MASTER'S THESIS. The impact of CRM on customer retention in electronic banking. Case of Iranian banks. Alireza Nili

Transcription

1 MASTER'S THESIS The impact of CRM on customer retention in electronic banking Case of Iranian banks Alireza Nili Master of Arts, Master programme Electronic Commerce Luleå University of Technology Department of Business Administration, Technology and Social Sciences

2 MASTER'S THESIS The Impact of CRM on Customer Retention in Electronic Banking: Case of Iranian Banks Supervisors: Dr. Abbas Keramati Dr. Anis Chelbi Prepared by: Alireza Nili International University of Chabahar Lulea University of Technology Department of Business Administration and Management Division of Industrial Marketing and e-commerce MSc PROGRAM IN E-COMMERCE [Joint]

3 I dedicate this thesis with all my love to my dear parents Alireza Nili December

4 Abstract The purpose of this study is to investigate the impact of CRM on customer retention in Iranian electronic banking. Iranian banks have used CRM more or less and proceeded to this issue with different rates of success. However, there are not available studies that investigated the impact of CRM in these banks especially in electronic environment. This study first reviews and analyzes the previous researches in the literature that investigate the relationship between CRM and firm performance and customer retention. Next, based on the RBV and process-oriented approach, we suggest a process-oriented framework that links CRM to customer retention. In order to be able to appropriately measure the extent of the usage of each of the research variables across organization, suitable dimensions for each of the research variables are defined based on the literature. In this regard, CRM resources are measured with five dimensions which are operational CRM technologies, collaborative CRM technologies, analytical CRM technologies, organizational, and human CRM resources. CRM processes are studied on both operational and management dimensions. Finally, customer retention programs of CRM are studied in four dimensions including: customer service, customization, community of customers, and loyalty programs. A questionnaire has been designed and used to collect data from 286 top bank managers of two governmental (Refah and Maskan) and two private (Saman and Pasargad banks) banks branches in the province of Tehran. The gathered data have been analyzed using quantitative methods based on a survey strategy. The results of data analysis show the direct effect of all CRM resources including technological and infrastructural resources on CRM processes and also the direct effect of CRM processes on customer retention programs of CRM (loyalty programs, customer service, customization, and creating community of customers) which themselves have direct effect on customer retention. Another contribution of this study is that it presents a comprehensive set of customer retention programs of CRM and it studies the correlation between each of them and customer retention. Keywords: Customer Relationship Management; CRM; customer retention; electronic banking; Iran 3

5 Acknowledgement Completion of this thesis would not have been possible without assistance and support from many individuals during the process of writing. I would like to take this opportunity to express my sincere appreciation and thanks to all of them. I would like to express my gratitude and appreciation to Dr. Abbas Keramati for his supervision, constant support and valuable guidance from the very early stage of this research. I would also gratefully extent my acknowledgment to Dr. Anis Chelbi for his intelligent advice, supervision, and support. Special thanks also go to all the respondents in Refah, Maskan, Saman, and Pasargad banks for their contribution and participation in this research. I also would like to thank all of International University of Chabahar and Lulea University of Technology faculty members for their support during this research. Finally, I would like to deeply thank my dear family, especially my wonderful parents for their never-ending love, encouragements and support throughout my entire life. I am deeply and forever indebted to them. Alireza Nili December

6 Table of contents Abstract Chapter 1: Introduction 1.1. Background Definition of Customer Relationship Management (CRM) and scope Customer retention The impact of CRM on customer retention Electronic banking Research problem Purpose of the research Outline of the thesis...18 Chapter 2: Literature Review 2.1. Introduction Customer Relationship Management (CRM) Goals of CRM The impact of CRM on customer retention Resource-based view and process-oriented approach in the CRM value creation model CRM and IT Resource-based view of the firm Technological CRM resources Communicational/collaborative CRM technologies Operational CRM technologies Analytical CRM technologies CRM technological activities in e-banking Infrastructural CRM resources Human CRM resources Organizational CRM resources Process-oriented approach The CRM process Operational CRM processes Management CRM processes CRM process capabilities Customer retention programs of CRM Summary of studies linking CRM to customer retention..42 Chapter 3: Frame of Reference 5

7 3.1. Introduction Resource-based View of the firm (RBV) Technological CRM resources Infrastructural CRM resources Process-oriented approach CRM processes Customer retention programs of CRM Research model Chapter 4: Research Methodology 4.1. Introduction Research approach Qualitative versus Quantitative Approach Deduction versus Induction Approach Purpose of the Research Exploratory research Descriptive research Explanatory research Research Strategy Time Horizon Data Collection method Sample Selection Probability Sampling Non-probability Sampling Sample size Questionnaire Design Pilot Test Data Analysis Validity and Reliability Validity Reliability 65 Chapter 5: Data Analysis 5.1. Introduction Descriptive Statistics Inferential Statistics Exploratory Factor Analysis The results of Exploratory Factor Analysis for the variable of "technological CRM resources" 76 6

8 The results of Exploratory Factor Analysis for the variable of "infrastructural CRM resources" The results of Exploratory Factor Analysis for the variable of "CRM processes" The results of Exploratory Factor Analysis for the variable of "customer retention programs of CRM" The results of Exploratory Factor Analysis for the variable of "customer retention" Structural Equation Model for research Hypotheses Tests Structural Equation Modeling Procedure Lateral Analysis Study of the correlation between each of the customer retention programs of CRM and customer retention using Pearson correlation Study of the difference between the governmental and private banks in relation to the research variables Study of the difference between the four banks (under study) in relation to the research variables by using ANOVA Conclusion 111 Chapter 6: Conclusion and Implications 6.1. Conclusion and implication Theoretical Implications Managerial Implications Research limitations and suggestions for future research.118 References.120 Appendix A 125 Appendix B 127 Appendix C 133 Appendix D 137 7

9 List of figures Figure 1.1 Outline of the thesis 19 Figure 2.1 The emerged framework investigation of CRM activities in Iranian banking industry (Keramati et al., 2009a, p. 213).31 Figure 2.2 Getting more customer interaction (Winer, 2001, p. 93) 35 Figure 2.3 CRM value generation process (Keramati et al., 2010, p. 1176) 38 Figure 2.4 Customer Retention Programs (Winer, 2001, p. 98)..39 Figure 2.5 Influencing factors of customer loyalty (Adapted from: Juan & Yan, 2009)..40 Figure 3.1 CRM value generation process (Keramati et al., 2010, p. 1176) 49 Figure 3.2 The research model.51 Figure 5.1 The distribution of the respondents' age.68 Figure 5.2 The distribution of the respondents' educational level...70 Figure 5.3 The distribution of the respondents' field of study.71 Figure 5.4 The frequency related to the respondents' job experience..72 Figure 5.5 The frequency related to type of the banks under study. 72 Figure 5.6 The frequency related to the belonging to each bank.73 Figure 5.7 The general form of structural equation model..96 Figure 5.8 Symbolization 97 Figure 5.9 The research conceptual model..98 Figure 5.10 Standard model.99 Figure 5.11 Diagram of significance..101 List of tables Table 2.1 Summary of studies linking CRM to customer retention. 43 Table 2.2 Analysis of studies linking CRM to customer retention..45 Table 4.1 Relevant situation for different research strategies (Yin et al., 1994, p.6) 56 Table 4.2 Descriptive statistics in sampling.61 Table 4.3 The sample size for each bank.62 Table 5.1 The frequency of the respondents' age Table 5.2 the frequency of respondents' gender...69 Table 5.3 The distribution of the respondents' educational level 69 Table 5.4 The frequency of the respondents' field of study.70 Table 5.5 the frequency related to the respondents' job experience 71 Table 5.6 The frequency related to type of the banks under study..72 Table 5.7 the frequency related to each bank under study in the sample 73 Table 5.8 The means of the four research variables for bank Refah...74 Table 5.9 The means of the four research variables for bank Maskan 74 Table 5.10 The means of the four research variables for bank Saman 75 Table 5.11 The means of the four research variables for bank Pasargad 75 Table 5.12 The means of the four research variables for governmental and private banks..76 Table 5.13 KMO and Bartlett's Test. 77 Table 5.14 Communalities

10 Table 5.15 Total Variance Explained..78 Table 5.16 Rotated Component Matrix 79 Table 5.17 KMO and Bartlett's Test. 80 Table 5.18 Communalities...80 Table 5.19 Total Variance Explained..81 Table 5.20 The communalities after deleting the questions with little communalities.82 Table 5.21 Total variance explained after deleting the questions with little communalities 82 Table 5.22 Rotated Component Matrix 83 Table 5.23 KMO and Bartlett's Test.84 Table 5.24 Communalities...84 Table 5.25 Total variance explained before deleting the question..85 Table 5.26 The communalities after deleting the question with little communalities...86 Table 5.27 Total variance explained after deleting the question with little communality.86 Table 5.28 Component Matrix.87 Table 5.29 KMO and Bartlett's test..88 Table 5.30 Communalities...88 Table 5.31 Total variance explained before deleting the question. 89 Table 5.32 The communalities after deleting the question with little communalities..90 Table 5.33 Total variance explained after deleting the questions with little communalities.91 Table 5.34 Rotated Component Matrix 92 Table 5.35 KMO and Bartlett's Test 92 Table 5.36 Communalities...93 Table 5.37 Total variance explained before deleting the question with little communality..93 Table 5.38 The communalities after deleting the question with little communality..94 Table 5.39 Total variance explained after deleting the question with little communality.94 Table 5.40 Component matrix.95 Table 5.41 Fit indices Table 5.42 Pearson correlation values Table 5.43 Group Statistics 104 Table 5.44 Independent Samples Test 105 Table 5.45 ANOVA Table 5.46 Multiple Comparisons

11 1. Introduction This chapter presents a research background regarding the concept of Customer Relationship Management (CRM) and its impact on customer retention in the electronic banking environment. It continues with problem discussion, the purpose of the research and the outline of the thesis Background This section presents the background of this study focusing on the central idea of CRM and its impact on customer retention in electronic banking Definition of Customer Relationship Management (CRM) and scope In the mid-twentieth century, increased product availability for consumers by mass marketing and mass production techniques, changed the purchasing process and competition view for the firms. Based on Fickel (1999), Chen and Popovich (2003) argue that as the customers have the ability to choose the items they want among the wide variety of services and products, companies lost track of market needs. Today, companies are trying to build better relationships to existing as well as new customers to increase customer loyalty and long-term retention. Therefore, many of them are trying hard to be more efficient in using technological resources and strategies of customer relationship management (CRM). Researchers also point out that technological applications of CRM link front office functions for example customer service, marketing, and sales and back office ones such as operations, financial, human resources and logistics with the customer touch points in the companies such as call centers, sales, , fax, and pagers which most of the times are managed by separate information systems. Different authors have described CRM in several ways which reflect variety of viewpoints about its definition. Some of them have described it as a philosophy or strategy, others as a process to maintain profitable customers (Zablah et al., 2004). However, according to Chan (2005) and Dimitriadis and Stevens (2008), most authors 11

12 consider CRM as the combination of information systems and strategies in order to make the organizations more customer-centric. In this view, CRM is not considered as only technology, but as a tool which enables the organizations to store and analyze the data about companies and their customers (Keramati et al., 2008). Customer relationship management is a comprehensive process and strategy for customer retention and acquisition to create superior value for them and for the company (Parvatiyar & Sheth, 2004). It can also be defined as a significant competitive strategy companies can use to focus on their customer's needs in the market and to integrate customer-centering path in the firms. Other authors regard CRM as a discipline focusing on development and automation of business process associated with management of relationships with the customers in management, sales, customer support and service (Chatterjee, 2000 cited in Shahin & Nikneshan, 2008). Based on Payne and Frow (2004) and Payne and Frow (2005), Dimitriadis and Stevens (2008) considers that CRM creates long-term and profitable relationships with stakeholders and customers by uniting the IT and appropriate strategies for relationship marketing. It supplies companies with this chance to use data to realize the customer's needs by cross-functional integration of operations, people, processes, and marketing capabilities that is enabled through applications, technology, and information. Regardless of the channel to interact with companies, CRM offers convenience of transactions, and customization for the customers (Gulati & Garino, 2000). According to Boulding et al. (2005, p. 156) and Dimitriadis and Stevens (2008), CRM is the result of integration and evolution of marketing ideas and technological resources, information, and organizational forms: "Not only does CRM build relationships and use systems to collect and analyze data, but it also includes the integration of all these activities across the firm, linking these activities to both firm and customer value, extending this integration along the value chain, and developing the capability of integrating these activities across the network of firms that collaborate to generate customer value, while creating shareholder value for the firm." 11

