Dynamics of Online Relationship Marketing: Relationship Quality and Customer Loyalty in Iranian banks

Purpose – The purpose of this paper is to show the dynamics of relationship marketing in e-banking by examining the association between relationship quality and online customer loyalty at different stages of the relationship lifecycle. Design/methodology/approach – A sample of 651 customers of Iranian banks in East Azerbaijan Province who had used online banking services was selected for the study after completing a questionnaire. The research hypotheses were tested using structural equation modeling and the AMOS23 software. Findings – The results showed that the level of the customer relationship determines the effect of relationship quality on customer loyalty in e-banking. Specifically, the effect of online commitment on customer loyalty decreases over time. In addition, as the relationship develops between the customer and the business, the influence of online trust on loyalty increases. Originality/value – The main contribution of this paper is it enriches the relationship marketing literature with respect to the dynamics of relationships by challenging the effectiveness of relationship marketing, especially the use of the same relational constructs (online satisfaction, trust, and commitment) for customers at different stages of the relationship lifecycle.


Introduction
The rise of internet commerce in the 1990s and its ever-increasing growth have been accompanied by tremendous developments in business environments, compelling firms to fight for survival in the highly competitive environment by entering electronic markets and adapting to the new conditions (Elliot, 2007). In addition, despite the rapid development of internet commerce and the need for businesses to enter the electronic market, a growing number of customers state that they are dissatisfied with their online shopping experiences. However, further research is required to gain a deeper perception of the factors affecting customers' evaluations regarding their online shopping behavior, which subsequently have a bearing on their loyalty (Luo, Ba, & Zhang, 2012).
The banking industry is no exception to this evolution as the internet has revolutionized industries throughout the world (Hussien & Aziz, 2013).
Although e-banking, in contrast to traditional banking, enables customers to undertake a wide range of banking activities at any time and place at a low cost (Amin, 2016), the elimination of the role of humans in providing electronic services presents a challenge in gaining customer loyalty (Amin, 2016;Brun, Rajaobelina, & Ricard, 2014). The staff of a service company can help their enterprise by establishing a close and intimate relationship with customers. In electronic services, customer relationships are established via electronic devices, and online relationship quality can play a central role in customer loyalty (Fong, 2015, Ozen, 2015Rafiq, Fulford, & Xiaoming, 2013;Shin, Chunga, Ohb, & Leec, 2013;Wang, Law, Guillet, & Hung, 2015).
Relationship quality is deemed to be a general evaluation of a relationship's power and its responsiveness to the needs and expectations of both parties based on successful encounters and events (Smith, 1998). Relationship quality is a multi-dimensional construct composed of several factors that reflect the general nature of the relationship between companies and customers. Despite the lack of a consensus on the dimensions and elements of quality, there is general agreement that satisfaction, trust, and commitment are key elements of relationship quality (Brun et al., 2014;Hennig-Thurau, Gwinner, & Gremler, 2002;Palmatier, Dant, Grewal, & Evans, 2006;Smith, 1998;Wang, Liang, & Wu., 2006). Also, according to Brun et al. (2014), commitment, trust, and satisfaction are constituent elements of relationship quality in the online context.
Most of the previous studies have examined the association between relationship quality and loyalty in a static state. However, according to the theory of dynamic relationship marketing, relationships have a similar lifecycle to products, and as time goes by, the relationship between a firm and its customers changes and enters a new level. At each level of this relationship, different relational constructs are needed to maintain the association (Zhang, Watson IV, Palmatier, & Dant, 2016a). This reveals that relationships are dynamic and firms therefore need to make different efforts at each stage of the relationship lifecycle to maintain their relationship and gain Akram Garepasha / Samad Aali / Alireza Bafandeh Zendeh / Soleyman Iranzadeh more customer value in the form of loyalty. Thus, academics and business managers need to conduct more empirical research in order to shed further light on the various impacts of relationship quality on customer loyalty during the relationship lifecycle in both online and offline contexts.
The present study, which is based on the relationship dynamics theory (Palmatier et al., 2013), proposes that the association between online relationship quality and customer loyalty varies at different stages of the relationship lifecycle (exploration, buildup, maturity, and decline).
This paper contributes to relationship marketing literature in the following ways. First, it suggests that online relationship quality is a driver of customer loyalty. Second, it implies that the influence of online relationship quality on customer loyalty may vary depending on the different stages of the relationship lifecycle. Third, it provides a real-life analysis of the proposed framework in e-banking services, suggesting that the direction and strength of the link between online relationship quality and customer loyalty change at different stages of the relationship lifecycle.

