Computers in Human Behavior 26 (2010) 1296–1304

Contents lists available at ScienceDirect

Computers in Human Behavior journal homepage: www.elsevier.com/locate/comphumbeh

Antecedents of customer satisfaction with online banking in China: The effects of experience
Cheolho Yoon *
Dept. of Business Administration, Mokpo National University, Republic of Korea

a r t i c l e

i n f o

a b s t r a c t
This study investigates the antecedents of customer satisfaction with online banking in China, and explores the effects of experience on the relationships between the antecedents and customer satisfaction. Based on prior research, the six antecedents of customer satisfaction – ease of use, design, speed, security, information content, and customer support service – are proposed, and the effects of experience on the relationships between these and customer satisfaction are analyzed using a structural multi-group (a high-experience group and a low-experience group) model. The results show that design, speed, security, information content, and customer support service have a significant influence on customer satisfaction in the high-experience group or the low-experience group, but ease of use does not have a significant influence on customer satisfaction in either of the groups. Also, the results relating to the influence of the level of customer experience reveal that the effects of design, security, speed, and information content on customer satisfaction were significantly higher in the highexperience group, whereas the effect of customer support service on customer satisfaction was significantly higher in the low-experience group. Ó 2010 Elsevier Ltd. All rights reserved.

Article history: Available online 24 April 2010 Keywords: Online banking Customer satisfaction Experience Multi-group analysis China

1. Introduction Since the first online banking services based on the Internet were provided by Stanford Federal Credit Union (SFCU) in October 1994 (Business Wire, 1995), online banking has spread rapidly in all countries of the world as a result of its convenience and ease in conducting transactions quickly and at a specific service level (Finextra, 2010). According to an International Data Corporation (IDC) report, the number of online banking users worldwide had already surpassed a hundred million by 2004 (Maenpaa, Kale, & Kuusela, 2008). Spurred by rapid growth of online banking usage, a number of the studies about online banking have been performed around the world; however, the research has mostly been confined to: the issues of online banking adoption (Cheng, Lam, & Yeung, 2006; Durkin, Jennings, & Mulholland, 2008; Lee, 2009; Yiu, Grant, & Edgar, 2007) and the customer attitude towards online banking usage (Joseph & Stone, 2003; Laforet & Li, 2005; Liao & Cheung, 2002). Now that online banking has several millions of active users worldwide, we need to shift the focus of research about customer satisfaction or customer loyalty – which increases current customer retention and life-time value of the customers (Maenpaa et al., 2008).

* Address: Dept. of Business Administration, Mokpo National University, 61 Dorim-ri, Chungkye-myun, Muan-gun, Chonnam 534-729, Republic of Korea. Tel.: +82 61 450 2611; fax: +82 61 450 6466. E-mail address: carlyoon@empal.com 0747-5632/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved. doi:10.1016/j.chb.2010.04.001

Customer satisfaction, the feeling of customers using a product or service, is one of most popular research topics in marketing and e-commerce studies, including its positive effects on customers’ repurchase and ‘word-of-mouth’ behaviors. Recent online banking studies, including Casaló, Flavián, and Guinalíu (2008), have found that customer satisfaction with previous online banking interactions have had a positive effect on both customer loyalty and positive word-of-mouth. One survey investigation (Foresee results survey, 2005) concluded that highly satisfied online bankers were nearly 39% more likely to purchase additional products and services from their bank than dissatisfied online banking customers. Thus, indentifying the antecedents of customer satisfaction with online banking is an important research subject in practice. Also, customers’ attitudes and behaviors may vary according to individual characteristics such as age, gender and experience. It has been proposed that experience, in particular, influences consumer (or user) behaviors in marketing, e-commerce and information systems studies. Several studies have examined the effects of experience on customers’ behavior, perceptions and behavioral intentions (Castaneda, Munoz-Leiva, & Luque, 2007; Rodgers, Negash, & Suk, 2005; Taylor & Todd, 1995). Although experience has received a lot of research attention in diverse contexts, little research has been conducted to explore the effects of experience in the online banking context. In 1997, China Merchants Bank was first to launch an Internet payment system in China; thereafter, online banking spread rapidly in China. Most retail banks in China now provide online