13 Dimitriadis and Stevens (2008) argue that all of the definitions stated underline that the firm's operations and processes are affected by the implementation and design of customer relationship management (CRM). In the study done by Kassanoff (2000), he also points out that CRM applications help companies assess the profitability and loyalty of their customers by some metrics like the number of repeated purchases and help to answer questions such as: Which products or services do the customers want? And how should the companies communicate with them? Also by these applications, customers save time and money as well as receiving better treatment, information, and efficient service regardless of the channel to contact the firm such as call centers and Internet. However, these strategies and applications are effective if possible risks such as inadequate ROI and budgets are well thought out and the outcome is creating competitive value for the customers and better quality of services and improved profitability with more reasonable price in comparison with the ones competitors offer. In this situation, one can say the organization is moving on the right path and is capable to be in a better position of the market (Zineldin, 2006) Customer retention In the competitive environments, customer retention has been shown to be a very important managerial issue in especially the markets with decreasing rate of acquiring new customers. It is admitted as the first goal of CRM because of its capability in delivering superior value for the customers and firms (Ahmad & Buttle, 2002). Therefore, it can be considered as the primary goal of the organizations practicing CRM strategies and applications (Ang & Buttle, 2006). Although the meaning of customer retention varies between different companies and organizations, they all have come to the conclusion that they can reach many economical objectives by focusing on customer retention such as collecting data about customers to better target and communicate or to customize the interactions with them (Ang & Buttle, 2006). As customers remain loyal, volumes of purchased items will grow and the costs of relationship maintenance reduces because both suppliers and customers know each 12

14 other better than past. Also, these loyal customers pay higher prices than the new ones, and are less likely to claim for discounts which we should offer to new customers in order to acquire them. However, to increase the net present value of these customers, all of the necessary conditions for CRM programs should be combined effectively (Ang & Buttle, 2006). Weinstein (2002, p. 259) argues that many firms spend most of their energy, time, and marketing budgets to gain new customers. However, because the cost of acquiring new customers is much more than the cost of retaining existing ones, customer retention is very important to most banks and organizations. It can be [up to] ten times more expensive to win a customer than to retain a customer and the cost of bringing a new customer to the same level of profitability as the lost one is up to 16 times more (Lindgreen et al., 2000, p. 295; Ang & Buttle, 2006). However, some organizational processes which can be associated with retaining the customers include the processes for: planning, customer satisfaction measurement, complaints-handling and quality assurance processes (Ang & Buttle, 2006). Finally, to be successful in implementing this program, any company should consider the related issues such as type of customers to be retained and the nature of product or services sold (Ahmad & Buttle, 2002) The impact of CRM on customer retention Many authors describe the final goal of CRM as customer retention which is the key to survive in competitive markets for any firm. Customer satisfaction and surviving in today's competitive environments require a heart to understand the customers; a brain to analyze; and hands to deliver the products to them (Ghahfarokhi & Zakaria, 2009). These researches also explain that customers are the source of revenue because they are the ones who pay the bills and generate profit. The role of customers is so vital for organizations and that is why customer relationship management (CRM) is born based on the recognition of the value customers have. Today, customers are the most significant factor in management of businesses because they are now able to choose the items they want from a variety of choices and therefore change the strategies and programs firms have. Hence, having enough 13

15 information about customer's needs and marketplace is a vital factor to interact with their unexpected behaviors and then act in such a way which is in the direction of companies' objectives (Ghahfarokhi & Zakaria, 2009). Now, firms know that absorbing and maintaining the customers is an art. They have realized that to reach this purpose using appropriate programs such as CRM which is a business philosophy that provides the organizations a vision to deal with their customers better, is inevitable. A CRM strategy delivers this vision by giving shape to customer service, marketing, and activities to analyze data and most organizations view the final purpose of this vision, maximization of relationships with their profitable customers by increasing the value of these relationships (Ghahfarokhi & Zakaria, 2009). Previous researches have shown that there is a strong dependency between customer satisfaction and customer retention (Winer, 2001). Considering customer retention, any firm needs to improve the relationships with its customers in a way that results in more customer satisfaction and therefore increased profits (Dwyer et al., 1987; Becker et al., 2009). CRM tries to retain the customers by its relationship programs whose final goal is to deliver a high level of customer satisfaction. Therefore, if companies want to retain their customers, they must develop and efficiently execute a comprehensive set of relationship programs to deliver a targeted performance customers expect. These programs include customization, customer service, community building, and loyalty programs (Winer, 2001). Based on CSO Insights (2006), Becker et al. (2009) add that most of the times, the implementation of these programs affect each of the firm's goals in a different way and may result in poor economical outcomes. For example, although many firms engage consultants and project members to implement CRM technological systems such as software applications and databases or to align the firms' structures and organizations, they can not be sure that focusing on these activities will guarantee reaching the final goal of CRM which is customer retention. In order to attain this goal of customer retention, in addition to the investments for CRM technologies, firms should consider their organizational and human resources and also relationship between processes and people because they can determine the degree to which CRM is adopted and supported. 14

16 Electronic banking Electronic banking project was launched for the first time in the UK in the early 1980s but it failed to gain considerable acceptance among people. However, rapid growth in electronic services generated renewed interest in them. For example, in the 1990s, UK banks made transactional e-banking services available for people. In fact during that time, a quarter of the responding banks with fully operational online systems provided e-services for their customers (Daniel, 1999; Ibrahim et al., 2006). Daniel (1999) argues that increased competitions in financial institutions have resulted in realizing the need to improve alternative delivery channels, the most recent one being electronic banking. He defines electronic banking as the provision of services or information for customers and the ability to buy products, gain access to accounts, and executes transactions via TV, computer, telephone, ATM He also points out that today; consumers have this ability to change banks only by pressing of a button. They can access to online intelligent agents which enable them to compare the services and conditions and therefore reduce the prices. In addition, he states that numbers of alternative delivery channels are increasing which causes fall in percentage of customers visiting bank branches. Besides, electronic banking has many other advantages for customers such as: - Convenience: Shopping, paying bills, buying, and transferring money from anywhere at any time. - Features: Electronic banking can be carried out at any time of the day or night. - Attractive Rates and Incentives: Banks offer attractive interest rates that are opened online. Many others also offer incentives, giveaways and special offers to customers for opening accounts online. - Consolidated Portfolio Interface: Most banks offer a seamless and consolidated interface to customers. Customers can use e-banking services 24 hours a day, without any limitation in location and standing in lines. These services include Automated Teller Machines, Personal Digital Assistants, Mobile Branches, Interactive voice recognition, Internet Banking, Point of Sale Devices, and Cell Phone Banking. 15

17 Although electronic banking has many benefits both for banks and customers, there are some problems in its implementation such as legal problems in digital signature and problems occurring for security systems (Alikhanzadeh, 2008) Research problem The interest in focusing on long-term and profitable customers and the need to better understand their behaviors have changed the view of marketers about the marketplace. Traditionally, they have been trying to acquire new customers and the currently other firm's customers, which needed heavy price-oriented promotions and mass advertising. However, today, customers have access to a variety of services and products and when they do not meet their needs easily; they can choose those institutions that provide them with fast and high quality products or services. Therefore, companies try to use unique strategies to retain their current customers instead of customer acquisition which needs more investments. To reach this purpose new and different tools and mindset are required (Winer, 2001). Also considering human and organizational resources as much as technological capabilities is necessary to manage good relationships with the customers (Keramati et al., 2008). As it has been said above, many studies show that the cost of acquiring new customers is 6 times much greater than the cost that firms pay for retaining current ones. Other researches also show the same result such as slight increase in customer retention rates results in substantial growth in income of firms. Also, other researches argue that the profit loyal customers generate, is over twice more than the profit generated by new ones (Winer, 2001). Customer relationship management (CRM) is now adopted as a necessity and its methods and experiences are applied in many industries because of its great role in becoming more customer focused to deal with the competitions between companies and to retain their current and loyal customers to gain more profit and reduce the costs (Newell, 2000). CRM is one of the most growing trends in banking industry these days, especially in electronic environment and high investments have been spent on its technologies in order to keep the customers satisfied. Also, it is considered as the top banks' 16

18 implementation programs priority which is today more seen in e-banking (Blery & Michalakopoulos, 2006). In today's competitive markets, an aggressive competition between banks is seen more than in the past. They have realized that relationship with customers and its management are significant factors to win this race especially in e-banking in which face to face interactions does not exist. The applications of customer-centric strategies and programs of customer relationship management (CRM) help banks to build longterm relationships with customers and result in increasing their income. Therefore, in the banking sector, CRM is of strategic significance because of the effects it has on customer satisfaction and retention which is the final goal in any successful businesses (Blery & Michalakopoulos, 2006). Although many evidences show the strong positive impact of CRM on organizational performance such as customer retention, there are many reports showing disappointing results of applying CRM to retain existing customers and developing relationships with them. By reviewing the previously published studies about CRM and firm performance, we found out that to find the origin of this problem and to propose an integrated framework from CRM investment to customer retention, resource based view (RBV) of the organization is extended into it by some researchers. In addition, some other authors have applied process oriented approach in their studies to enhance and improve this path. In this thesis we have also used these two approaches to gain insight about the real impact of CRM on customer retention (Richards & Jones, 2008; Keramati et al., 2010) Based upon the above discussion and on the limited amount of researches available about CRM and its impact on customer retention in electronic banking especially for Iranian banks, we have formulated the research problem of this thesis as follows: "What is the impact of CRM on customer retention in electronic banking in Iranian banks?" 1.3. Purpose of the research From the discussions above, the aim of this thesis is to understand the relationship between CRM resources, CRM processes, customer retention programs of CRM (as 17

19 the CRM process capabilities), and customer retention in electronic banking of Iranian banks. Therefore, briefly, this study tries to investigate the impact of CRM on customer retention in Iranian electronic banking based on RBV and process oriented approach. Iranian banks have used CRM and its strategies more or less and proceeded to this issue with different rates of success. However, there are not available studies that investigated the impact of CRM on customer retention in these banks especially in electronic environment. This research aims at getting a deeper insight to this issue regardless of size and age of the banks Outline of the thesis As it is presented in figure 1.1, this research is divided into six chapters. The first chapter which is the introduction presents the background of the research, the research problem, and research objectives. Chapter 2 which presents a literature review provides the reader with an overview on the main previously published papers related to CRM and customer retention. Chapter 3 deals with the frame of reference. It presents the adopted models for the study, formulates the research hypothesizes, and concludes the final research model. The fourth chapter presents the methodology. It describes the methodological choices made in this work and it also examines its validity and reliability. Chapter 5 analyzes the empirical data which are collected by the means of questionnaire to test the research hypothesizes. This study ends with chapter 6 in which the conclusions related to the formulated hypotheses will be stated. This chapter ends with managerial and theoretical implications and further research perspectives within the area of customer relationship management in electronic banking sector. Figure 1.1 presents the outline of the thesis. 18

20 Chapter 6 Conclusion and Implications Chapter 4 Chapter 5 Methodology Data Analysis Data Presentation Chapter 3 Frame of Reference Chapter 2 Literature Review Chapter 1 Introduction Fig. 1.1: Outline of the thesis 19

21 2. Literature Review 2.1. Introduction In this chapter, based on the research problem presented in the introduction chapter, we will review an essential part of the literature concerning CRM and its effects on firm performance; especially, the 'customer retention'. In doing so, we start by defining CRM and specifying its goals. Next, the two well known approaches namely, the resource-based view (RBV) of the firm and the process-oriented approach, are discussed. Then, we examine the CRM process and process capabilities as well as customer retention programs of CRM. At the end, a summary of the literature linking CRM to 'customer retention' is provided Customer Relationship Management (CRM) Increased product availability and mass production techniques have given this chance to customers to be able to choose the items they really desire among the variety of products. Therefore, focusing on customers' expectations is the most important factor for firms to survive in today market places. On the other hand, knowing customer's needs and problems helps the companies to acquire and retain them easier and with less cost (Dimitriadis & Stevens, 2008). Customer relationship management is built on relationship marketing philosophy and redefines the relationship between companies and their customers. Some researchers have defined CRM as a competitive strategy companies adopt to focus on their customer's needs, but others regard it as a discipline to concentrate on development and automation of business process in companies. However, despite the variety of definitions of CRM, they all intend to build customer relationship to create superior value for both the customers and firms (Chatterjee, 2000 cited in Shahin & Nikneshan, 2008). Today, companies have realized that the cost of acquiring new customers is far greater than the cost of retaining existing profitable customers. Therefore, with the help of CRM strategies, they are trying to build better and customized relationships with 21

