“ The relationship between e-CRM and customer loyalty : a Kenyan Commercial Bank case study

Since customer loyalty is key, especially in the highly competitive commercial banking environment, this article evaluated the effects of features of electronic customer relationship management (e-CRM) on customer loyalty. Using a crosssectional survey design, data were collected from a convenience sample of customers of a major international Kenyan bank using self-administered questionnaires. The findings based on correlation and multiple regression analyses, revealed that pre-service, during (the) service and post transactional e-CRM features have a positive and significant relationship with loyalty, and that the pre-service and during the service features significantly predict loyalty. Thus, enhancing e-CRM practices could be a strategic competitive tool to impact the banks’ relationship with their customers.

Customer relationship management (CRM) assists an institution to tap into a range of business benefits such as profitability, customer satisfaction, allegiance (Koçoglu and Kirmaci, 2012; Nguyen and Mutum, 2012; Peštek and Lalović, 2012) while enabling effective communication between the firm and the customer (Gikenye, 2011; Gikenye and Ocholla, 2014).CRM is vital to service quality and a quick response to market changes (Jamali, Moshabaki, Aramoon and Alimohammadi, 2013).Due to technological advances, the use of e-CRM as a marketing paradigm is increasing (Bahrami, Ghorbani and Arabzad, 2012; Thuo, Kibera, K'Obonyo and Wainaina, 2011), and it is used as a strategy to enhance organizational flexibility (Dubihlela and Khosa, 2014; Jamali et al., 2013;Jih, 2011;Salehi, Kheyrmand and Faraghian, 2015;Sigala, 2011).E-CRM is defined as 'technology-centred' relationship marketing and ensuing advantages, incorporating traditional customer relationship management (CRM) tactics and e-business market place applications used by an organization to maintain customer relations (Dubihlela and Khosa, 2014; Harrigan, Ramsey and Ibbotson, 2012; Keshvari, 2012; Kim-Soon and Zulkifli, 2012; Salehi et al., 2015).Webbased technologies use the Internet to enable organizations to attract new customers, analyze their preferences and behaviors, customize support services while increasing services and value benefits, retain customers, and craft tactics to encourage loyalty (Ata and Toker, 2012).In essence, e-CRM is used by staff at all levels of business to enable them to interact with customers electronically (Lam, Cheung and Lau, 2013; Tauni, Khan, Durrani and Aslam, 2014).The Internet and web technologies facilitate the implementation of e-CRM (Dolly and Pruthi, 2014; Kűster, Vila and Canales, 2016), since they are web-based interaction platforms between the institution and their customers (Abdulfattah, 2012;Grover, 2011).Therefore, the features that make up e-CRM remain critical for managing customer relationships online, and building and maintaining loyalty.These features include concrete websites or Internet-based tools integrated into organizational systems, which, when properly customized, enable the required interaction with the customer (Grover, 2011).Thus, the relationship with the customer cannot be realized through the Internet without effective e-CRM, and, in order to succeed, relationship management, therefore, remains an initiative that requires commitment, a correct mindset, and an overarching strategic CRM ethos (Govender, 2004).Some researchers (Ata and Toker, 2012; Azila and Noor, 2011; Salehi et al., 2015) have suggested that numerous business benefits accrue from the implementation of e-CRM, which enhances customer satisfaction and loyalty.In the light of the aforementioned, this paper explores how commercial banks may use technology and e-CRM to manage customer relationships and achieve customer loyalty.