Relationship quality
Relationship quality can be defined as a multi-dimensional construct related to a customer's general assessment of his/her relationship with a ser vice provider at a specific time based on all previous interactions with that provider (Keating, Alpert, Kriz, & Quazi, 2011). Despite the lack of a consensus regarding the components and dimensions of relationship quality in the literature, a shared line can be identified between various conceptualizations, in that different researchers have proposed satisfaction, commitment, and trust as the key components of relationship quality in the traditional context (Brun et al., 2014;De Wulf, Odekerkenschroder, & Iacobucci, 2001;Hennig-Thurau et al., 2002;Palmatier et al., 2006;Rafiq et al., 2013;Vesel & Zabkar, 2010). Similarly, relationship quality in the online environment has three dimensions, online trust, online satisfaction, and online commitment, which indicate the overall power of the quality of the relationship between online vendors and their customers. In brief, according to the literature, trust, commitment, and satisfaction are among the most important aspects of traditional relationship marketing. Several studies have shown the importance of these three dimensions in online business environments (Brun et al., 2014;Fang et al., 2016).
In the online context, a number of authors have stressed a conceptualization of relationship quality with independent dimensions, considering it to be made up of components of trust, commitment, and satisfaction (Arcand, Promtep, Brun, & Rajaobelina, 2017;Brun et al., 2014;Chung & Shin, 2010;Walsh, Hennig-Thurau, Sassenberg, & Bornemann, 2010). Thus, in the present study, we have conceptualized online relationship quality using three key dimensions of trust, satisfaction, and commitment as related but independent constructs.

Online trust
Given the increasing importance of e-commerce, trust in the digital world has received increasing attention from marketing experts and academia (Beldad, Jong, & Steehouder, 2010). Specifically, online trust is defined as the interplay of positive beliefs or expectations concerning the competency, integrity, and benevolence of a company in an online setting (McKnight et al., 2002).
Many researchers posit that trust is one of the main factors that determine consumers' Dynamics of Online Relationship Marketing: Relationship Quality and Customer Loyalty in Iranian banks initial and sustained use of e-banking services (Lichtenstein & Williamson, 2006;Rexha, Kingshott, & Shang, 2003;Suh & Han, 2002). However, negative features of online transactions such as a lack of control, risky decision making, a lack of physical contact with online companies, and an absence of tangible capabilities in online exchanges cannot be overlooked (Shin et al., 2013). Therefore, in the context of relationship quality, trust represents one of the prerequisites for success in e-commerce. That is, trust triggers new transactions, while its absence creates a barrier against new transactions. Also, trust is one of the determinants of the online environment as it helps to maintain customers and develop long-term relationships with them (Sahney, Ghosh, & Shrivastava, 2013). Moreover, it is also an important factor in determining customers' intentions to make online purchases and remain loyal to e-commerce (Pengnate & Sarathy, 2017).2.1.2 Online commitment According to Berry and Parasurman (1991), relationships are created based on mutual commitment. Morgan and Hunt (1994) view commitment as being at the heart of successful long-term relationships. Given that commitment is a basic variable in measuring the future of seller-purchaser relationships, most studies on relationship marketing have treated it as an important dimension of relationship quality (Brun et al., 2014;Cambra-Fierro, Melero-Polo, & Javier 2018;De Wulf et al., 2001;De Wulf et al., 2003;Hennig-Thurau et al., 2002;Palmatier, 2006;Roberts, Varki, & Brodie 2003;See Dorsch, 1998;Wang et al., 2006). Berry and Parasurman (1991) define commitment as a vital part of a successful relationship that can foster loyalty. In general, studies on relationship marketing tend to treat commitment in terms of emotional commitment. Such commitment is usually assumed to be an attitudinal construct (Bansal, Irving, & Taylor, 2004;Fullerton, 2003;Gundlach, Achrol, & Mentzer 1995).
In the online environment, commitment refers to a type of relationship and tendency that is comparable to emotional commitment. Online commitment is defined as the consumer's desire to continue communication with an online vendor (Rafiq et al., 2013).