22 existing customers in order to increase customer satisfaction and build customer loyalty (Ang & Buttle, 2006). Based on Payne and Frow (2004) and Payne and Frow (2005), Dimitriadis and Stevens (2008) emphasizes that CRM creates long-term and profitable relationships with stakeholders and customers by uniting IT technologies and appropriate strategies. Here, the outcome is a mutually beneficial relationship between company and customers that leads to customer loyalty and therefore more profitability for the companies. Winer (2001) considers the customer retention as the final goal of firms practicing CRM programs. He suggests that if firms want to deliver their customers the performance they expect, a comprehensive set of relationship programs is a necessity. He implies that these programs include customization, customer service, rewards programs, community building, and loyalty programs. Ghavami and Olyaei (2006) argue that CRM is one of the key processes in any company and its implementation needs capital investments to integrate marketing, strategy, and technology. Using CRM, companies can achieve competitive advantage provided it is well implemented. In case it is not, customers may leave the company and never come back again. Now, a key question is: what are the elements of a successful CRM program? To answer this question, based on Payne (2001), Ghavami and Olyaei (2006) explain the following four elements: 1) Strategy assessment process 2) Value creation process 3) Multi channel integration process 4) Data repository process The first process must be done for both business strategy and customer strategy. The components of a business strategy include business vision and competitive characteristics and the components of customer strategy include customer characteristics, choice, granularity, and customers segmentation (Ghavami & Olyaei, 2006). 21

23 The second process consists in the value customers receive, the value organization receives, and the customer segment life time value analysis. The values customers receive from a relationship with a company include value proposition and value assessment. Moreover, in a well planed and customized relationship with customers, organizations receive the values of acquisition economics and retention economics (Ghavami & Olyaei, 2006). The third process is multi channel integration process. In this step, the components which CRM tries to integrate and manage include e-commerce, m-commerce, direct marketing, sales force, telephony, and outlets. Finally, the last process contains the functions of data analysis, information systems, front office applications such as customer service and back office applications such as human resources and logistics (Ghavami & Olyaei, 2006) Goals of CRM Based on the study done by Swift (2001), Persson (2004, p.11), believes that by implementing CRM strategies, firms can achieve many goals such as: - Reducing costs of sales Because the relationship with customers will be more efficient and current customers become more responsive, the costs of sales reduce. - Reducing costs of acquiring new customers Because of savings on marketing, relationship programs, services, and so on, the cost of acquiring new customers will be reduced. - Increased customer satisfaction and profitability - Decreasing the need to acquire so many new customers The number of loyal customers increase, therefore the need to gain so many new customers decreases. - Evaluating profitability of customers Companies can know the profitable customers and the ones who might become profitable. This is a very significant factor, since any business must focus on acquiring and retaining the profitable customers to generate profit and reduce costs. 22

24 - Higher customer retention rates Customer loyalty and retention increases, therefore they will stay longer and generate more profit for firms. Based on Newell (2000), Persson (2004) discusses that a real value to a firm is the value it creates for the customers as well as the value its customers deliver back to that firm. It is necessary to say that this value lies in the customer knowledge and in the way the firms use this knowledge in managing relationships with their customers. If CRM is applied in the right way, it can be an efficient tool that generates profits for firms. By transforming customer data into knowledge intended to build better relationships with profitable customers, CRM creates more customer loyalty and therefore more profits for companies. Ultimately, CRM is about creating mutual and collaborative satisfying relationships between the company and its customers to increase customer loyalty and satisfaction The impact of CRM on customer retention Although many practitioners have provided evidences of the positive effects of CRM on organizational outcome such as customer retention, there are many reports showing disappointing and poor outcomes (Richards & Jones, 2008; Keramati et al., 2010). To find the reason of this problem, many academicians and practitioners tried to identify the origin of this situation and found two problems referred to the performance of CRM. The first reason is that many firms view CRM as an IT solution (Reinartz et al., 2004; Keramati et al., 2010). In this regard, academicians believe that although it is easy to acquire IT in the industry, it does not lead to success in business strategies and does not confer competitive advantage. This finding is the same with CRM technology. Companies can buy the same CRM technologies which their competitors use and not have the same results (Lawson-Body & Limayem, 2004; Keramati et al., 2010). The second reason is referred to the concept of CRM meaning that it should not be considered only in the context of technology. According to Zablah et al. (2004), there are five main concepts on CRM including technology, strategy, philosophy, process, and capability which represent an important insight toward an integrated framework linking CRM to organizational outcome. Indeed, many models have been developed 23

25 to propose how CRM affects the firm performances which are different in conceptualization of key dimensions of CRM and the relationships among them (Keramati et al., 2010). By reviewing the IT and organizational outcome literature, we found out that in order to explain the productivity paradox of IT and to propose an integrated framework from CRM investment to customer retention in firms, some academicians have emphasized on considering the resource based view (RBV) of the organization. Also, some practitioners have applied process oriented approach in their studies to improve and develop this path. Indeed, by using these two approaches, firms can be able to identify important resources in the CRM processes implementations and can find the appropriate mechanisms of CRM value creation for the firm. Therefore to find the real impact of CRM on firm performance such as customer retention which is the most important factor for firms to have better financial outcomes, the consideration of these two concepts is necessary (Levesque & McDougall, 1996; Keramati et al., 2010) Resource-based view and process-oriented approach in the CRM value creation model In this section, we discuss theoretical backgrounds as well as adoption and extension of CRM and IT, resource based view of the firm (RBV), and the process oriented approach CRM and IT While many authors such as Keramati et al. (2010) consider CRM as a customer orientation strategy, they all explain that core organizational IT capability is the basis of CRM and have described IT as information-enabled relationship marketing. This can be the reason of this problem that why many practitioners and managers have seen CRM as only a technology solution (Keramati et al., 2010, p. 1171). In many recent papers on CRM, we found evident emphasizes on accepting CRM as a strategy. Payne and Frow (2005) and Keramati et al. (2010, p. 1171) describe CRM as a continuum which "on the one side, is about the implementation of a specific technology solution, and, on the other side, is a strategy for managing customer relationships to create shareholder value". Therefore, CRM can be defined as a 24

26 marketing strategy to develop beneficial relationships between the company and its customers with the help of IT. As it has been mentioned before, to answer this question of why investments in information technology do not lead to expected organizational outcome, some theories such as a theory based framework which is referred to as the resource based view (RBV) in the organizations have been developed (Keramati et al., 2010). The attributes of RBV is relevant to CRM for some reasons. The first reason is that CRM is rooted in information systems and marketing. Indeed, we can claim that RBV has important roles in the both mentioned fields. First, in the field of IT, RBV has been applied to investigate the ability of IT in providing competitive advantage for organizations. Second, in the marketing field, using RBV, organizations are able to assess the competitive advantage provided by the processes which transform the resources to valuable results both for firms and customers. The second reason for the relevance of RBV to CRM is that the people aspect has gained considerable attention in the definition of CRM (Reinartz et al., 2004; Keramati et al., 2010) and RBV also stresses the importance of people in the field of strategic human resources management for which the most important components are employees' skills and behavior in the organization. Finally, the third reason is that, the aim of CRM is creation of superior value both for organizations (from economical aspect) and their customers (Greve & Albers, 2006; Keramati et al., 2010). According to Keramati et al. (2010, p. 1171), "The RBV combines the underlying principles of economics with a management perspective" Resource-based view of the firm Resource-based view of the firm (RBV) emphasizes on the firm resources and views them as valuable firm assets. As mentioned, technology can be easily purchased and duplicated by any entrants in the industry; therefore it can not provide competitive advantage for the organizations. Many relevant researches also imply that there is no correlation between investments in technology and organizational outcomes, or if there is a correlation, it is negative. This contention can be considered as a reason for weak effect of CRM technology on firm's efficient and sustainable relationship building with customers. On the other hand, recent researches such as the one done by 25

27 Keramati et al. (2010) confirm this idea and imply that if firms want to achieve competitive advantage, they should compete on the basis of unique, rare and valuable corporate resources. Through these researches, two main tangible and intangible resources which are complementary to IT namely as human and organizational resources are identified. Indeed, an efficient use of these resources can lead to firm capabilities and therefore expected organizational outcomes. In this regard, we should note that capabilities measure the organization's ability to assemble these resources effectively to achieve a specific objective (Coltman, 2007; Keramati et al., 2010). Is the RBV applicable in CRM strategies? Reinartz et al. (2004), Zablah et al. (2004), Coltman (2007) and Keramati et al. (2010) conclude in their studies that the findings below imply a positive answer to the mentioned question: 1) The overemphasis on technology investments addressed by the studies about the applications of IT and CRM. 2) The concept of capability which refers to efficient use of all resources and as the result firms will be able to form their performance toward customer-focus strategies and sustainable competitive advantage. 3) In the view of RBV, a selective and path-dependent process can improve the firm capabilities and this is consistent with the goal of CRM. The above mentioned literature discusses the usefulness and application of RBV in the study of CRM. However, to understand why the RBV is better to use than other theories in this context, some researchers such as Keramati et al. (2010) argue that the theory of RBV focuses on the main drivers of competitive advantage which is in the direction of the main objective of CRM that is creation of superior value for the firm and customers. In this regard, some other references view the RBV as a theory applied in industrial organization economics. In these studies, RBV is compared with five theories that have played an important role in the evolution of industrial organization. These theories include Bain-type IO, neoclassical economics, transaction cost economics, and the Schumpeterian and Chicago responses. RBV has at least one similarity and one distinction with the mentioned theories, Keramati et al. (2010, pp ) say that the RBV reflects a strong cumulative IO heritage and is at the same time unique in that it incorporates major departure from each of the five 26

28 theories. In addition to this reason, many other authors such as Lockström (2007) have also based their study on the RBV because of its powerful logic and comprehensiveness Technological CRM resources CRM technologies are the medium and tools which enable the firms to get the appropriate information to the right person at the right time (Massey et al., 2001). Based on the firm strategy on CRM program, Keramati et al. (2009b) classifies technological resources into: - IT infrastructure including the electronic services across the company. - The combination of hardware and software to implement business programs by use of the infrastructure such as the electronic systems for sales and also the analysis tools. CRM technologies in themselves are classified into three classes of communicational, operational and analytical technologies. As it has been said previously, communicational technologies make interactions between the firm and its customers and help driving the sales through communicational channels such as call centers. Customer facing applications are related to operational segment which automate customer service, marketing, and sales. And finally, analytical technologies uses information sources in the best possible way for better gaining insight of customer behavior patterns and then make more personalized communication with them (Keramati et al., 2009b; Keramati et al., 2010) Communicational/collaborative CRM technologies Using communicational CRM, communication between the firm and the customers become possible. This can be done via the e-services provided for customers to contact the firm. It also enables suppliers to interact with each other effectively. Moreover, it offers tools and knowledge to employees, partners and suppliers and helps driving the sales through communicational channels (Keramati et al., 2009a). The special features of communicational CRM include enabling the firm to communicate with the customers at any time and any location. They also allow 27

29 solving customer's questions, fast service response to customer's requirements, increase in quality of services, customer satisfaction and therefore retaining the current customers (Keramati et al., 2009a; Reinartz et al., 2004) Operational CRM technologies A typical operational CRM is the contact center and its management. To interact with customers, customer data should be collected through all possible touch points such as contact center, web, and sales force. Then, these data must be stored in particular databases. By this contact management system, tracking any information related to contact with customers becomes easier and more efficient (Keramati et al., 2009a). It is also established that by using a contact management system, operational CRM improves marketing, customer services, sales force automation and so on. It automates customer contact and customer-facing processes. In addition, it supports their interaction in service, sales, and marketing. Ultimately, operational CRM is a great help in personalizing relationships with customers and realizing their needs. This has been improved by deploying Intranet and Extranet to distribute the information within the organization and between the organization and all the partners via all touch points (Keramati et al., 2009a; Keramati et al., 2010). The functions of operational CRM include sales force automation (SFA) to focus more on sales, customer service and support to provide better and personalized services for the customers, and also enterprise marketing automation (EMA) to evaluate profitability of customer segmentations (Keramati et al., 2009a; Deng, 2009) Analytical CRM technologies Academic researchers argue that to increase competitiveness of a firm and reach its objectives, customer satisfaction is a necessity. To improve it, companies should ensure customer's expectations are understood and met. Therefore, they need a measurement system fed by data gained from electronic systems or directly from customers. In the next step, to support identifying customer behavior and their loyalty level, these stored data must be analyzed by specific analytical systems (Reynolds, 2002; Keramati et al., 2009a). 28

