and quality. As a result, online retailers should compete on every competitive priority, including cost and quality, simultaneously. Our research finds that resource bundling positively associates with both sales and profit, which managers might find useful.
Finally, the concept of E-SRB and its performance impact provides a direction for managers in terms of how to attach quality to their e-service process. A great number of e-service quality factors that can confuse managers have been identified based on the gap between expectation and perception from the customer standpoint (Parasuraman, Zeithaml, and Malhotra 2005). The concept of E-SRB connects to e-service quality, but it is measured according to e-service providers. Hence, E-SRB can provide a direct prescription to managers for improving an e-service system with minimum confusion. With respect to the challenge of satisfying online customers, where customer expectations vary according to demographic characteristics and change in response to technological advances (Parasuraman 2000; Xue and Harker 2002), our research findings suggest that managers should mobilize a set of three resources simultaneously. In the examination of such a complementary effect, managers might want to heed the uniqueness of an e-service context. First, a service resides in an open system, wherein the production and consumption occurs simultaneously, which is different from manufacturing (Fitzsimmons and Fitzsimmons 2008). In the case of e-services, a changing technological environment mixed with volatile customer loyalty produces a significant need for coexistence of resources surrounding web-users, technologies, and service delivery processes, a directive that is well embedded in E-SRB.
Future Research
This study opens the door for future research on resource bundling, the scope of which can vary, spanning a variety of operational contexts. The current study has used only one industry sector, but application of the bundling principle to different e-service sectors will lead to new insights because different types of products demand distinct customer preferences, website functionalities, and operational standards (Boyer, Hallowell, and Roth 2002; Konuş, Verhoef, and Neslin 2008).
Separation of back-office and front-office operations in a service system demonstrates an absolute necessity of resource bundling between the two heterogeneous operations. Chase (1981) introduced the concept of “back-office” and “front-office” service operations and advised service firms to have separate strategies for each of them. According to Chase, back-office operations require manufacturing principles to increase process efficiency, while front-office operations need more customer-oriented strategies to satisfy customers. However, Sousa and Voss (2006) warn of the asymmetric research convention in the e-service area that focuses only on website characteristics in the front-office. Embedding service quality in each back-office operation is still important because it influences performance of front-office operations and thus customer satisfaction (Field, Heim, and Sinha 2004). The current study is one of the first trials that examine resources from a holistic perspective, but by presuming two different operations—front and back, their complementarities would be an interesting future research topic.
In the data collection stage, we recognized that many small online stores have at least two business channels—the physical store and the website. In multichannel management, channel conflict is clearly an important managerial issue. Channel conflict is driven by consumers’ comparative channel experience and evaluation (Falk et al. 2007), which implies that too much emphasis on E-SRB can result in channel conflict since many store-visiting customers can turn instead to the e-channel. We reason that one of the key determinants of channel conflict is the e-channel’s market coverage. The market coverage is expected to negatively influence the level of channel conflict since an e-channel with wide market coverage can attract many customers from a significant distance away. Given this negative relationship, it might be interesting to examine which resources are associated with the e-channel’s market coverage.
This research has focused distinctly on a view of resource bundling as an antecedent of e-channel success. Looking closely at small service providers, we have provided empirical evidence that E-SRB positively influences e-channel performance, thereby indicating that the collection of three resources—e-market acuity, e-IT competence, and e-service agility—can help retailers to effectively respond to the needs of the market. New insights are still to come concerning the factors that grow an e-service system, but the goal of our present research is to draw managerial and scholarly attention to the concept of resource bundling.
Footnotes
Acknowledgment
The authors thank the editor and the three anonymous reviewers for their constructive comments, which allowed us to improve the article during the review process.
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
The author(s) received no financial support for the research, authorship, and/or publication of this article.
References
Supplementary Material
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