Online satisfaction
As another dimension of relationship quality, customer satisfaction plays an important role in competitive environments due to its impact on loyalty (Brun et al., 2014). Online customer satisfaction represents the level at which customers assess their online purchases in general (Al-Hawari, 2014). Internet users often rely on the quality of online systems and the information provided on the website when assessing their online shopping experiences to compensate for the lack of physical contact in traditional transactions. Thus, in online purchases, consumer satisfaction is often associated with the website and its quality rather than the actual items on sale (Brun et al., 2014).

Online loyalty
Loyalty plays a pivotal role in the survival and development of e-commerce (Chen, 2012) as it is a major driver of continued contact with the organization (Rafiq et al., 2013). Online loyalty can be described as a desirable customer attitude and commitment to online shopping, which convinces the customer to repeat his/her shopping behavior (Toufaily & Pons, 2017). Online loyalty of customers to a bank indicates their intention to revisit the bank's website and consider reusing a given product and service in the future (Amin, 2016). Therefore, it can be said that retaining current customers and strengthening their loyalty are main tasks of service providers seeking to gain a competitive advantage (Chen & Wang, 2016), as obtaining loyal customers on the internet can be a serious challenge (Chang & Wang, 2011).
E-loyalty is a concept that has been extensively discussed in online banking literature (e.g. Mohsin & Aftab, 2013;Al-Hawari, 2014). It is important to focus on online customer loyalty to online banking in order to maintain relationships with customers. In this context, Akram Garepasha / Samad Aali / Alireza Bafandeh Zendeh / Soleyman Iranzadeh highly loyal customers tend to visit a website frequently and suggest that site to others (Amin, Isa, & Fontaine 2013)

.2.3 Relationship lifecycle
The existing models of the relationship development process reveal that marketing relationships are inherently dynamic. These models are derived from the interpersonal relationship literature and demonstrate differences in the various stages of the relationship lifecycle (Hansen, Beitelspacher, & Deit, 2013). Many studies that assume relationships involve a dynamic process are inspired by the study from Dwyer, Schurr, and Oh (1987). Chris and Karen (2005) assessed the customer relationship lifecycle based on the time and intensity of the relationship. This lifecycle represents the evolution of relationships over time along with their intensity at each stage. According to Dwyer et al. (1987), lifecycle generally covers five stages of evolution: awareness, exploration, development, commitment, and dissolution. Jap and Ganesan (2000) also specify five stages in the evolution of a lifecycle: awareness, exploration, buildup, maturity, and decline. Despite the different titles used for the various stages, there is a general consensus that the level of each relationship aspect varies from one stage to another.
Since this study looks into the relationship between a bank and its existing customers, the awareness stage has been eliminated as the bankcustomer relationship is yet to be established at that stage. Accordingly, as suggested by Jap and Ganesan (2000), the stages of exploration, buildup, maturity, and decline of the relationship have been considered as the four phases of the relationship in this paper. Reichheld and Phil (2000) contend that to gain the customer's loyalty, it is necessary to obtain their trust. Lau and Lee (1999) found a significant positive correlation between customer trust in a certain brand and their loyalty to that brand. Likewise, Chaudhuri and Holbrook (2001) discovered a significant relationship between trust in a brand and customers' attitudinal and behavioral loyalty.

Online trust and online loyalty
Furthermore, the studies by Wang et al. (2015), Polites, Williams, Karahanna, and Seligman (2012), Ribbink et al. (2004), Safa and Ismail (2013), and Shin et al. (2013) demonstrate the impact of trust on loyalty. Therefore, the first hypothesis of this research is presented as follows: H1: Online trust has a positive effect on loyalty in e-banking services.