30 Globally, the most important goal of analytical CRM is to use information sources properly to better gaining insight of customer behavior patterns and then have more personalized communications with them. As a result, because customers are offered more customized products that fit their buying profiles, they become more satisfied and therefore more loyal to the company (Keramati et al., 2009a; Reinartz et al., 2004). Now a key question is the following: what is the relationship between these characteristics: communicational, operational and analytical? To answer this question we can say that for instance, without data gathered through operational CRM, analytical CRM can not drive strategic and tactical decision making about customer acquisition or maintenance. Therefore, if firms want to drive the customer life cycle effectively, they must consider all the characteristics of CRM and the available business intelligence at the same time (Keramati et al., 2009a; Keramati et al., 2010) CRM technological activities in e-banking In the past few years, many banks in the world have applied data mining technologies and data warehouses to provide clear information about interactions with their customers and to realize what types of customers they are dealing with. In order to target right groups of customers, these banks also have segmented their customers based on their behavior patterns and loyalty (Keramati et al., 2009a). Iranian banks have also undertaken different CRM aspects. A number of them are using operational CRM while others use communicational one. Some of these activities (e.g. check balances and accounts records) are for observation and some other such as customer service are for account controls. The latter include ordering checkbooks, changing the accounts, bill payments to third parties, sending messages and paying credit-card bills, fund transfer, etc. (Keramati et al., 2009a; Reinartz et al., 2004). In general, the applications of CRM include management and automation of call center, marketing activities, contacts, campaign, s, field service, data 29

31 warehouses, SFA, knowledge management, and customization of the products and services (Reynolds, 2002). A considerable part of the customer data can be collected from the customers interacting through different channels. This data is classified into two main parts which are transactional data and non-transactional ones. The first part is about transaction time, place and amount, but the second part is more about feedbacks gained from customer's propositions or complaints. To communicate with customers more efficiently, both of the mentioned parts should be combined into a customer data profile (Keramati et al., 2009a; Kim et al., 2010). Based on the discussion above, Keramati et al. (2009a) divides customer information into the following three types: Customer's information including their transactional and personal information. The information provided for the customers such as the product information that customers perceived as useful. Non-transactional customer feedback information such as customer's suggestions, complaints and so on. It should be noted that the collected information through different channels must be integrated in a proper database. In the next step, with the help of operational CRM and considering the gained feedbacks, analytical CRM creates various customer segments to help designing proper customer strategies. All in all, to perform this process efficiently, all of these activities should be arranged and conducted one after another (Keramati et al., 2009a; Reinartz et al., 2004). Using CRM techniques, banks can be more able to find the profitable customers who have considerable current and future net value for them, know them and understand their needs better, and retaining and improving the value of their business in a costless and efficient way (Keramati et al., 2009a). As the final result, banks will be more capable to improve their customer satisfaction and retention programs; therefore, customers are more eager to have long-term relationship with them which results in 31

32 Touch points - Call center automation - management - Automated teller machine (ATM) - Internet banking (CSS) Customer service and support - Contact management - Field service automation - Credit card Data collection - Data warehousing - Knowledge management Clients Decision making (SFA) Sales force automation higher customer loyalty and profitability; reduced cost of sales; reduced costs of acquiring new customers and decreased need to acquire so many of them; buying more products and therefore increased long-term value (Keramati et al., 2009a; Greve & Albers, 2006). Keramati et al. (2009a, p. 213) summarize the emerged investigation of communicational CRM, operational CRM and analytical CRM activities in the context of electronic banking in Iran in figure 2.1. Front office Back office Enterprise marketing automation (EMA) - Personalized marketing - Advise, peripheral services - Cross selling Communicational Operational Analytical Fig. 2.1 The emerged framework investigation of CRM activities in Iranian banking industry Reference: Keramati et al. (2009a, p. 213) Infrastructural CRM resources As mentioned before, CRM is a marketing strategy and technology is its non-strategic aspect. On the other hand, the infrastructure of CRM in the organizations is formed based on the non-technological CRM resources which are called the 'infrastructural 31

33 CRM resources' and are classified into human CRM resources and organizational CRM resources (Keramati et al., 2009b; Keramati et al., 2010) Human CRM resources Human CRM resources consist of the employees' ability to work well with the existing CRM programs. In addition, they are about customer facing and noncustomer facing employees' attitudes, technical skills, and ability to convert customer data to knowledge (Keramati et al., 2009b; Keramati et al., 2010). Reinartz et al. (2004) suggest that customers would rather communicate with employees than electronic systems. This implies the reason why academic researchers claim that among all CRM resources, human resources is the first priority to be considered. Based on Rigby et al. (2002), Keramati et al. (2009b) confirm this idea by mentioning that CRM can be implemented more efficiently simply by motivating employees to be more aware of customer's expectations Organizational CRM resources Based on Greenberg (2004), Keramati et al. (2009b) argue that many companies consider CRM as a project of implementing IT technologies. They emphasize that CRM is a program not a project because projects are implemented in limited time, but for CRM there is no end. These researchers add, to implement a successful CRM program, some elements are required including setting CRM goals, defining incentive systems, having a customer centric philosophy, training the employees, and top management commitment Process-oriented approach According to Keramati et al. (2008), while several studies have adopted and applied the RBV approach, it has some limitations, for example, it does not talk about how and through which mechanism the resources should be applied and what the best ways of managing them are. In this regard, the process-oriented approach examines the impact of Information technology on intermediate business process. Mooney et al. (1996) and Keramati et al. (2008, p. 1280) state that "to evaluate IT business value, 32

34 the key business processes within each core business area must be identified and the linkages and contributions of IT to those processes should be defined". Eng (2004) argues that in the context of CRM, the applications of the process-oriented approach and RBV are relevant. Keramati et al. (2010) confirm this idea stating the following reasons: First, a process management orientation is a necessity for a CRM program to be successful; therefore, to ensure the efficient use of firm resources toward the creation of expected firm performance, managers must focus on CRM processes effectively. Second, firms are able to enhance their process capabilities by strategic approaches and long-term view to their resources. Mooney et al. (1996) and Keramati et al. (2010) compared organizations which efficiently deployed IT with other ones to test the ways by which IT influences the processes leading to process capability in firms. The result was that the firms which absorbed and applied IT as a differentiated resource have enjoyed better process capabilities and firm performance out of it. This research result implies the necessity of aligning the process oriented approach with the resource based view of the firm (RBV) The CRM process A process is the way we do a work and in businesses, it is linked sets of activities to create value for customers. According to Dwyer and Tanner (2005) and Ghavami and Olyaei (2006), a CRM process involves four steps including market segmentation, designing a communication strategy, implementing the strategy, and finally the evaluation of that strategy. These authors explain that in the first step of a CRM process which is market segmentation, companies segment and profile the similar groups of consumers and then they customize the products meeting that group's expectations. Besides, when segmenting the customer groups, firms should consider the way customers want to contact them. For example, some buyers in a segment would rather order over the web, but some others prefer to buy in another way (Dwyer & Tanner, 2005). 33

35 To understand which way of contact the customers in each segment prefer, companies can track customers response to earlier contacts. This can be done by constructing a database of information about the customers and then investigate which customers responded well to which way of contact in the past. For example if some customers have bought some items frequently online, they are more likely to buy in this way in the future (Dwyer & Tanner, 2005; Ghavami & Olyaei, 2006). According to Winer (2001), a business database should contain the following information: Customer Contacts: This includes all customers and the organization initiated contacts. Transactional information: This kind of information includes comprehensive and detailed customers' purchase history. Response to Marketing activities: to know whether or not the customers responded to direct contacts. Descriptive Information: This is more used for customer segmentation and analyzing the collected information. This author also argues that in data collection and creating a database, the biggest challenge is to create opportunities for communication between firms and customers especially when the firms use intermediaries such as physical stores that prevent direct contact. Winer (2001, p. 93) illustrates the general problems in creating a database in figure 2.2. For example, in this figure, banks have many direct communications with customers; thus, creating a database is an easy for them. On the other hand, this job is very hard for the lower right-hand quadrant due to the indirect communications these firms have with their customers. 34

36 Customer Interaction Direct Indirect Banks Airlines High Telecom Packaged goods Interaction Frequency Low Retail Personal Computers Internet Infrastructure Drugs Furniture Autos Fig. 2.2 Getting more customer interaction Reference: Winer (2001, p. 93) The collected data in database should then be analyzed through statistical methods such as cluster analysis or through "lifetime customer value" (LCV), click stream analysis, etc. The purpose of these analyses, especially LCV, is that each customer must be analyzed in terms of profitability to the company; therefore, companies can realize which customers to target more efficiently (Winer, 2001, p. 94). Different results can be gained through these analyses. For example, if purchasing behaviors of the customers have been analyzed, the customers in the highest purchasing rate are the first priority to focus in loyalty programs. The final goal is to select long-term profitable customers among all existing ones (Winer, 2001). In the next step, a strategy for communication with customers will be designed. In this strategy, several different communication channels such as and direct mail will be used (Dwyer & Tanner, 2005). Many academicians have suggested organizations to dialogue with their customers through these communication channels rather than communicating with them through mass media such as television or radio. Also, they say Internet can be deployed as a technological tool to facilitate individual relationship building with customers. 35

37 Meanwhile, based on the customers' profiles and their buying behaviors, companies should make offers such as purchasing discounts (Dwyer & Tanner, 2005). In the third step, the strategy is implemented and then in the final step, it is evaluated. Measures of performance are considered for evaluation. Customer satisfaction is the most important one. Depending on the results, companies may change their customer segmentations and strategies (Dwyer & Tanner, 2005). Keramati et al. (2010) classify CRM processes into operational CRM and management CRM processes. They explain that operational CRM processes include the CRM's tasks in a firm's value chain, but management CRM processes are about all CRM's tasks related to the administration, allocation, and control of all resources in firms. By dividing CRM processes into these two main classes, organizations become able to determine the CRM activities under each one more efficiently Operational CRM processes According to Zablah et al. (2004) and Keramati et al. (2010), those customer processes found at the operational level of the business are called operational CRM processes which themselves are classified into operational knowledge processes and operational interaction processes. The above authors explain that the first sub-process involves the tasks associated to customer knowledge at the operational level. Collecting customer data through different channels and disseminating customer knowledge between the firms and their customers is an important example for this process. On the other hand, the operational interaction processes consist of different activities from the activities done in the first process. Based on the studies done by the mentioned researchers, using these processes, firms consider the available intelligence to establish and develop relationships with customers Management CRM processes Based on Greve and Albers (2006) and Keramati et al. (2010), all the activities which refer to the activities programmed to create market intelligence and improve decision making in allocating the resources, developing new products, and so on are explained in the context of management CRM processes. An important example of these 36

38 activities in this context is the realization of customer behavioral pattern which has significant effects on decision making and therefore on the firm performance CRM process capabilities Coltman (2007) and Keramati et al. (2010) argue that capabilities are intermediate transformation ability between all the resources (technological and infrastructural) and the organization's mission. Capabilities measure the organization's ability in effective combination of resources to meet its goal; therefore, they approximately reflect the concept of efficiency. The same authors believe that since capabilities have mediating effect between firm's resources and performance, they are not observable; therefore they become hard to imitate. Process capabilities can also be regarded in the context of CRM. Keramati et al. (2010, p. 1177) define the CRM process capabilities as: "the process abilities that are gained by effectively applying CRM resources to CRM processes and that enable firms to create superior value for their organization as well as their customers.". Regarding this definition, we can conclude that the best level to assess the CRM process capabilities is the process level. Like CRM processes, these capabilities are classified as management CRM process capabilities such as new product development capabilities and operational CRM process capabilities such as customer support capabilities. Keramati et al. (2010, p. 1176) summarize the relationship between CRM resources, CRM processes, CRM process capabilities, and organizational performance in figure

39 Fig. 2.3 CRM value generation process Reference: Keramati et al., (2010, p. 1176) As previously mentioned, process capabilities have mediating effect between firm's resources and performance to meet a specific objective. This study investigates the impact of CRM resources and processes on customer retention (as the organizational performance); therefore, to attain this objective, firms should implement a comprehensive set of relationship programs with their customers as their process capability. Winer (2001) calls these programs as 'customer retention programs of CRM' Customer retention programs of CRM As it has been discussed before, the overall objective of relationship programs is delivering a higher level of customer satisfaction than similar companies do. Therefore, companies must measure their customers' satisfaction levels and improve their relationship programs which help to deliver products and services beyond the customer expectation (Winer, 2001; Farquhar, 2004; Arbore & Busacca, 2009) and therefore contribute to retain the customers. Winer (2001) believes that if firms want to efficiently retain their profitable customers, they need to implement a comprehensive set of relationship programs including: frequency/loyalty programs, customer service, customization, and community building. He states these programs are determinants of customer satisfaction which finally leads to customer retention (figure 2.4). 38