Literature review and research hypotheses
1.1.The service relationship.There are three levels of a service experience, namely, pre-service, during the service sale and after service, which help fortify relationships while increasing overall customer satisfaction (Khalifa and Shen, 2009).According to the Transaction Cycle Framework (TCF) (Feinberg and Kadam, 2002), the three features in e-CRM, namely, pre-transaction, during transaction and post-transaction, are encountered by the customer during a service purchase, at the time of purchase, and after the purchase (Belch and Belch, 2010).The transaction cycle reinforces the three-stage relationships while increasing the overall customer satisfaction.The consumer's behavior with regard to a service enables a marketer to establish a products' or service's position, so as to increase its consumption (Belch and Belch, 2010), as well as appreciating buying processes (Siwach and Dahiy, 2009).Thus, it can be stated that e-CRM depends upon an understanding of consumer buying behavior (Grover, 2011, p. 37).
Pre-service transaction features include sign-in or log in capabilities, information customization and personalization, and information search capabilities, including several items, for example, site customization, local search engine and chat in the pre-purchase stage will lead to site traffic, pre-purchase satisfaction, trust and retention (Peštek and Lalović, 2012;Rozita, 2012;Tian and Wang, 2014).Personalization of communication while making it efficient between seller and buyer, as well as of a product or a service, will draw customers towards the firm (Lam, Cheung and Lau, 2013; Mogharabi, Akbarabadi, Mirnezhad and Kariznoee, 2014), since they get good experience (Nikhashemi, Paim and Haque, Khatibi and Tarofder, 2013; Wells, Valacich and Hess, 2011) and support their final purchase decisions (Zeithaml, Bitner and Gremler, 2012).According to Zeithaml et al. (2012), online information efficiency is referred to as eservice quality, since a customer who is satisfied with the e-service may purchase the product or service.The organization would, therefore, be able to get more information about the customer and package this information as per individual customer's needs and desires (Ha and Janda, 2014; Hung, Chen and Huang, 2014; Peng, Wang, He and Tang, 2015).Customized alerts as a pre-transaction feature can enable the customers to pre-specify and automatically receive information about new services, which the banks release (Tanveer, 2009).Different e-CRM features during transaction stage can influence a customer's decision to complete the online transaction.Customers' education is vital at this point (Khalifa and Shen, 2009).The customers' education includes guidance provided on the procedures to purchase a service or product, criteria to consider, and how to evaluate them.It is at this stage that service suppliers and customers agree on the conditions of their transaction based on their negotiation, and, then, the transaction is completed.As such, security and privacy need to be considered to reduce any perceived risk and give clients sufficient confidence and a greater feeling of security in performing online transactions (Kim, Chung and Lee, 2011; Kűster et al., 2016;Olupot and Kituyi, 2013;Rozita, 2012).Marketing tools such as emails, site personalization, loyalty schemes, price information, and reward schemes strengthen the relationship between a company and a customer (Yoon, Choi and Sohn, 2008).Online surveys assist in attitude judgment and possible evaluation of customer behavior significant for website customization (Park, Cho and Rao, 2012), as such, the organization is obliged to provide information pertaining purchase conditions.Post-service e-CRM features comprises frequently asked questions (FAQs), problem solving and online feedback (Khalifa and Shen, 2009), which essentially entails customer service.Usually a help desk is established, where all customer correspondence issues with regard to a service or product are channeled thereby, creating a 'personal interaction' with the organization (Peštek and Lalović, 2012).New technological tools such as the Internet, wireless communications, speech recognition, and video help organizations are used to integrate all customer interactions on a central platform providing customers with more control on the services they want (Abdulfattah, 2012).Some researchers, for example, Olupot and Kituyi (2013) have stressed components such as FAQs, complaints ability feature, problem solving feature (online self-help functionality), feedback channels, order tracking and online communities.Since customer satisfaction is a postpurchase experience (Assouad and Overby, 2016), the aforementioned features are critical for increasing the customers' post-purchase satisfaction via one-to-one communication and support from the company's website.Therefore, e-service quality will be assessed based on the competence and enthusiasm of service providers to respond to the customer's problems after the purchase and, their willingness to inform customers about their special offers and complementary services or products (Park et al., 2012).Good e-service quality at the post e-CRM stage is believed to increase online customer's eloyalty (Alim and Ozuem, 2014; Koçoglu and Kirmaci, 2012; Lam et al., 2013).