Online commitment and online loyalty
Commitment is a key term for researchers and marketers who study online and web-based environments (Park & Kim, 2003;Ozen, 2015), online retailing (Rafiq et al. 2013), internet shopping (Chung & Shin, 2010), e-banking (Mukherjee & Nath, 2007;Sanchez-Franco, 2009;Keating et al, 2011 ), and banking (Rahman & Ramli, 2016). Most of these studies have focused on the association between commitment, satisfaction, trust, loyalty, and intention to purchase/repurchase, among other things. In the marketing research, commitment is a major variable that distinguishes between loyal and nonloyal customers. That is, commitment signifies the tendency to continue a relationship and ensures its sustainability (Rafiq et al, 2013). Based on the above, the second hypothesis is formulated: H2: Online commitment has a positive impact on loyalty to e-banking services. Bitner (1990) argues that satisfaction is an antecedent to loyalty. Also, Oliva, Oliver, and MacMillan (1992) have shown that satisfaction and loyalty are significantly correlated.

Moderating effect of the relationship lifecycle on the quality-loyalty link
Previous studies have explored the effect of relationship quality on customer loyalty from a static perspective. Also, in a number of studies that have examined this relationship from a dynamic perspective by introducing the relationship lifecycle variable, this dynamism has been confined to a mere analysis of the status of trust and commitment at different stages of the relationship lifecycle, with few examining the relationship lifecycle as a moderating variable in the interplay between relationship quality and customer loyalty. For instance, Jap and Anderson (2007) investigated the role of trust in four stages of the relationship lifecycle (exploration, buildup, maturity, and decline), concluding that the highest trust in the producer was observed in the buildup, then the maturity, and then the exploration stages, and the lowest trust was observed in the decline stage. Furthermore, Palmatier et al. (2013) demonstrated that commitment improves until the fourth year of the relationship but tends to decrease between the fourth and the sixth years.
Some of the relationship constructs (e.g. trust, coherence norms, and commitment) develop over a longer period of time while others grow at faster rates. For this reason, different constructs of the relationship may become dominant at different stages of it (Dwyer et al., 1987). For example, trust may be highly active at the early stages of a relationship as the buyer and seller are less likely to build a long-term relationship unless they are able to assess the extent to which they trust each other. However, trust may become inactive (latent) in the following stages and not receive the same level of management attention (Wilson, 1995). Nonetheless, this does not diminish the importance of trust in the later stages.
The scant empirical studies in both areas of marketing, i.e., industrial and consumer marketing, show that the association between relationship quality dimensions and customer performance outcomes such as loyalty changes at different levels of the relationship, but the mechanism of this change is not the same. For instance, Hibbard, Jonathan, Frederic, Rajiv, and Iacobucci (2001) argue that as the relationship ages, the importance of commitment in predicting relationship performance declines, but the role of trust in predicting relationship performance increases, until the third and fourth stages of the relationship, when it declines again. However, Zhang et al. (2016a) showed that trust and commitment are strengthened along with customer performance up to the third stage of the relationship lifecycle, but drop in the fourth stage (relationship decline). Also, the findings of Verhoef, Franses, and Hoekstra (2002) suggest that the age of the relationship increases the positive effect of satisfaction and commitment on the number of services purchased. Cambra-Fierro et al. (2018) showed that in the buildup and maturity stages, the impact of relationship quality on customer value co-creation was stronger than in the decline stage.
The different results suggest that further research is required to understand the mechanism by which the relationship quality dimensions affect relationship performance, especially customer loyalty. Therefore, the following hypotheses are put forward:   (2009), using four items. Online satisfaction was also measured by three items suggested by Fang et al. (2016) and Wang, Wang, and Liu (2016). Electronic customer loyalty was evaluated by six items derived from the studies of Alonso-Almeida, Bernardo, Llach and Marimon (2014), Amin (2016), Bernardo, Marimon, and Alonso-Almeida (2012), and Toufaily and Pons (2017). Finally, the relationship lifecycle was measured by a fourpoint scale proposed by Jap and Ganesan (2000).
A total of 19 items were used to measure the constructs in this model. Of these items, four were removed at the assessment stage as their factor loadings were below the acceptable level Dynamics of Online Relationship Marketing: Relationship Quality and Customer Loyalty in Iranian banks (three items related to online commitment and one related to online loyalty), and so 15 items were included in the final model. Table 2 displays the measurement items and research constructs along with their sources. All of the measurement scales reflected the underlying constructs, for which a 7-point Likert scale was used (ranging from "strongly disagree" to "strongly agree"). The relationship lifecycle was measured by four choice items on a nominal scale, with each choice representing a stage of the relationship lifecycle (i.e. exploration, buildup, maturity, and decline).
The respondents were asked to specify the level of their relationship with the bank by choosing one of the options. In previous studies, besides the relationship lifecycle, other parameters such as the age or length of the relationship have been used to measure the relationship stage. However, as indicated in the literature, the relationship lifecycle is the most effective method to determine the level or stage of the relationship (Jap & Ganesan, 2000). The relationship lifecycle suggests that relationship formation is an evolutionary process while the relationship age approach overlooks temporal variations by assuming that all relationships in the lifecycle move at an equal rate (Palmatier et al., 2013). Thus, age is not a suitable criterion to measure the stages of the relationship. Some relationships may reach the stage of maturity, whereas others might still remain in the stage of development even after several years (Eggert et al., 2006).
The final draft of the questionnaire was prepared after a thorough review of the related literature. Bank managers' views were taken into account and we conducted interviews with a number of bank customers. In the final step, the opinions of marketing experts were sought. The goal of this step was to assess the measures adopted in the study. The first draft of the questionnaire was constantly modified throughout these steps. The revised version was sent via email to 750 customers of Iranian banks in East Azerbaijan Province who had used e-bank services, 651 of whom agreed to fill out the questionnaire. This yielded a response rate of 86%, which was adequate for structural equation modeling (SEM). In terms of gender, 51.9% of the respondents were female and 37.8% were in the age range of 21 to 30 years. (See Table 1 for the participants' demographic information).