40 Frequency/ Loyalty Programs Customer Service Customer Relationship Management: Satisfaction Customization Rewards Program Community Building Fig. 2.4 Customer Retention Programs Reference: Winer (2001, p. 98) Loyalty/Frequency Programs Loyalty program which is also called frequency program provides rewards to targeted consumers in order to encourage them to buy products repeatedly. These programs have become competitive necessities for the firms in the marketplace especially those with a decreasing rate of customer retention. Recent studies have also shown that the most successful retailers in the top sectors such as department stores, grocery, drugstores category, and mass merchandisers design these programs (Winer, 2001; Verhoef, 2003; Juan & Yan, 2009). Customer loyalty in service industry consists of three main dimensions including affective loyalty, behavioral loyalty, and cognitive loyalty. Behavioral and affective loyalties are formed based on cognitive loyalty which itself is influenced by company's reputation, brand recognition, its public image and most importantly, corporate social responsibilities. On the other hand, in the intermittent service industry, customer satisfaction, service quality and customer perceived value (CPV) 39

41 are the main factors affecting customer loyalty. Through the moderation of customer satisfaction, both of CPV and service quality affect Customer loyalty indirectly (Farquhar & Panther, 2008; Juan & Yan, 2009). In descending order of significance, affective loyalty is influenced by customer satisfaction, cognitive loyalty, customer perceived value, and service quality. Moreover, determinants of behavioral loyalty include customer satisfaction, service quality, affective loyalty, customer perceived value, and cognitive loyalty. And finally, the only influencing factor of cognitive loyalty is the service quality (Juan & Yan, 2009). Figure 2.5 illustrates the influencing factors of customer loyalty in service industry. Customer Perceived Value Cognitive Loyalty Service Quality Customer Loyalty Affective Loyalty Customer Satisfaction Behavior Loyalty Fig. 2.5 Influencing factors of customer loyalty Adapted from: Juan & Yan (2009, p. 68) Many academicians believe that there are some considerable problems with loyalty programs such as the need for high investments and the difficulty to correct mistakes. There is confusion about whether these programs improve the customer retention or average spending behavior, and most considerably, it is not easy to gain competitive advantage by these programs. However, loyalty programs can work effectively if firms increase the switching costs and build barriers to entry (Winer, 2001; Deng et al., 2009; Juan & Yan, 2009). 41

42 Customer Service Generally, a customer service encounter is defined as the contact or "touch points" that a customer has with a company. It has the potential to improve customer strategy or to have the opposite effect. Because the targeted customers are most valuable customers to the companies and they are able to choose their favorite item among a variety of products in the marketplace, customer service must be the first priority within each company (Winer, 2001, p. 99; Venetis & Ghauri, 2004; Smith, 2006). There are two types of services including Reactive and Proactive services by which companies can enhance their customer service (Winer, 2001; Smith, 2006). Reactive services are applied when a customer faces problem such as product failure or question about a bill and contacts the firm to solve it or get more information about it. Today, most organizations especially the banks have used the technology and also trained their employee to deal with these situations through s, faxback systems, telephone, etc. Proactive services are the situations where the firms do not wait for customers to contact them and are determined to hold a dialogue with their customers prior to use reactive solutions such as complaint handling. This can be done well where the sales force or other employees dealing with the customers are trained and interested to realize customers' expectations (Winer, 2001; Venetis & Ghauri, 2004; Smith, 2006; Arbore & Busacca, 2009). A number of systems leveraging the Internet assist both of these two types of services. For example, some Web-based service providers such as HumanCIick provide their customers with the ability of real time interactions with service representatives (Winer, 2001; Smith, 2006). Community Using the Web, both online and offline companies can build online networks of customers in which they are able to exchange information about the products and also can interact between themselves and the company easier and more personally. This network of customers is called community. In this situation, the customer is committed to the company and therefore he or she is less likely to leave the family of customers. In addition, by creating these communities, companies can gain more accurate information about specific products in a market and then make it with the 41

43 characteristics or the quality that customers expect (Stauss et al., 2001; Winer, 2001; Guo et al., 2009). Customization Customization goes beyond communicating with customers and is also about the creation of products for individuals. For example, some companies such as Dell and Nike have established online processes for creation of customized and personal products and services based on individual customer's taste. In this way, each customer is able to choose a product from a list or order the item he or she really wants (Winer, 2001; Ahmad & Buttle, 2002; Farquhar, 2004). Academicians call such customization, "versioning" which is easy and cheap to do. The point that should be noticed is that versioning is easier to do for services and intangible products than for physical products; however companies can use the additional information gained from customers to tailor at least the appearance of products (Winer, 2001; Farquhar, 2004; Smith, 2006) Summary of studies linking CRM to customer retention The coordination and integrity between different components of CRM have a significant effect on CRM value creation for the organizations; therefore, to be able to investigate the real value generation process of CRM, its different components should be aligned and linked together. Many researches have been conducted to empirically and theoretically examine the value process of CRM such as its impact on firm performance. We have discussed the concept of 'customer retention' and its importance for firms in the previous chapter and also in the first part of this chapter. In this section, a summary of our literature survey of studies linking CRM to customer retention (as the considered firm performance in this study) is presented. Table 2.1 provides this review. Besides, table 2.2 exhibits the analysis of these studies based on addressed or not addressed important issues from the viewpoint of this research. As it can be noticed, some of these addressed issues are more technology-oriented while others are more strategic. However, to have a broader view, the important issues which have not been addressed in these studies are brought, too. 42

44 Table 2.1: Summary of studies linking CRM to customer retention Reference CRM components Mediators Findings Winer (2001) Customer retention programs of CRM: - Loyalty programs - Customization - Customer service - Community building - Customer satisfaction - Customer retention - Each of the loyalty programs, customization, customer service, and community building affects customer satisfaction significantly. - Customer satisfaction has a great impact on customer retention. Lüneborg and Nielsen (2003) - IT knowledge - Inter-firm cooperation - Customer-focusing technology usage - Organizational size - There is a positive relationship between IT knowledge and customer-focusing technology usage. - Customer-focusing technology affects customer retention. - The effect of Inter-firm cooperation on performance in large banks is more than small banks. - The association of organizational size with the relationship between customer-focusing technology and customer relationship outcome (such as customer retention) is negative. Verhoef (2003) RMIs: - Loyalty programs -Direct mailing - Customer retention - Commitment has a significant effect on customer retention. - RMIs (Loyalty programs and direct mailing) affect customer retention. Eng (2004) - Industry attractiveness - Resource advantage of customer portfolio - Long-term value of customer portfolio - The correlation between industry attractiveness and resource advantage of customer portfolio is positive and considerable. - There is a weak positive correlation between long term value of customer portfolio and customer retention. Reinartz et al. (2004) Customer-facing level of CRM process: - Relationship initiation - Relationship maintenance - Relationship termination - CRM technology - Organizational alignment - Only for initiation and maintenance level CRM process implementation associates better customer retention. - A CRM-compatible organizational alignment plays the role of moderator between CRM processes and firm performances (such as customer retention). - Using large CRM technology proportion does not affect customer retention as it is expected. Wang et al. (2004) - Customer value - Customer satisfaction - Customer value (customer's perception) is a determinant of customer satisfaction and customer loyalty. - There is a positive relationship between customer satisfaction and customer retention. Roh et al. (2005) - Process fit - Quality of customer Table information 2.1: (continued) - Efficiency - Customer satisfaction - Among CRM initiatives, process fit is the only one that has direct effect on performance. - CRM elements influences efficiency and efficiency has significant impact on customer satisfaction - Customer satisfaction affects customer retention. 43

45 Table 2.1: (continued) Greve and Albers (2006) - CRM technology - Top management commitment - CRM orientation - Organizational alignment - Customer heterogeneity - CRM activities - Use of CRM technology affects performance indirectly (including customer retention). - Top management commitment and customer orientation only affect the retention performance. - Except customer orientation and top management commitment, all variables affect each of initiation, maintenance, and retention performances directly. Coltman (2007) - Superior CRM capabilities - Reactive and proactive market orientation - CRM capability affects customer retention. - The relationship between CRM capability and customer retention is partially mediated by proactive market orientation, but reactive market orientation does not show a significant mediation impact. Keramati et al. (2009b) CRM resources: - Technological resources - Human resources - Organizational resources - Firm performance (including customer retention) - Technological CRM resources affect organizational performance (such as customer retention) when they are aligned with infrastructural resources. - Deployment of all CRM resources leads to better customer retention. Becker et al. (2009) - Technological implementation of CRM - Organizational implementation of CRM - Customer acquisition - Customer maintenance and retention - CRM implementation impacts customer retention only if the appropriate company stakeholders support it adequately. - Customer heterogeneity impacts customer retention rate significantly. - The effect of technological implementations with employee support on performance is significant and positive for both the initiation and the maintenance aspect. Kim et al. (2010) - Organizational resources - Organizational capability for CRM implementation - Implementation of enterprise CRM - Customer retention - Relationship expanding - While the integration of organizational resources and capabilities (analytical and operational capability, and information resources) may impact retention process, they are not enough to provide desired firm performance of acquiring new customers and improving relationships with them. Keramati et al. (2010) - CRM resources - CRM processes - CRM process capabilities - Firm performance (such as customer retention) - CRM resources (Technological, human, and organizational resources) positively affect CRM processes. - The extent of affected CRM processes is directly associated with improved CRM process capabilities. - There is a positive association between CRM process capabilities and firm performance. 44

46 Table 2.2: Analysis of studies linking CRM to customer retention Reference Addressed issues Not addressed issues Winer (2001) Lüneborg and Nielsen (2003) Verhoef (2003) Eng (2004) - Addresses the relationship between customer satisfaction and customer retention. - Specifies all the determinants of customer retention and gives them in the form of a comprehensive set of customer retention programs. - Discusses CRM processes broadly. - Measures significant IT related capabilities (for example information analysis skills) from the view of RBV. - Addresses two levels of organizational outcome. The first level which is customer-focusing technology usage measures these capabilities as market orientation and frontline support. - Investigates the impact of relationship marketing instruments (RMIs) and customers relationship perceptions (CRPs) on customer retention. - Tries to find out whether different variables of RMIs and CRPs affect customer retention. - examines the effects of three significant strategic views in attaining customer performance (including customer retention). - Emphasizes on the scales measuring the specified three constructs. - The variable of competitive characteristics is examined as an important indicator of industrial view. - Focuses on suggesting a comprehensive set of customer retention programs; As a result, the significance and priority of each program have gone unnoticed. - It does not say that which program affects customer retention directly and which program affects it indirectly. - Organizational resources are not addressed enough. - While investigates the relationship between adoption of CRM and customer performance, it does not address the way of relating this adoption to performance. - Does not pay attention to CRM processes. - Many human and organizational aspects are not considered. - Human aspect of CRM especially employees are not addressed. - Management CRM processes are not considered. - It does not directly address CRM and its applications. Reinartz et al. (2004) - Specifies and assesses the impacts of attractive three stages of CRM processes at customer-facing level for practitioners. - evaluates the impact of technological CRM resources on the relationship between CRM processes and organizational outcome (including customer retention). - Measures some company performances such as customer retention. - Considers industry as a control variable and say that benefits of CRM such as customer retention are approximately the same among different industries. - Evaluates the relationship between CRM processes and firm outcome (including customer retention) directly. However, it does not address the benefits gained by CRM which improve the customer retention. Wang et al. (2004) - Measures behavioral outcomes of CRM such as emotional value from customer's view and relates them to customer retention. - Focuses on behavioral variables of CRM performance and does not mention how they are created and help to retaining existing customers. Roh et al. (2005) - emphasizes on important aspects of CRM (for example CRM systems and customer information). - It argues why the CRM system affects profitability and customer retention by considering efficiency aspects of the CRM system. - The emphasis of the research is on the technological resources of CRM and as it is obvious, the considered efficiency aspects are only from technology perspective, therefore, they are not enough to enhance the customer retention. - The study does not address the people aspect in CRM programs. 45