E-CRM.
Adoption of e-CRM has been a recent strategy in the majority of business organizations, more particularly the developing countries, and, as such, managers have strategically invested in on-line technologies while emphasizing the building and maintaining of worthy linkages with profitable customers (Harrigan et al., 2012;Nguyen and Mutum, 2012;Sigala, 2011;Thuo et al., 2011).Merging technology, processes and other business activities around the customer has facilitated CRM applications (Abdulfattah, 2012), enabling organizations to recognize the best customer, increase their satisfaction and loyalty (Alhaiou, 2011;Tian and Wang, 2014).Effective use of relationship marketing can help to create competitive advantage (Harrigan et al., 2012;Karimi and Sarkhosh, 2006, p. 706;Thuo et al., 2011); and firms that have implemented fundamental e-CRM practices are reaping numerous benefits ranging from superior customer service, improved profitability, sales, reduced operational costs, enlarged customer base and a broader market share (Jamali et al., 2013;Kihara, 2015).The aforementioned benefits emanating from e-CRM practices are associated with customer satisfaction (Alim and Ozuem, 2014;Rabbai, 2013).It is, therefore, vital to create more value to customers since this nurtures interpersonal ties that lead to repeat purchases that benefit the business and consumer marketing.
Recent studies (Abdulfattah, 2012;Rahimiparvar, 2014) on the use of e-CRM in building customer relationships established that it affects online customer satisfaction, service quality and retention (Khalifa and Shen, 2009; Tian and Wang, 2014), and promotes the development of an attractive virtual community which further enhance satisfaction (Alim and Ozuem, 2014).Researchers (Koçoglu and Kirmaci, 2012) have also identified e-CRM antecedents of loyalty in the financial services industry.Therefore, an organization's resources such as organization-customer interface, flexibility, speed of response, etc., need to be positively augmented in the relationship between e-CRM practices and marketing throughput.

E-CRM and customer loyalty.
The advent of new technologies has led to a shift from CRM to e-CRM, and with increasing global penetration of the internet, e-CRM has become a more popular communication tool and relationship-building platform (Lam et al., 2013).Organizations, therefore, are keen to deploy different types of e-CRM strategies to attract, maintain, and enhance customer relationships, which contribute to loyalty and the organization's success (Chess Media Group and Lieberman, 2010; Yun and Good, 2007).However, relationship management is an initiative that requires commitment, strategic CRM tenets and a correct mindset in order for it to succeed (Govender, 2004).Tanveer (2009) postulates that e-CRM infrastructure provides support to valuable customers to remain loyal, since information stored in the e-CRM database assists an organization to look at the actual cost of attracting and retaining customers.The firm can also access new international customers and seize valuable data essential to the firm's competitiveness and market share (Harrigan, Ramsey and Ibbotson, 2009).Azila and Noor (2011) assert that the association between e-CRM and customer loyalty means that the more customers are satisfied, they repur-chase and spread positive word-of-mouth about the service and the provider, which tends to result in longer relationships, trust in, and commitment to the service provider.As such, loyalty will continue to play crucial roles in the organization's competitiveness and profitability (Rahman, 2006).This is supported by some researchers (Gorondutse, Hilman and Nasidi, 2014; Hayes, 2008; Khan and Fasih, 2014; Olupot and Kituyi, 2013) who have suggested that e-CRM impacts on loyalty.Alhaiou (2011) who studied the relationship between e-CRM features and e-loyalty on online shoppers at the different stages of transaction cycle argued that the use of e-CRM in building consumer relationships affects online consumer satisfaction and loyalty.A similar study by Abdulfattah (2012) investigated the effects of various e-CRM features at different stages of the transaction cycle, on customer satisfaction on banks' websites.The aforementioned researcher also established that e-CRM influences customer relationships and enhances online customer satisfaction and service quality.Rabbai (2013) confirmed the effect of e-CRM on customer loyalty, while Alim and Ozuem (2014) concluded that e-CRM is effective in strengthening relationships with customers and promoting the development of an attractive virtual community, which further enhances satisfaction.