Analysis Approach
The data analysis was performed by SEM in the Amos 23 software using a 2-step approach (Anderson & Gerbing, 1988). Delineating the patterns of relationships between constructs was the primary focus of the study; therefore, a correlation matrix was used to estimate the structural model (Hair, Black, Babin, Anderson, & Tatham, 1995). Cronbach's alpha was utilized to evaluate the internal consistency of the scales. With an alpha value greater than 0.70, all measurements confirmed the reliability of the model.
Confirmatory factor analysis (CFA) was used to determine whether the number of factors and loadings of the measured items were consistent with expectations, based on previous research and the theory. Also, CFA was performed to assess the overall validity of the measurement model. The final measurement model indicated a good fit of the 15-item model with χ2 = 243.946; df = 84; p < 0.001; CFI = 0.973; RMSEA = 0.054; NFI = 0.959; and TLI = 0.966. Additionally, the results confirmed the construct validity of the measurement model.
The internal consistency, validity, and reliability of the scales were further examined using three indicators: composite reliability (CR), Cronbach's alpha (α), and average variance extracted (AVE). The acceptable values for CR, AVE, and Cronbach's alpha are above 0.5 (Anderson & Gerbing, 1988), 0.7 (Fornell & Larcker, 1981), and 0.7 (Nunnally & Bernstein, 1994), respectively. As shown in Table 2, CR, AVE, and Cronbach's alpha values were calculated for each construct, with the results indicating that all constructs are within the acceptable range. Therefore, it can be stated that the constructs had acceptable reliability.
Moreover, to assess validity, the square root of the AVE was compared to all inter-factor correlation coefficients. As shown in Table 3, the least squares were computed by comparing correlations between each pair of constructs and the square root of the corresponding AVE values of all constructs (Fornell & Larcker, 1981).
The factor loadings and fitness of the indices are presented in Table 2. To improve the model fitness and obtain a factor loading of greater than 0.5, four items were removed, leaving a 15item questionnaire. Although the loading factor of item 7 was 0.42 (< 0.5), the effect of this item on lowering the value was more than the other model indicators. Thus, as suggested by Tabachink and Fidel (1996), who consider factor loadings greater than 0.32 as acceptable, this item was retained in the questionnaire (Meyers, Gamest, & Goarin, 2006).  The square roots of the AVE for the different constructs are displayed on the diagonal line in Table 3. As shown in the table, the values of the square roots of the AVE for all constructs were greater than the inter-construct correlations (Fornell & Larcker, 1981), which verified the discriminant validity. Hence, the good construct validity of the measurement model is confirmed. Note. The numbers on the diagonal are the square root of the AVE. Table 4 and Figure 2 illustrate the hypothesized relationships along with a summary of the hypotheses supported by the results. Table 4 shows that all three aspects of online relationship quality (online trust, satisfaction, and commitment) have a positive effect on customer loyalty to the e-banking services. The goodnessof-fit indices showed the adequate fitness of the model, though a significant chi-square was obtained (χ 2 = 243.946, df = 84, P < 0.001, N = 651). Given that the likelihood ratio based on the chi-square is sensitive to the sample size (Byrne, 2001), a relative chi-square statistic (χ 2 / df ) is often used as a goodness-of-fit measure. The estimated value of χ 2 /df in this study was 2.904, which is lower than the threshold limit of 5 (Hair et al., 1998). The GFI, AGFI, NFI, CFI, TLI, and RSMEA were 0.952, 0.931, 0.959, 0.973, 0.966, and 0.054, respectively.