47 Table 2.2: (continued) Greve and Albers (2006) - Assesses the impact of technological CRM resources on organizational outcome. - Moreover, it views top management commitment and customer orientation as aspects of CRM. - It measures CRM performance including customer retention at three stages of CRM processes. - It does not pay enough attention to human resources of CRM. For example it does not consider employee experiences. - Determines the factors affecting customer retention, but it does not say why these effects happen. Coltman (2007) - Evaluates the resource capabilities of CRM based on RBV. - By considering the four perspectives of the balanced scorecard, it assesses organizational outcome. - Investigates the mediating effect of conversion feasibility (i.e., the organization's capability in following the specified path when the CRM resource capabilities are built) on the benefits gained by CRM. - Since it evaluates the CRM resources too broadly, some of CRM aspects about customer retention (such as efficiency) are not considered. - CRM processes are not addressed directly. Keramati et al. (2009b) Becker et al. (2009) Kim et al. (2010) Keramati et al. (2010) - Investigates the role of technological, human, and organizational resources of CRM on firm performance including customer retention based on RBV. - Addresses and discuses the important role of human and organizational CRM resources and concludes that without efficient use of these two resources, investments and large use of IT and CRM technology can not lead to expected performance. - Investigates what effect firms can expect CRM to have on customer retention and how this effect can be leveraged. - Presents a conceptual model that tries to find the relationship between organizational and technological resources. Also, the impact of management and employee support on CRM process-related outcome is addressed. - Measures performance in terms of the initiation, maintenance, and retention of gained by CRM. - Considers information resource and analytical capability as the organizational resources and capabilities and talks about the impact of integrating them on customer retention. - Addresses the technology, process, and people aspects in implementing a successful customer retention strategy. - Proposes a framework to investigate the relationship between CRM resources, CRM processes and process capabilities, and firm performance based on RBV and process oriented approach. - Emphasizes on the importance of human and organizational resources of CRM and aligning them to IT and CRM technology. - Evaluates the effect of all CRM resources on CRM processes. - Discusses CRM processes and CRM process capabilities on both operational and management processes. - This study focuses on the role of CRM resources based on RBV; therefore, it does not pay enough attention to CRM processes. - This research emphasizes on the effect of technological and organizational CRM resources on customer retention, but does not pay enough attention to human CRM resources. - The capabilities of firm in CRM implementation is discussed enough, but the CRM process capability is not general discussed. - It does not examine CRM and its applications directly and deeply. - Customer facing level of CRM process is not discussed directly. 46

48 3. Frame of Reference 3.1. Introduction In the previous chapter, we presented a literature survey of several studies linking CRM to customer retention and elaborated on their important addressed and not addressed issues from the viewpoint of this study. The two models presented by Winer (2001) and Keramati et al. (2010) appear as the most suitable ones for this thesis. In fact, in comparison with other studies, they enable the researcher to test all CRM resources and comprehensive set of customer retention programs with regard to firm performance and most importantly customer retention. Also these models are found to be the best for collecting data from banks employees working in electronic banking section in Iran. In addition, using these models, measuring the effectiveness of CRM resources, processes, and retention programs on customer retention in the mentioned environment becomes easier and more accurate. These models were fully explained in the previous chapter. We will consider them as a reference to state our research hypotheses. They are presented again in this chapter to be combined as the final research model for this thesis. In this chapter, by gaining insight from the RBV and the process-oriented approach, we propose a framework to investigate the relationship between CRM resources, CRM processes, customer retention programs of CRM (as the CRM process capabilities), and finally customer retention as the result of these programs. Thus, in a broad view, we will be investigating the impact of CRM on customer retention. The dimensions of each of the constructs have been specified and the research hypothesizes are formulated Resource-based View of the firm (RBV) RBV emphasizes on all the firm resources and sees them as the specific and valuable assets. According to many academicians if organizations want to achieve competitive advantage, they should compete based on unique corporate resources. Thus, two 47

49 tangible and intangible resources which are complementary to CRM technology and also not duplicable by competitors, namely as human and organizational resources are introduced by researchers and practitioners (Keramati et al., 2010) Technological CRM resources As mentioned, CRM technologies are the tools enabling the firms to get the right information to the right person at the right time and are divided into three parts of communicational, operational and analytical technologies (Keramati et al., 2010) Infrastructural CRM resources Based on Keramati et al. (2010), CRM is a marketing strategy and its infrastructure is based on the non-technological CRM resources which academicians call 'infrastructural CRM resources'. They are classified into human CRM resources and organizational CRM resources. Human CRM resources Human CRM resources consist of customer facing and non-customer facing employees' attitudes, technical skills, and ability to convert customer data to knowledge. Academic researchers claim that among all CRM resources, human resources are the most important ones (Keramati et al., 2010). Organizational CRM resources Some researchers such as Keramati et al. (2010) emphasize that to implement a successful CRM program, some changes in culture of firms are required. For example, the firms must consider some elements including training the employees, and top management commitment. Our first hypothesis related to this aspect is the following: H1. CRM resources have direct effect on CRM processes Process-oriented approach The process-oriented approach examines the impact of information technologies on intermediate business process. Indeed, a process management orientation is a 48

50 necessity for a CRM program to be successful; therefore, to ensure the efficient use of firm resources toward the creation of expected firm performance, managers must focus on CRM processes effectively (Keramati et al., 2010) CRM processes According to Keramati et al. (2010), CRM processes can be divided into operational CRM processes and management CRM processes. They explain that operational CRM processes include the CRM's tasks in a firm's value chain, but management CRM processes are about all CRM's tasks related to the administration, allocation, and control of all resources in firms. Based on this study, the following hypothesis can be developed: H2. CRM processes have direct effect on customer retention programs of CRM. Figure 3.1 illustrates the CRM value creation process and also the relationship between CRM resources, CRM processes, CRM process capabilities, and the organizational performance. Fig. 3.1 CRM value generation process Reference: Keramati et al., (2010, p. 1176) 3.4. Customer retention programs of CRM As it has been discussed in the previous chapter, process capabilities have mediating effect between firm's resources and performance to meet a specific objective. This research considers the customer retention as the final firm performance and investigates the impact of CRM resources, processes, and customer retention programs of CRM (as the CRM process capabilities) on customer retention (as the 49

51 organizational performance). Winer (2001) believes that if firms want to efficiently retain their profitable customers, they need to implement a comprehensive set of relationship programs including: frequency/loyalty programs, customer service, customization, and community building. Loyalty/Frequency Programs These programs provide rewards to targeted customers in order to encourage them to buy or demand for services repeatedly. However, loyalty programs can work more effectively if firms increase customer switching costs and build barriers to entry (Winer, 2001). Customer Service Customer services are two types including Reactive and Proactive services. Reactive services are deployed by firms when a customer faces problem such as product failure or question about a bill and contacts the company to inform and solve them. Proactive services are used when the firms do not wait for customers to contact them and are determined in holding a dialogue with their customers prior to use reactive solutions such as complaint handling (Winer, 2001). Community Community is online network of customers that firms build it in order to enable their customers to exchange information about the products or services and also interact between themselves and the company easier and more personally (Winer, 2001). Customization Winer (2001) states that customization is about the creation of products and services based on each customer's taste. In this way, each customer is able to choose a product from a list or order the item with the favorite quality or characteristics. Therefore, the final formulated hypothesis is the following: H3. Customer retention programs of CRM have direct effect on customer retention. 51

52 3.5. Research model Considering the above mentioned approaches, programs, and hypothesizes, figure 3.2 proposes the conceptual model for the impact of CRM on customer retention in electronic banking of Iranian banks. Technological CRM Resources Infrastructural CRM Resources CRM Processes Customer retention programs of CRM Customer Retention Fig. 3.2 The research model 51

53 4. Methodology 4.1. Introduction In this chapter, we present and describe the procedure and the methods that have been used to conduct this work. The research purpose, research approach, and the research strategy used in this thesis are also described and justified. Furthermore, the data collection, sample selection, questionnaire design, the ways to evaluate the validity and reliability of the obtained results will be discussed Research approach Qualitative versus Quantitative Approach According to Hair et al. (2007, pp ), there are two types of research approaches available to researchers, namely quantitative and qualitative. The data in quantitative approach are numbers and lends itself to statistical analysis in order to imply the characteristics of something. An important point to consider is that structure, representativeness, and providing objectivity are important strengths of quantitative research. This approach provides objectivity because the respondents are the ones who provide the numbers; therefore researcher's opinion does not have any impact on testing the hypothesis. In a general sense, this approach is used in explanatory researches. Furthermore, it allows generalization and enables the researcher to predict the future. The data in qualitative approach is generally collected by observation or unstructured interviews. They are usually in the form of words, phrases and pictures. It is important to say that this approach provides a deeper understanding of the phenomenon the researcher wants to find and therefore furnishes a holistic view. Also, because judgment has an important role in this approach, it is difficult to replicate the findings. Using this approach, the researcher can choose small sample sizes. Moreover, the author will have less concern about the representativeness of the results (Hair et al., 2007, pp ; Saunders et al., 2009, p. 482). 52

54 From the above discussion and since the results are presented by digits and they are analyzed by statistical methods, we can conclude that quantitative approach proves to be suitable for this research Deduction versus Induction Approach In another classification, there are two main approaches which are: deductive and inductive approaches. In the deductive approach, the researcher develops a theory or hypothesis and then designs a strategy to test it. This approach is more discussed in natural sciences and scientific principals where the explanation, anticipation of phenomena and the prediction of their occurrence are presented by laws. Deduction is a highly structured approach in which there is the search to explain causal relationships; the data are quantitative and they are collected through samples of sufficiently large size to generalize the results. It is clear that in this situation, the researcher is independent of the subject under study (Hair et al., 2007; Saunders et al., 2009, pp ). In the induction approach, the collected data are analyzed and as a result, a theory will be formulated and developed. This approach tries to gain a close understanding of the concepts humans attach to events. Inductive researches use a more flexible structure in which the researcher is a part of that research and collects the qualitative data through a small sample size; therefore there is not much concern with the generalization (Hair et al., 2007; Saunders et al., 2009, pp ). In the present work, we have developed some hypothesizes and designed a strategy to test them. On the other hand, the data we used are quantitative and they were collected through a large sample. Hence, we can say that the deductive approach is the one used for this study Purpose of the Research The way a researcher asks the research question would result in a descriptive or an explanatory research, or both of them. Considering the research question, a researcher definitely must think about the research purpose which is classified as exploratory, descriptive and explanatory. However; Academicians believe that a research may 53

55 have more than one purpose which may change over time (Zikmund, 2000; Saunders et al., 2009, p. 139) Exploratory research Exploratory research is useful when there is little information about the problem or when the research questions are vague. This kind of research is also useful to discover new patterns, ideas, relationships and so on; therefore this research is not appropriate to test specific hypotheses. It should be mentioned that this design relies more on a qualitative approach; however, usage of quantitative techniques is possible, too. The great advantage of an exploratory study is that it is flexible. It means the researchers using this design, can change their direction when a new insight occurs to them (Hair et al., 2007, pp ; Saunders et al., 2009, pp ). Based on Saunders et al. (2009, p. 140), there are three main principal ways to conduct an exploratory research as. They include: - Searching the literature; - Interviewing focus group; - Interviewing the 'experts' Descriptive research Sounders et al. (2009, p. 140) state: "The object of descriptive research is to portray an accurate profile of persons, events, or situations". Descriptive research is a useful design to obtain the data describing the characteristics of something; thus the questions starting with who, when, how, which, why, and what can be answered by this research design. In this situation, the researcher should know the phenomena very well before collecting the data (Hair et al., 2007; Saunders et al., 2009). Using descriptive research, the researcher can test the hypothesis. Also the data is usually collected through structured observation or interviews asking structured questions. To analyze these data, descriptive research uses descriptive statistics which include frequency counts, measures of central tendency, variations and so on. (Hair et al., 2007, p. 155; Saunders et al., 2009, p. 140). 54

56 Explanatory research Explanatory researches study a problem or a situation in order to finding out the relationship between dependent and independent variables. Explanatory research is the most appropriate design to test whether one variable causes or determines the value of another. For example, does a change in variable X (CRM) cause a change in variable Y (customer retention)? (Hair et al., 2007, p. 160; Saunders et al., 2009, p. 140). Hair et al. (2007, p. 160) explain the four conditions a researcher looks for in testing these relationships as: 1. The time sequence, which means occurring the cause before the effect. 2. Covariance: which means changing of cause is associated with changing the effect. 3. Organization association: meaning that the relationship is true and is not because of another variable affecting the dependent and independent variables. 4. Theoretical support: The existence of the relationship between variables is supported by a logical explanation. Hence, considering the purpose and the research question of the present work, we can claim that this research is mainly explanatory. In addition, because the data is collected through a questionnaire and is used to find out the relation between the dependent and independent variables of this study, we can conclude that it is explanatory Research Strategy Research strategy is a plan showing the way how the research goes on and how the research questions will be answered. Different research strategies can be used for explanatory, descriptive, and exploratory researches. Also, some of them belong to inductive approach, others to the deductive approach. However, researcher's choice of a research strategy depends on the research question, available time and resources (Zikmund, 2000; Saunders et al., 2009, p. 141). 55

57 Yin (1994) has identified five research strategies including experiment, survey, case study, archival records, and history. He has distinguished these strategies according to the three criteria: form of research question, the extent of control a researcher has on actual behavioral events and also the degree of focus on contemporary events. These conditions for each research strategy are illustrated in table 4.1. Table 4.1: Relevant Situation for Different Research Strategies (Yin, 1994, p. 6) Form of research Requires control Focuses on Strategy question over behavioral contemporary events events Experiment How, Why Yes Yes Survey Who, What, Where, No Yes How many, How much Archival Analysis Who, What, Where, No Yes/No How many, How much History How, Why No No Case study How, Why No Yes The purpose of an experiment strategy is to study the casual relationships. It tends to be used in explanatory and exploratory research and can answer 'how' and 'why' questions. The survey strategy is usually used in deductive approach and allows the researcher to collect quantitative data. It is also the most popular strategy to answer who, what, where, how many and how much questions. Thus, it tends to be used in descriptive and exploratory research (Saunders et al., 2009, pp ). Another research strategy is 'case study'. Saunders et al. (2009, p. 145) state that case study is: "a strategy for doing research which involves an empirical investigation of a particular contemporary phenomenon within its real life context using multiple sources of evidence". This strategy is very appropriate to answer the 'how' and 'why' questions and very useful for exploratory and explanatory research. 56