Conceptual framework
The literature revealed that the characteristics of three stages of a transaction cycle, namely, pre-transaction, during-transaction and post-transaction which constitute e-CRM (independent variable), impact on customer e-loyalty (dependent variable).The dependent variable (customer e-loyalty) is concerned with the purchase and user behavior, attitudes towards the bank and the purchasing or service consumption situation.Recent studies (Abdulfattah, 2012; Sivaraks, Kirairit and Tang, 2011) postulated that e-CRM features influence customer satisfaction, service quality and loyalty.In addition, Alim and Ozuem (2014) and also Peštek and Lalović (2012) reiterated that e-CRM characteristics are effective in reinforcing relationships with customers, and promoting the development of an attractive virtual community which significantly impacts satisfaction and loyalty.Whereas the moderator variable in the e-CRM-eloyalty relationship is customer satisfaction, the mediating variable (includes factors such as Internet accessibility, network setup, perceived IT usage, previous experience with the bank, etc.) explains the relation between e-CRM and loyalty.This implies that it is not necessarily true that the existence of these transactional features of a bank's electronic platform could lead to loyalty, since other features may come into play to bridge the gap.An example would be the lack of Internet accessibility, which would limit a customer's use of the e-CRM features, and negative perceptions about ICTs would adversely impact its role in enhancing customer loyalty.Previous experience with the bank and/or its employees would also come into play depending on whether they were positive or negative.Therefore, to validate the effect of the e-CRM dimensions on customer e-loyalty, the findings of this study may shed light on the interrelationships among the variables, and help the banking industry to improve the quality of e-CRM activities, thus, enhancing customer loyalty.Figure 1 illustrates the conceptual model of the study, which will be empirically evaluated.

Research design.
A approach using a cross-sectional survey design was implemented, because it is an appropriate strategy in establishing relationships between variables (Creswell, 2014).A bank-intercept survey was used to collect data.

Population, sample and sampling procedures.
The research population consisted of all customers of the Standard Chartered Bank residing in the capital city of Kenya.The unit of analysis was all customers registered and using Internet technology-based applications provided by the bank to access their services.Since a complete list of all customers within the headquarters of the bank was not freely available to the researchers, numerous visits were made to the select branch to conduct the survey among the customers who were at the bank during that time and who met the criteria.The aforementioned bank's headquarters was selected for its convenience due to the difficulty of collecting data from other branches about their e-CRM performance.In addition, e-CRM had recently been relaunched by the bank, therefore, it was an opportunity to assess its performance from the customers' standpoint.One of the researchers is an employee of the bank, which made it easier to get permission to conduct the study.The sample was selected using convenience sampling method where a 'bankintercept survey' strategy was employed in order to gain a representative sample of the customers that constituted a percentage of the total population (Gorard, 2013).Convenience sampling designs, generally, have been embraced in service marketing and social science research (Etikan, Musa, Alkassim, 2016).Therefore, customers who visited the bank for services were invited to participate, and those who agreed to the request were provided with a questionnaire to complete.In addition, the statistical methods applied to test the hypotheses needed to be sufficient enough to perform the relevant statistical analyses.This research comprised three predictor variables, and since it is suggested that for multiple regression analysis the desired level is between 15 and 20 observations for each variable (Hair, Black, Babin and Anderson, 2010), the researchers aimed at a sample size of not less than 60.The questionnaires were distributed at 'opportunistic' bank locations such as in the ATM lobby and banking hall.A total of 78 customers were conveniently surveyed and included in the final sample and their responses werefit for analysis (Lin and Jones, 1997).