Evaluation of lifecycle moderator model
To test the effect of the moderator variable on the relationship lifecycle, multi-group SEM was used and the results are presented in Tables 5 and 6. The results of Table 5 reveal that the effect of online trust on customer loyalty is significant at a 99% confidence interval in the buildup and maturity stages and at a 90% confidence interval in the exploration stage, but it did not have any significant effect on customer online loyalty in the decline stage (β = 0.159; p> 0.1). Considering the significance of the changes in β2 in Table 6 (Δβ2 = 6.614; p <0.1), Hypothesis 4 is confirmed, i.e. the effect of online trust on customer loyalty differs at various stages of the relationship lifecycle.
The analysis of the beta coefficients obtained in the four stages showed that online trust has the greatest effect on customer loyalty in the maturity (β = 0.485), buildup (β = 0.258), and exploration (β = 0.179) stages, respectively. More precisely, as the relationship ages, the impact of online trust on customer loyalty rises, but as the relationship reaches the decline stage, the impact of online trust on loyalty drops to a minimum.
According to the results of the analysis, in the exploration, buildup, and maturity stages, the effect of online commitment on customer loyalty is significant at a 99% confidence interval.
However, in the decline stage, online commitment does not have any effect on customer loyalty (β = 0.041; p > 0.1).
Considering the significance of the changes in β2 in Table 6 (Δβ2 = 8.901; p <0.05), Hypothesis 5 is confirmed, meaning that the effect of online commitment on customer loyalty differs at various stages of the relationship lifecycle.
In addition, using the estimated β coefficients, the differences between the four subsamples can be detected. The values of the exploration stage are higher, followed by those of the buildup and maturity stages, respectively. In other words, with increased duration of the relationship, the impact of online commitment on customer loyalty declines.
With regard to Hypothesis 6, it can be concluded that although online satisfaction had a significant effect on customer loyalty in the buildup and maturity stages, considering that the changes in χ2 in Table 6 (Δχ2 = 6.614; p <0.1) are not significant, Hypothesis 6 is rejected. That is, the effect of online satisfaction on customer loyalty is similar at different stages of the lifecycle.