C. Yoon / Computers in Human Behavior 26 (2010) 1296–1304

1297

banking as add-on services to existing branch activities (Laforet & Li, 2005). In 2008, the number of online banking users in China was 58 million, an increase of 45% from the previous half-year. The overall percentage of the number of users in the country has reached 19.3% (China Internet Network Information Center, 2009) and IDC predicted that online banking in China would increase rapidly from 2008 to 2012, with a compound annual growth rate (CAGRs) of 23.1% (International Data Corporation, 2008). Given the size of the online banking user community and the recent developments in online banking in China (as mentioned above), it is important to study this area since it represents a large part of the Internet economy – China surpassed the United States in its number of Internet users in June 2008, and is now ranked first in the world (China Internet Network Information Center, 2009). Accordingly, the purpose of this study is to investigate the antecedents of customer satisfaction with online banking in China and explore the effects of experience on the relationships between the antecedents and customer satisfaction. To attain the objectives of this study, we first identify the antecedents of customer satisfaction with online banking based upon a theoretical review and logical reasoning, and we then analyze the effects of experience on the relationships between the antecedents and customer satisfaction using a structural multi-group (a high-experience group and a low-experience group) model. The next section of this study will introduce the antecedents of customer satisfaction with online banking. Our hypotheses will be explained in the third section. The research design and survey results will be presented in the fourth and the fifth sections, respectively. In the sixth section, we will discuss our results and will explain the implications of this study and further research directions. 2. Antecedents of customer satisfaction with online banking The concept of customer satisfaction occupies a central position in marketing thought and practice (Churchill & Surprenant, 1982). Researchers have focused on antecedents of customer satisfaction because customer satisfaction is generally assumed to be a significant determinant of repeat sales, positive ‘‘word-of-mouth”, and consumer loyalty (Bearden & Teel, 1983). Although the subject of satisfaction has been discussed extensively in the literature of information systems, e-commerce and marketing (see DeLone & McLean, 1992; Liu, He, Gao, & Xie, 2008), the exploration of antecedents to customer satisfaction with online banking is still in its infancy. Recently, some researchers (e.g., Liao & Cheung 2008) have started to investigate the determinants of customer satisfaction with online banking. Yet, there is no current consensus on the antecedents or determinants of customer satisfaction with online banking. Given that online banking deals with information systems, e-commerce, and marketing activities, we propose conducting a literature review of the antecedents of customer satisfaction with online banking. 2.1. Ease of use In information systems literature, ease of use has been regarded as a factor that influences users’ satisfaction (Doll & Torkzadeh, 1988; McHaney & Cronan, 1998), a measurement of system quality (DeLone & McLean, 1992), and a determinant of IT adoption (Davis, 1989). These phenomena have been supported by e-commerce studies. That is, in those studies that propose a key dimension of web quality (Aladwani & Palvia, 2002; Palmer, 2002) or find determinants of customer satisfaction (Pikkarainen, Pikkarainen, Karjaluoto, & Pahnila, 2006), ease of use is frequently used. In particular, Liao and Cheung (2008) proposed and empirically tested ease of use as a measurement of consumer satisfaction with online