58 Archival research is another research strategy which uses recent as well as historical documents and records as the principal source of data. Archival research enables the researcher to answer the research questions which focus upon the past and changes over time, be they explanatory, exploratory or descriptive and tries to answer the Who, What, Where, How many, and how much questions (Saunders et al., 2009, p. 150). History is the final strategy which does not deal with contemporary events and there is no need to control the behavioral events. It is also an appropriate strategy to answer the 'how' and 'why' questions (Saunders et al., 2009). Since in this study we want to investigate the impact of CRM on customer retention in e-banking of Iranian banks from employee's perspective and also as the main question of this research is the form of 'what', survey seems to be the best research strategy for this study in which we are using a questionnaire Time Horizon In planning a research, the following question is important 'is this research done in a particular time or it is done over a given period'. According to Sounders et al. (2009, p. 155), a study of particular phenomena taken at a particular time is a cross sectional research. This study often employs the survey strategy; however, it may also be used in qualitative approaches. On the other hand, those studies taking place over time are called longitudinal studies. Their main strength is their capacity in studying change and development (Saunders et al., 2009, p. 155). Investigating the impact of CRM on customer retention in electronic banking is conducted in a limited period of time; so we can conclude that this study is a crosssectional research Data Collection method Zikmund (2000) and Sounders et al. (2009, p. 256) say there are two classifications for collected data which are: primary and secondary data. Primary data can be collected for instance through interview, observation, and questionnaire. On the other 57

59 hand, secondary data is the information collected from the studies done before and can be collected from the Internet or libraries. Sounders et al. (2009, p. 360) state that questionnaire is one of the most widely used techniques to collect data within the survey strategy and since each respondent answers the same set of questions, it is an efficient technique of gathering responses from a large sample. For this study, primary data seems to be the most suitable one and because our research strategy is survey and also the sample is large, we used the questionnaire as the instrument to collect the primary data in this research. Our questionnaire has been distributed to the top managers in each selected bank branch. To increase the response rate, meeting was the first method to give the questionnaires. Due to the time and accessibility limitations, some of the respondents received the questionnaires by along with an explanation letter Sample Selection Ideally, a researcher would like to obtain the data from all members of the population. However; it is almost impossible; therefore a small subset of the population which must be representative of all the members in that population is drawn. This subset which mirrors the characteristics of population is called sample (Zikmund, 2000; Hair et al., 2007, p. 170). According to Hair et al. (2007, p. 171), a set of well defined procedures to obtain a representative sample is as follows: 1. Defining the population under investigation (The complete group of elements relevant to the research). 2. Determining the appropriate sample frame (a complete list of all the elements in that population). 3. Sampling method selection. 4. Calculation of the sample size. 5. Implementation of the sampling plan. There are two broad categories of sampling methods available including probability and non-probability methods. Both of these methods are explained in details below. 58

60 Probability Sampling According to Hair et al. (2007, p. 175) and Saunders et al. (2009, p. 214), drawing a probability sample is based on this premise that each element in the population has a known and nonzero probability of being selected. Also, in this method selected samples are usually large to be representative of the population; therefore, with a specified level of confidence, the findings can be generalized to the population under investigation. Probability sampling methods are classified as Simple random sampling, Systematic sampling, Stratified sampling, Cluster sampling, and Multistage sampling. Simple random sampling In the simple random sampling (sometimes called just random sampling), all the elements in the population have equal chances of being selected. In this method, the resulting sample is representative of the population if its calculated size is sufficiently large. It is also an appropriate method for a geographically dispersed area if the researcher uses an alternative data collection technique such as telephone interviewing or online questionnaires (Hair et al., 2007, pp ; Saunders et al., 2009, p. 222). Systematic sampling In systematic sampling, an initial starting point on a list is randomly selected and after that, the researcher selects every n th element in the sampling frame. Thus, to draw the sample, we should first calculate the sample size and the sampling interval. This method is useful for geographically dispersed cases only if face to face contact in not required. Also, using this sampling method, the researcher can obtain representative data (Hair et al., 2007, p. 177; Saunders et al., 2009, p. 226). Stratified sampling To use this method, the sampling frame must be divided into distinct and relatively homogeneous subgroups which are called strata. This is usually done by the researcher's past experience or specified by the client. Then the researcher determines the size of total sample and the size for each individual strata. The composite of the samples taken from the strata shapes the stratified sample and to select the elements of this sample, systematic or simple random samples of the strata of the population must 59

61 be drawn. The strength of this method is that it helps increasing the accuracy of the sample information (Hair et al., 2007, p. 178). Cluster sampling In this method, the researcher views the population as made up of heterogeneous groups, each of them is called a cluster. A cluster group can be for example geographic areas, households, firms, and so on. It should be noted that cluster sampling can produce representative data if it is done properly. It is important to say that in this method the sampling frame is the list of clusters rather than a list of individual elements in the population. In order to increase the representativeness of the sample, the researcher must increase the number of sub-areas (Hair et al., 2007, p. 180; Saunders et al., 2009, p. 230). Multi-stage sampling Multi-stage sampling which is also called Multi-stage cluster sampling is a development of cluster sampling method. This method is useful when the population is a large geographical region and face to face contact with the elements is needed or it is hard to construct a sampling frame for that region; however, it can be used for not geographically based discrete groups (Hair et al., 2007, p. 181; Saunders et al., 2009, p. 231) Non-probability Sampling Hair et al. (2007, pp ) explain that in this technique, the probability of each element in the population is not known and the selected sample is not necessarily representative of the population statistically. In this situation, to select the elements in the sample the researcher uses expert judgment, experience, and convenience. Therefore, unlike probability samples, the results can not be generalized to the population (Hair et al., 2007, pp ). These authors also discuss that the most common types of non-probability sampling techniques include Convenience sampling, Judgment sampling, Snowball sampling, Self selection sampling, and Quota sampling. As it has been mentioned before, the population of this survey is all Iranian banks' managers (the top manager in each branch). However, because of the time limitation, 61

62 we limited this population to all the top banks' mangers in the province of Tehran which is the heart of financial activities. In addition, according to Bank and Insurance Manifest agency of Iran (2010), it has the highest rate of e-banking usage. In the next level, among all of the mentioned banks, we selected four of them including two governmental and two private banks. Among governmental banks, bank Refah and bank Maskan and among private banks, bank Pasargad and bank Saman have been chosen. We chose these banks because they are old and in comparison with other banks they are more pioneering in e-banking. In addition, they have the greatest number of branches in Iran and have been recognized as the most effective banks in e- banking context in Iran. In the final level, we used simple random sampling to choose the branches of each bank to distribute the questionnaires to the respondents with equal chances of being selected. Therefore, according to the definitions of each sampling method in the previous section, we have used a probability stratified sampling method Sample size: Determining an efficient sample size is of great significance in any research. This is because too small samples may lead to inaccurate results, while samples that are too large may waste time and resources (Hair et al., 2007, p. 182). According to Azar and Momeni (2005), research populations are divided into two kinds: restricted and unrestricted. Since banks have restricted number of managers, our research population will be restricted. Therefore, we have used restricted sampling formula (see below) to calculate the sampling size. Table 4.2 presents the descriptive statistics in sampling. In this research, first we run 30 questionnaires (shown with N in table 4.2) and then a sample size is calculated. Table 4.2: Descriptive statistics in sampling Descriptive Statistics N Minimum Maximum Mean Std. Deviation Variance TOTAL Valid N (listwise) 30 61

63 According to this table, the calculated value for standard deviation (σ) is and the value for variance is.385. These values are very important for calculation of suitable sample size using the formula presented by Azar and Momeni (2005) (see below). In this formula, n is the total sample size for this research. Also total population is shown by N. In this research, the total population is sum of the numbers of all banks' branches under study. The number of the branches of each bank is presented in table 4.3. Thus, as it can be vividly seen, N is equal to 553. Also, it should be said that, for the 95% level of confidence, z value is n N z ( N 1) z ( ) n 2 2 (553 1)(0.05) 1.96 ( ) Therefore, the total sample size based on the formula for this research is equal to 286. Table 4.3 presents the sample size for each of the four banks (Refah, Maskan, Pasargad, and Saman). In this table, the sample size for each bank is calculated as follows: (The percentages of units in the research population) * (the total sample size) = the sample size for each bank Table 4.3: The sample size for each bank Sampling method Number of the branches of each bank The percentages of units in the research population The total sample size based on the formula The sample size for each bank Refah: * 286 = 76 Stratified Maskan: 184 Pasargad: 159 Saman: * 286 = * 286 = * 286 = 33 Total

64 4.9. Questionnaire Design The main measure instrument used in this study is the questionnaire. In this section, we describe how the questionnaire is designed and developed. Scales for this survey were developed through the literature reviews and the research model. Appendix A exhibits the resources from which the scales were drawn. The questionnaire includes two types of questions. The first type consists of demographic questions about the top managers in each branch and the second type includes the different questions concerning their attitude about the impact of CRM on customer retention in e-banking in the branch they are working in. These questions are classified into questions about technological CRM resources, Infrastructural CRM resources, CRM processes, customer retention programs of CRM, and customer retention. All of the scales in the second type were measured using 7-point Likert scales (strongly agree, agree, somehow agree, neutral, somehow disagree, disagree, and strongly disagree). In the next step, in order to improve the validity and reliability of the data, the first version of the questionnaire has been modified after submitting it to eight academic and bank experts as a pilot test. The final English version of the questionnaire is shown in appendix B and its Farsi version in appendix C Pilot Test We conducted a pilot test in order to evaluate the respondent' comprehension of the questionnaire and estimate the average time to complete it. To do this, first we translated the questionnaire to Farsi. This version of the questionnaire was distributed to eight decision makers in banking industry of Iran from Refah bank, Maskan bank, and Saman bank. From the results, we realized that there were some modifications that should be performed for some questions. In addition, the wording and relevancy of questions were checked and based on that, some other questions were modified, too. 63

65 4.11. Data Analysis In order to make effective decisions in the business world, managers and decision makers must be able to use statistical data analysis techniques and methods (Zikmund, 2000). In this research, data analysis is done in two parts by descriptive and interferential statistics. In the first part, the demography of statistical sample is investigated and the results are shown in frequency tables and diagrams. However, in the inferential part, the data are analyzed using a series of steps which have to be followed. To do so, the first step is doing Kaiser-Meyer-Olkin and Bartlett's test (Momeni, 2006) to measure the adequacy of the sampling. After that, the exploratory factor analysis will be performed. Indeed, when using factor analysis the researcher can gain better results when applying confirmatory factor analysis and structural equation model. This is because the numbers of factors, total variance explained, and communalities of the questions can be gained from factor analysis. Thus, the researcher can delete the questions with little communalities. The final step consists in studying and testing the research hypothesizes using Structural Equation Model (Confirmatory Factor Analysis). To do so, the following steps will be followed: 1. Model expression 2. Model estimation 3. Correction of the model (if needed) 4. Hypothesis test In the structural model we want to understand whether the relationships between the latent traits which are taken from the theory are confirmed by data collected from the sample or not. After defining the latent and evident variables (gained after doing the factor analysis), the research conceptual model will be presented. After that, we will study the accuracy of the measuring model by its indices and correct the model if it is needed (Kalantari, 2009). Studying the significance of the relationships with t-value and studying the correlation are the next steps. To check whether the model is a proper one (accuracy of the measuring model) or not, some indices including chi-square relative to its degree of freedom, root mean square error of approximation (RMSEA), p-value, goodness of fit index (GFI) and also adjusted goodness of fit index (AGFI), comparative fit index (CFI), and normed fit index (NFI) will be studied. To investigate the mentioned issues, the standard model 64