Discussion and Managerial Implications
The main goal of this paper was to investigate the link between relationship quality and customer loyalty at different stages of the relationship lifecycle in the e-banking context. The increasing use of the internet in businesses and the development of a highly competitive and challenging environment in various industries, including the banking industry, along with the problems inherent to this kind of business, including a lack of physical contact with online companies, a lack of touch capabilities in online exchanges, and an absence of control and decisionmaking in online commerce, have driven online companies to seek a competitive advantage and establish loyalty and long-term relationships with customers. Hence, in addition to focusing on the quality of services, they pay greater attention to the quality of their relationships with customers in a bid to establish relationships based on trust and mutual commitment and to attain customer satisfaction. A number of conclusions can be drawn from the findings of this study.
The results of testing the hypotheses showed that online satisfaction, commitment, and trust have positive effects on online loyalty. These results are consistent with those reported in previous studies (Amin, 2016;Giovanis et al., 2015;Levy, 2014;Rahman & Ramli, 2016;Sharifi & Esfidani, 2014;Wang et al., 2015). In other words, increased customer commitment, satisfaction, and trust lead to greater loyalty to online banking in the long run. However, a meticulous analysis of the results indicates that the effect of satisfaction on loyalty is greater than that of commitment and trust. Thus, banks should develop strategies to improve the quality of online relationships to further enhance online loyalty. In this regard, managers are advised to build confidence by fulfilling promises made by the bank in the online environment, to foster trust in the information and services provided, to implement accurate banking transactions, to reinforce customer commitment to online services by offering comprehensive and convenient services, and to meet customers' expectations in order to nurture their loyalty to the online services.
A variety of results were obtained from the different stages of the relationship lifecycle. The effect of online trust on loyalty was not significant in the exploration stage of the relationship, but it increased in the later stages as further communication was established and a larger number of customers used the online services. As such, the effect of trust in the maturity and buildup stages of the relationship was higher than in the exploratory stage.
In other words, customer trust in online services grows with the development of the relationship lifecycle. Finally, in the decline stage of the relationship, the impact of trust on loyalty falls to its lowest level. These finding are consistent with the results reported by Dwyer et al. (1987), Hibbard et al. (2001), and Zhang, Li, Wang, and Wang, (2016b).

Dynamics of Online Relationship Marketing: Relationship Quality and Customer Loyalty in Iranian banks
However, regarding the impact of online commitment on online loyalty, the results indicated that as customers move further along the relationship lifecycle, the impact of online commitment on customer loyalty diminishes. Although it may appear to go against common sense, this finding is consistent with some empirical studies in this field. According to Hibbard et al. (2001), as relationships age, the importance of relationship quality variables in predicting customer performance declines. Moorman et al. (1992) argue that relationshipbased partnerships become obsolete over time, and therefore the neutrality of the relationship is reduced. With both sides' expectations and opportunism rising, the continuation of the relationship is threatened.
Therefore, in addition to improving online relationship quality by strengthening online loyalty in customers, banks have to consider other options too. According to our results, the amount of trust in early stages of the relationship is trivial, but it grows over time as the relationship develops, so banks need to pay greater attention to customers in the first and second stages (exploration and buildup) to drive customers into the maturity stage of the relationship. For customers who are in the maturity phase, banks should try to maintain trust and enhance customer loyalty by providing services that are more attractive than those offered by competitors. Nonetheless, given the diminishing effect of commitment on loyalty at later stages, banks need to look for other solutions to ensure loyalty in the long run (e.g. through appreciation or the provision of customized services to customers along with the current basic services), in a bid to draw their attention and ensure their loyalty in the long term. Banks should also identify important customers during the relationship lifecycle and formulate appropriate marketing strategies in fitting with various groups of customers to maintain and develop their relationship.

Limitations and Further Research
This research had a number of limitations. First, the dynamics of the relationship were assessed by adding the lifecycle variable to the model. This was undertaken despite the fact that our study was cross-sectional and the research data were collected during a specific time interval. Accordingly, further insights into the dynamics of the relationship and possibly different outcomes could be obtained in a longitudinal study. Therefore, interested researchers are advised to collect and analyze data over several years. Second, the results of the present research are exclusive to the financial services sector, so caution must be practiced in generalizing and applying them to other services and industries. To ensure the generalizability of the results, researchers need to implement this study empirically in other services in future research. Finally, to develop a new avenue of research, researchers are advised to examine the mediating effect of other variables, such as customer involvement.