Data collection instrument.
A selfadministered questionnaire was used to collect data from the bank's customers.The questionnaire required the respondents to supply background data, namely, their gender, age, level of education, years of dealing with the bank, duration of use of the bank's Internet/electronic-based services, and frequency of use of these services every month.Further sections consisted of questions measuring pre-, at-, and post-services e-CRM, and e-loyalty at each stage.The questions in the aforementioned sections comprised six items for each stage of e-CRM, developed on a five-point scale with 1 = strongly disagree and 5= strongly agree.E-loyalty was also measured on a 5-point Likert scale where 1 = not at all influential and 5 = extremely influential.

Pilot study.
Prior to primary data collection, a pilot study was conducted in August 2016 among ten participant customers (Saunders, Lewis and Thornhill, 2012; Zikmund and Babin, 2010) who were randomly selected from outside the final sample to assure internal consistency out of their responses of the final instrument (Sekaran and Bougie, 2016).Consequently, Cronbach alphas (α) for study variables were as follows: pre-service features and during service features had 6 components and their alphas were .754and .607,respectively.Post-service features had 4 components (α = .741);pre-service loyalty (components 4, α = .834);during service loyalty (components 4, α = .829);and post-service loyalty (components 4, α = .840).The aforementioned reliability results demonstrate that the Cronbach coefficient alphas were acceptable, as suggested by Hair et al. (2010), implying that the measurement instruments were fairly reliable.The reliability analysis permitted us to proceed with the main study.

Ethical considerations.
The study adhered to necessary ethical guidelines (Sekaran and Bougie, 2010, 2016) of information confidentiality and privacy of the participants that was solely the responsibility of the researcher.Informed consent was sought from respondents by disclosing the procedures of the survey and respondents were informed that participation was voluntary and that they were free to withdraw at any time without any consequences.Common ethical research requirements on anonymity, independent work, storage and plagiarism were also observed (Creswell, 2014, p. 92).

Hypotheses
In light of the literature review, this paper explores if and how commercial banks (in Kenya) may use e-CRM practices as a marketing strategy and keep customers loyal by using one bank as a case study.Thus, to examine various relationships, it is hypothesized that: H1: pre-service transaction features have no significant relationship with pre-service customer loyalty at the commercial banks; H2: there is no statistically significant influence of 'during service transaction' features on commercial bank customer loyalty; H3: post-transaction service features have no significant relationship with post-service customer loyalty at the commercial bank.

Findings
Table 1 show that the majority (53.3%) of the customers (respondents) were male and (30.7%) were aged between 36 to 40 years.With regard to their education level, the vast majority (60%) had a bachelor's degree and (25.3%) had a master's degree, and the majority (45.3%) had been with the bank for 6 to10 years.Furthermore, 52.0% have been using the bank's electronic services between 1-3 years and often uses e-CRM services in a month.Source: survey data 2017, SPSS output.