banking. Therefore, ease of use is an antecedent of customer satisfaction with online banking. 2.2. Transaction speed Transaction speed (often referred to response time) has received attention in the context of information systems and e-commerce due to an increase in focus on the efficiency of operational resources. Therefore, it is frequently discussed as an important factor of user satisfaction in information systems (DeLone & McLean, 1992; Srinivasan, 1985) or commercial web site evaluations (Aladwani & Palvia, 2002). Liao and Cheung (2002) argued that in advanced societies, consumers tend to be highly sensitive to the speed of service delivery, thus transaction speed, as a time-saving feature – is an essential consideration for customers’ satisfaction with self service technologies such Automated Teller Machines (ATMs) and is expected to have a positive influence on customer satisfaction with online banking. 2.3. Design Usability and design are important to human–computer interaction (HCI) because they influence user satisfaction and task performance when using a computer (Palmer, 2002). Usability refers to the extent to which the user and the system can communicate clearly and without misunderstanding through the interface (Benbunan-Fich, 2001). Good web site design enhances usability, thus affecting the success of the websites. Ranganathan and Ganapathy (2002) argued that a web site’s design plays an important role in attracting, sustaining and retaining the interest of a consumer in the site. Also, several studies have empirically verified that the design of a web affects users’ satisfaction (Liu et al., 2008; Zviran, Glezer, & Avni, 2006). Hence, the design of the online banking web site may also have a positive influence on customer satisfaction. 2.4. Security Because e-commerce is based on the Internet, which is an open network, security is an important factor for e-commerce applications. Despite various technical advancements in Internet security such as cryptography, digital signatures and certificates, and authentication, consumers are still concerned about the security of monetary transactions when using the Internet (Ranganathan & Ganapathy, 2002). Monetary transactions over the Internet are the main role of online banking; therefore, if the security concern of online banking is removed, customers’ satisfaction with online banking may increase. Several studies (Jun & Cai, 2001; Liao & Cheung, 2008) also argued that security was a significant determinant of online banking. To summarize, the security of online banking may have a positive influence on customers’ satisfaction. 2.5. Information content The key characteristics of a web site may be categorized into either content or design (Huizingh, 2000). Content refers to the information offered in the web site. The contents play an important role in influencing the behaviors of consumers. Thus, a lot of studies have selected content (or information content) as a measurement of web site quality (Ranganathan & Ganapathy, 2002). Although originally online banking focused on the function that performs bank transactions in Internet, most online banking web sites now provide a variety of information areas such as investment, real estate, and personal financial planning information. The information content of online banking may therefore have a positive influence on customers’ satisfaction.

1298

C. Yoon / Computers in Human Behavior 26 (2010) 1296–1304

2.6. Customer support service According to Joseph and Stone (2003), customer support services may be defined as the ability to provide feedback on problems about the purchased goods and services, or the purchase process itself, and to the ability to solve customer problems and concerns. Thus, customer support services can be regarded as the responsiveness dimension of service quality. Several studies have empirically tested responsiveness as a determinant of service quality or customer satisfaction with online banking (Jun & Cai, 2001; Liao & Cheung, 2008). These studies showed that responsiveness increased customers’ satisfaction. Hence, the quality of online banking customer support services may have a positive influence on customers’ satisfaction. 3. The moderating effects of experience This study includes level of experience as a potential moderating variable because a review of the literature shows that people with a high level of experience are distinct from those with a low level of experience in terms of the knowledge structure (Mitchell & Dacin, 1996; Söderlund, 2002). Differences in the structure of knowledge are reflected in different activities related to information processing, such as problem solving, reasoning and induction, forming opinions, and recall and recognition of information (Hernandez Maestro, Munoz Gallego, & Santos Requejo, 2007). This implies that customers evaluate their satisfaction with online banking differently depending on their experience. The existing literature indicates that experience is closely related to expertise (Rodgers et al., 2005) and self-efficacy (Bandura, 1986). Expertise is defined as the ability to perform product-related tasks successfully (Alba & Hutchinson, 1987), which helps people use a particular product very easily and comfortably. Selfefficacy, another construct, is defined as one’s sense of competence and confidence in performing behaviors to achieve a desired outcome (Suldo & Shaffer, 2007). In information systems studies, it has also been shown to enhance users’ perception of ease of use (Venkatesh & Davis, 1996). Consequently, the more the customer experiences, the higher evaluates ease of use given the effects of both expertise and self-efficacy. These high evaluations can cause a decrease in the importance of ease of use in determining customer satisfaction. In a study that explores the role of prior experience in TAM (technology acceptance model), Taylor and Todd (1995) argued that users without experience focus first on ease of use, and after overcoming concerns about use, then focus their attention on perceived usefulness. Castaneda et al. (2007) also found that experience has a moderating effect on web acceptance and that the relationship between ease of use and behavior is stronger for users with limited experience. The results of these studies suggest that ease of use is more important for users with low experience. Therefore, we establish the following hypothesis.  Hypothesis 1 (H1). The effect of ease of use on customer satisfaction is significantly higher in online banking users with low experience. If users do not overcome their concerns about using a web site, the design and transaction speed of the web site are unimportant. Castaneda et al. (2007) argued that, for individuals with limited user experience with a system, ease of use was evaluated as being a strong determinant of customer satisfaction, however, experienced users evaluated a system in a more in-depth way, specifically, they would value web site design or transaction speed more than less experienced users would. Therefore, we establish the following two hypotheses.