Diagnostic test results.
Since simple linear and multiple regression analyses were conducted, the suitability of the data for the aforementioned procedures were investigated and the assumptions for multiple regression were considered.First, the measures, especially for the dependent variable (customer loyalty), were converted into a continuous scale, and Shapiro-Wilk's W test, which is recommended for small and medium samples (Garson, 2012), was used to test the normality of the data.
Considering that the data is normal when the Shapiro-Wilk (W) p>.05, it is evident from Table 2 that all the variables had normally distributed data, since there were no statistical significant differences noted in any of the variables with their corresponding normal scores.Further, the tests of assumptions of multicollinearity performed using correlation (Table 3) reveal that all the correlation coefficients were less than 0.8, implying that the population data were free of singularity, and that there was no multicollinearity.However, Huck (2010) recommends that inter-correlation among the independent variable beyond .08 is a sign of multi-collinearity and data should be subject to further scrutiny.Tolerance and the Variance Inflation Factor (VIF) tests were conducted to examine the existence of any violation of multi-collinearity assumptions.Table 4 shows that the collinearity conditions were met, given that each of the variables had adequate tolerance value > .10,and the VIF <10.The Durbin-Watson statistic of independence of observation was 1.893, also implying that the data were not auto-correlated (Ayinde, Apata and Alaba, 2012).The result clearly supports the proposition that the e-CRM construct is an antecedent to loyalty.Another interesting finding is that although post-service e-CRM is positively related to loyalty, it is not a significant predictor of customer loyalty (p>0.05).Therefore, it is concluded that customer loyalty in the commercial bank is largely dependent on the pre-service and during service transactional features.In theory, this implies that the e-CRM features are important criterion for a customer of the bank to complete a transaction, hence this increases their loyalty.
The findings have also revealed that e-CRM is positively related to customer loyalty, which is significantly influenced by the pre-service and during service features, but not by the post-service features.This advances our understanding of the influence of e-CRM on loyalty for example, enhancing marketing strategy in general.It is concluded that future investigation of the relationship between the two constructs should take into account moderating factors like customer satisfaction, age, gender etc. that may have an effect on the relationship.
This study was limited in terms of the sampling, since a convenience sample was used.The positive yet non-significant effect of the post-service features on loyalty needs to be examined further.The potential exists to replicate the study with a larger sample of the participants employing mixed methods and more sophisticated analysis techniques such as structural equation modeling (SEM).
Tauni et al. (2014) postulated significant relationships between CRM and customer retention.Prior marketing research conducted in the financial services sector has shown that e-CRM features are antecedents of customer loyalty (Koçoglu and Kirmaci, 2012; Sivaraks et al., 2011).

Figure 2
further supports this relationship, since the scatter plot diagram pattern indicates positive correlation between the variables, and the line of best fit further reveals that there was correlation between the variables.Regression estimates and the Analysis of Variance (ANOVA) of the level of influence of pre-service e-CRM features on customer loyalty showed that the pre-service e-CRM features were a significant predictor of customer loyalty [F (1, 73) = 28.6,p < .001,R 2 = .282].

Fig. 2 . 5 . 3 . 5 . 4 .
Fig. 2. Scatter plot: pre-service e-CRM and customer loyalty 5.3.Influence of during (the) service e-CRM features on customer loyalty.Figure 3 reveals that the during (the) service e-CRM features have a significant positive relationship with customer loyalty [r =.556; p<0.05], therefore, hypothesis 2 is rejected.The results support that the proposition that during (the) service e-CRM is an antecedent to loyalty, with high levels of during (the) service e-CRM features associated with enhanced customer loyalty at the commercial banks(Yoon, Choi and Sohn, 2008).The calculated coefficient of determination showed that the during (the) service e-CRM features were a significant predictor of customer loyalty to the bank [F (1, 73) = 32.721,p< .001,R 2 = .310],and accounted for much of the variability in loyalty among the bank customers.5.4.Influence of post service e-CRM features on customer loyalty.A statistically significant [r = .398,p<.05] correlation exists between the postservice e-CRM and customer loyalty; thus, hypothesis 3 is rejected.It is also evident that the postservice e-CRM feature is a significant predicator of customer loyalty [F (1, 73) = 13.701,p< .05].Several researchers, interalia, Alim and Ozuem (2014),Koçoglu and Kirmaci (2012) andLam et al. (2013), also supported the positive impact of post-services e-CRM features on customer loyalty.

Table 1 .
Bio data of the respondents

Table 2 .
Tests of normality *.This is a lower bound of the true significance.a. Lilliefors Significance Correction.

Table 4 .
Tolerance and Variance Inflation Factor (VIF) statistics between the pre-service e-CRM features and customer loyalty (r=.531, p<.05), with an increase in the pre-service e-CRM features associated to an increase customer loyalty and vice versa.Since the aforementioned results did not support hypothesis 1, it was concluded that pre-service e-CRM features have a significant influence on customer loyalty.This finding supports that of previous researchers, interalia, Rozita (2012), Wells et al. (2011) and Zeithaml et al. (2012).