 Hypothesis 2 (H2). The effect of design on customer satisfaction is significantly higher in online banking users with high levels of experience.  Hypothesis 3 (H3). The effect of transaction speed on customer satisfaction is significantly higher in online banking users with high levels of experience. In many previous studies, security of financial transactions and personal data has been identified as determinants of online banking adoption (Durkin et al., 2008; Liao & Cheung 2002). The importance of security for the customer’s attitude is determined by the consumers’ perception of risk to privacy and financial transactions over the Internet, which is an open network. In studies related to online shopping, trust is seen as a crucial factor to reduce perceived risk (Pavlou, 2003), which is enhanced significantly by familiarity (often referred to as user experience) with an online vendor (Gefen, 2000). Consequently, it is deduced that the more customers transact with an online bank, the less they have concerns over security and privacy issues. Miyazaki and Fernandez (2001) concluded that higher level of Internet experience resulted in lower perceived risk and fewer concerns regarding security and online fraud. Therefore, we establish the following hypothesis.  Hypothesis 4 (H4). The effect of security on customer satisfaction is significantly higher in online banking users with low levels of experience. According to the heuristic-systematic model (Chen & Chaiken, 1999), heuristic processing takes place in situations in which people are not motivated or not able to think thoroughly about the contents of a message. In contrast, systematic processing occurs in situations in which participants are highly motivated and able to scrutinize a message. Thus, when users evaluate a website, users with low levels of experience will process the website heuristically, drawing on those aspects that are easy to evaluate (ease of use), whereas users with higher levels of experience will assess the more detailed aspects of the website such as information content (Castaneda et al., 2007). Therefore, we establish the following hypothesis.  Hypothesis 5 (H5). The effect of information content on customer satisfaction is significantly higher in online banking users with high levels of experience. Experience refers to personal knowledge or skill derived from participation or observation. Thus, individuals with lower levels of experience may have more problems in interacting with online banking than individuals with high levels of experience. Customer support service means to provide feedback promptly on problems regarding the online banking process, or solving the customer’s dissatisfaction (Joseph & Stone, 2003). This implies the fact that users with low levels of experience may depend on customer support services more, and they may therefore also evaluate it as a more important feature, than users with high levels of experience. Therefore, we establish the following hypothesis.  Hypothesis 6 (H6). The effect of customer support service on customer satisfaction is significantly higher in online banking users with low levels of experience. 4. Research methodology 4.1. Data collection Since the primary users of online banking are college students and white collar workers in China (China Internet Network Information Center, 2009), the respondents were chosen randomly from

C. Yoon / Computers in Human Behavior 26 (2010) 1296–1304

1299

university campuses and on the streets. A total of 300 questionnaires were distributed, among which 224 usable questionnaires were collected and used in the analysis. The respondents were equally distributed between male (114) and female (114). Approximately, 96% of the respondents were over the age of 20, and 69% of the respondents had over 1 year of Internet banking experience. Detailed descriptive statistics relating to the respondents’ characteristics are shown in Table 1. 4.2. Measurements’ development The questionnaire used for data collection contained scales to measure the various constructs of the research model. The measurements were adapted from studies by Aladwani and Palvia (2002), Joseph and Stone (2003), Suh and Han (2003), Yoon and Kim (2009), and Rod, Ashill, Shao, and Carruthers (2009). Although we adapted most of the items from previous studies to establish the content validity of the measurements, some items were newly developed for this study. A pilot test was, therefore, conducted with university students to validate the measurement items. The wording of the survey items was modified based on the results of the pilot test and the advice of IT consultants. All items of the questionnaire are shown in Appendix A. Individuals indicated their agreement or disagreement with the survey items using a seven-point scale. Data analysis proceeded in two stages. First, a validity test on the research measurements was conducted by confirmatory factor analysis (CFA). Second, an analysis of the structural multi-group model was used to test the associations in the research model. 5. Results The structural equation modeling (SEM) approach was used to validate the research model. Partial least squares (PLS-Graph Version 3.0) was employed to perform the analysis. PLS employs a component-based approach for estimation, and places minimal restrictions on sample size and residual distributions. PLS also supports a pooled significance test for a structural multi-group (Chin, 2000), which is becoming the most common approach in IS studies (Qureshi & Compeau, 2009). Thus, PLS was used to accommodate the presence of relatively lean sample sizes and multi-group analysis. 5.1. Reliability and validity of measurement items Partial least squares (PLS) can tests the convergent and discriminant validity of the scales. In a confirmatory factor analysis (CFA) by PLS, convergent validity is shown when a measurement loads highly (coefficient above 0.60) and loads very significantly (t-val-

ues well within the 0.01 level), on its assigned construct (Bagozzi & Yi, 1988). Table 2 shows the factor loadings of the measurement items and t-values. The factor loadings of all items surpass the recommended level of 0.60, demonstrating convergent validity, and all t-values are also above 1.96 (Gefen & Straub, 2005). Discriminant validity is shown when: (1) measurement items load more strongly on their assigned construct than on the other constructs in a CFA; and, (2) the square root of the average variance extracted (AVE) of each construct is larger than its correlations with the other constructs (Gefen & Straub, 2005). As shown in Table 2, all the measurement items loaded considerably more strongly on their respective factor than on the other constructs. Table 3 shows the square root of AVE and the inter-construct correlations. A comparison of the correlation with the squared AVE shows that all correlations between two constructs are less than the squared AVE of both groups. In order to assess the reliability of measurement items, we compute composite construct reliability coefficients. Composite reliabilities range from 0.849 (for ease of use) to 0.929 (for customer satisfaction), which exceed the recommended level of 0.70 (Bagozzi & Yi, 1988). AVEs range from 0.653 (for ease of use) to 0.814 (for customer satisfaction), which also exceed the recommended level of 0.50 (Fornell & Larcker, 1981). The results, therefore, demonstrate a reasonable reliability level of the measured items. 5.2. Common method bias Since data were self-reported for all of the variables – six dependent variables and one independent variable – there were possible sources of common method bias, including consistency motif and social desirability (Podsakoff, MacKenzie, Lee, & Podsakoff, 2003). We used an approach that is used commonly in IS studies (Klein, Rai, & Straub, 2007; Liang, Saraf, Hu, & Xue, 2007; Ke & Zhang, 2009), and performed an analysis to assess common method bias using PLS. We did this using the following procedures. First, a common method factor was added in the PLS model, whose indicators include all of the principal construct items. Then variances of each indicator were calculated, substantively explained by the principal constructs and by the common method factor. The squared values of the common method factor loadings were interpreted as the percentage of indicator variance caused by method, whereas the squared loadings of substantive constructs were interpreted as the percentage of indicator variance explained by substantive constructs. As shown in Table 4, the average of variances substantively explained by the principle constructs is 0.718, while the average of variance explained by the method-based factor was 0.008. The subsequent ratio of substantive variance to method variance is 90:1, with no significant method factor loadings detected for all but two items with significant at the 0.01 level. This analysis supports the conclusion that common method bias did not impact our results (Klein et al., 2007). 5.3. Hypothesis testing results In order to test the hypotheses, we divided all the respondents into two groups: individuals with more than one online banking use per month (high-experience group), and all the rest (low-experience group), and we established structural multi-group models for the groups. A customer who performs more than one online banking per month may be regarded as a high-experience user because a large percentage of bank transactions happen on a monthly basis. Having assessed the structural multi-group model, we examined the coefficients of the causal relationships between con-

Table 1 Descriptive statistics of respondents’ characteristics. Measure Gender Age Value Male Female Below 20 20–29 30–39 Over 40