66 will be presented. In the next step, we study the significance of the results in the model. After completing the all mentioned stages in presenting the model, the primary conceptual model is hereby confirmed. Finally, paying attention to the model with significant numbers and standard estimation model, the research hypothesizes will be studied Validity and Reliability In conducting a research the researcher always tries to minimize the measurement error. This error is reduced when the variables accurately and consistently represent and measure the concept. Accuracy is related to the term 'validity' while consistency is referred to 'reliability' (Hair et al., 2007, p. 240) Validity The extent to which a construct measures what it is intended to measure is called 'validity'. No measurement error occurs when the construct has perfect validity (Hair et al., 2007, p. 246). To ensure the validity of this research the approaches mentioned below have been adopted: - To make sure that the measurement scales were adapted appropriately, the questionnaire has been translated into Farsi. - The questionnaire has been reviewed by the supervisors of this work and banks experts to remove and correct the potential problems before sending it to the respondents (content validity). - To check the construct validity of the questionnaire and also to find out if all indicators of each variable (construct) measure what is expected, 'exploratory factor analysis' (calculated in next chapter) has been used. The calculations for this section lead to satisfactory results Reliability According to Hair et al. (2007, p. 241), if the repeated application of a survey instrument results in consistent scores, we can consider it reliable. They also state: "reliability is concerned with the consistency of the research findings". In other 65

67 words, a research can be considered reliable, if its measuring procedure yields the same results on repeated trials (Saunders et al., 2009, p. 156). In this research, Cronbach's alpha has been used to measure the reliability of the items. As it is shown below, the calculated Cronbach's alpha is equal to.970 which is a very good result. The calculations for this section are brought in appendix D. Reliability statistics Cronbach's Alpha N of Items

68 5. Data Analysis 5.1. Introduction In order to analyze the collected data, we should convert them into valuable information by using statistical tests. Indeed, in any research, data analysis constitutes one of the most important parts. It is of great significance in studying the accuracy of hypothesizes. This chapter is drawn up in two parts: descriptive and interferential statistics. In the descriptive part, the demography of statistical sample is tackled and in the inferential part the following issues are studied: 1. Kaiser-Meyer-Olkin and Bartlett's test to measure the adequacy of the sampling to use for exploratory factor analysis (EFA). 2. Exploratory factor analysis for deleting the questions with little factor load and communalities (for more preparation to do the confirmatory factor analysis). 3. Study and test the research hypothesizes using Structural Equation Model (Confirmatory Factor Analysis) 4. Lateral Analyses: - Investigating the correlation between each of the four customer retention programs of CRM and customer retention by using Pearson correlation. - Two-sample t-test with independent samples to study the difference between the respondents' point of view in the two groups of banks (private and governmental banks) - Analysis of variance or comparing the means of some communities to investigate the difference in respondents' point of views in the four groups of banks (for each of the research variables) It should be noted that the software packages SPSS 17 and Lisrel 8.5 have been used to perform the calculations mentioned above. 67

69 5.2. Descriptive Statistics In this section we have studied the variables of respondents' gender, age, educational level, field of study, and also the type of the banks and the belonging to each bank in the research sample. The obtained results have been represented by frequency tables and graphical methods. In addition, a primary comparison of the four banks under study in relation to the research variables is done. Age: Table 5.1 represents the frequency of the respondents' age. As shown, 0.7% of the respondents are between 21 to 30, 13.3% are between 31 to 40, 31.8% are between 41 to 50, and 54.2% are over 51. Figure 5.1 illustrates these results. Table 5.1: The frequency of the respondents' age Frequency Percentage Cumulative Percentage 21 to to Valid 41 to x>= Total Figure 5.1: The distribution of the respondents' age 68

70 Gender: Table 5.2 represents the frequency of respondents' gender. Table 5.2: the frequency of respondents' gender Frequency Percentage Cumulative Percentage MALE As it can be clearly seen, all the respondents are male. Educational level: Table 5.3 represents the distribution of the respondents' educational level. Table 5.3: The distribution of the respondents' educational level EDU Frequency Percentage Cumulative Percentage Diploma A.D Valid BA MSc Total As it can be seen in table 5.3, 8% of the respondents have high school diploma, 20.6% have Associate degree (AD), 66.8% have Bachelor degree, and 4.5% have Master degree (MSc). Figure 5.2 shows the distribution of the educational level of the respondents. 69

71 Figure 5.2: The distribution of the respondents' educational level Field of study Table 5.4 shows the frequency of the respondents' field of study. As we can vividly see, 1.7% of the respondents' fields of study are banking, 4.9% are economics, 37.4% are management, 11.5% are accounting, 3.8% are computer engineering, and 40.6% are other fields such as statistics and electronic engineering. Table 5.4: The frequency of the respondents' field of study Frequency Percentage Cumulative Percentage banking economic management Valid accounting computer other Total

72 Figure 5.3 shows the distribution of the respondents' field of study. Figure 5.3: The distribution of the respondents' field of study Job experience According to table 5.5, %28.7 of the respondents' job experience is under 5 years, 15.7% are between 6 to 10 years, 9.8% are between 11 to 15 years, 14.3% are between 15 to 20 years, and 31.5 are over 21 years. Table 5.5: the frequency related to the respondents' job experience Frequency Percentage Cumulative Percentage X<= TO Valid 11 TO TO X>= Total Figure 5.4 illustrates these findings. 71

73 Figure 5.4: The frequency related to the respondents' job experience Type of the bank Table 5.6 represents the frequency related to type of the banks under study and figure 5.5 illustrates these results. Table 5.6: The frequency related to type of the banks under study Frequency Percentage Cumulative Percentage Gov Valid private Total Figure 5.5: The frequency related to type of the banks under study Having looked at figure 5.5, we can easily perceive that about 59.8% of the respondents are working in governmental banks and about 40.2% are working in private banks. 72

74 The belonging to each bank Table 5.7 represents the frequency related to each bank in the sample. As this table reveals, 26.6% of the respondents are working in bank Refah, 33.2% in bank Maskan, 11.5% in bank Saman, and 28.7% are working in bank Pasargad. Figure 5.6 shows these results. Table 5.7: the frequency related to the belonging to each bank Frequency Percentage Cumulative Percentage refah Maskan Valid Saman Pasargad Total Figure 5.6: The frequency related to the belonging to each bank A primary comparison of the four banks under study in relation to the research variables - Means of the four research variables for bank Refah: The means of the four research variables for bank Refah are represented in table 5.8 (the scores are out of 7) 73

75 Table 5.8: The means of the four research variables for bank Refah Descriptive Statistics Refah N Minimum Maximum Mean Std. Deviation CRM resources CRM processes Retention programs of CRM Customer retention Valid N (listwise) 76 Means of the four research variables for bank Maskan - The means of the four research variables for bank Maskan are represented in table 5.9 (the scores are out of 7). Table 5.9: The means of the four research variables for bank Maskan Descriptive Statistics Maskan N Minimum Maximum Mean Std. Deviation CRM resources CRM processes Retention programs of CRM Customer retention Valid N (listwise) 95 - Means of the four research variables for bank Saman The means of the four research variables for bank Saman are represented in table 5.10 (the scores are out of 7). 74

76 Table 5.10: The means of the four research variables for bank Saman Descriptive Statistics N Minimum Maximum Mean Std. Deviation CRM resources CRM processes Retention programs of CRM Customer retention Valid N (listwise) 33 - Means of the four research variables for bank Pasargad The means of the four research variables for bank Pasargad are represented in table 5.11 (the scores are out of 7). Table 5.11: The means of the four research variables for bank Pasargad Descriptive Statistics N Minimum Maximum Mean Std. Deviation CRM resources CRM processes Retention programs of CRM Customer retention Valid N (listwise) 82 - Means of the four research variables for the two groups of governmental and private banks The means of the four research variables for the two groups of banks (governmental and private) are represented in table 5.12 (the scores are out of 7). 75

77 Table 5.12: The means of the four research variables for governmental and private banks Group Statistics Bank type N Mean Std. Deviation Std. Error Mean CRM resources CRM processes Retention programs of CRM Customer retention gov private gov private gov private gov private Since all of the scales in the questionnaire were measured using 7-point Likert scales and all of the scores in table 5.12 are more than the average score (4), we can conclude that both the governmental and private banks are in favorable condition concerning the four research variables Inferential Statistics: Exploratory Factor Analysis Factor validity is a form of construct validity gained through factor analysis. Factor analysis is a statistical technique with lots of usage in humanities. In fact, it seems that usage of factor analysis in the researches in which test and questionnaire are used is essential (Kalantari, 2009). It should be noted that since number of factors, total variance explained, and communalities of the questions can be gained from factor analysis, in this part of the research, we aim at calculating the communalities and deleting the questions with little communalities. Indeed when using factor analysis; the researcher can gain better results when applying confirmatory factor analysis and structural equation Model The results of Exploratory Factor Analysis for the variable of "technological CRM resources" Table 5.13 represents the result of KMO and Bartlett's test for the data related to the variable of 'technological CRM resources'. 76

78 Table 5.13: KMO and Bartlett's Test Kaiser-Meyer-Olkin Measure of Sampling Adequacy..958 Bartlett's Test of Sphericity Approx. Chi-Square 5.580E3 df 171 Sig..000 Since KMO is greater than 0.7 and Bartlett's test significance number is less than 0.05 (sig<0.05), we can say that the data is proper for doing factor analysis. Table 5.14 shows the communalities of the questions related to 'technological CRM resources'. Table 5.14: Communalities Initial Extraction col col col col col col col col col op op op op op anal anal anal anal anal Extraction Method: Principal Component Analysis. 77

79 Since the numbers of communalities in table 5.14 are greater than 0.05, all the questions related to the variable of 'technology' are proper in the process of factor analysis; therefore, no question is deleted. The total variance explained is presented in Table Table 5.15: Total Variance Explained Component Initial Eigenvalues % of Cumulative Total Variance % Extraction Sums of Squared Loadings % of Cumulative Total Variance % Rotation Sums of Squared Loadings % of Cumulative Total Variance % Extraction Method: Principal Component Analysis. The total variance explained in table 5.15 shows that these questions totally form three factors and these three factors explain and cover about % of the variance of 'technology'. This number indicates the very good power of explanation of research questions for 'technology'. Later, we will deal with factor loads and naming each 78

80 component (factor loads greater than 0.5 are acceptable). Table 5.16 presents the Rotated Component Matrix. Table 5.16: Rotated Component Matrix a Component Analytical Operationa l collaborativ e col col col col col col col col col op op op op op anal anal anal anal anal Extraction Method: Principal Component Analysis. Rotation Method: Varimax with Kaiser Normalization. a. Rotation converged in 6 iterations. 79

81 The results of Exploratory Factor Analysis for the variable of "infrastructural CRM resources" The result of KMO and Bartlett's Test for the data related to the variable of 'infrastructural CRM resources' is shown in table Table 5.17: KMO and Bartlett's Test Kaiser-Meyer-Olkin Measure of Sampling Adequacy..892 Bartlett's Test of Sphericity Approx. Chi-Square 1.321E3 df 55 Sig..000 According to table 5.17, the data is proper for doing the factor analysis. This is because KMO is greater than 0.7 and Bartlett's test significance number is less than 0.05 (sig<0.05). The communalities of the questions for the variable of 'infrastructural CRM resources' are presented in table Table 5.18: Communalities Initial Extraction HR HR HR HR HR org org org org org org Extraction Method: Principal Component Analysis. 81

82 Table 5.18 shows the inadequacy of some questions because of a lack of harmony with the other questions within the process of factor analysis (the first and the fifth questions of human CRM resources and the fifth question of organizational CRM resources). As we can clearly see, the communalities of the mentioned questions are less than Total variance explained is brought in table Table 5.19: Total Variance Explained Initial Eigenvalues Component % of Cumulative Total Variance % Extraction Sums of Squared Loadings % of Cumulative Total Variance % Rotation Sums of Squared Loadings % of Cumulative Total Variance % Extraction Method: Principal Component Analysis. The total variance explained in above table (before deleting the questions with little communalities) shows that these questions totally form two factors which explain and cover about % of the variance of 'infrastructural CRM resources'. After deleting the questions with little communalities (less than 0.5), we will deal with the mentioned issues again. Table 5.20 represents the communalities of the questions after deleting the questions with little communalities. 81

83 Table 5.20 Communalities of the questions after deleting the questions with little communalities Communalities Initial Extraction HR HR HR org org org org org Extraction Method: Principal Component Analysis. Since the numbers of communalities in table 5.20 is more than 0.5, these questions (the remaining questions) are suitable in the process of factor analysis. Table 5.21 represents the total variance explained after deleting the questions with little communalities. Table 5.21: total variance explained after deleting the questions with little communalities Initial Eigenvalues Component % of Cumulative Total Variance % Extraction Sums of Squared Loadings % of Cumulative Total Variance % Rotation Sums of Squared Loadings % of Cumulative Total Variance % Extraction Method: Principal Component Analysis. The total variance explained in table 5.21 (after deleting the questions with little communalities) shows that these questions totally form two factors and these two 82

84 factors explain and cover about 67.5% of the variance of 'Infrastructural CRM resources'. Comparing the two 'total variance explained' tables, it can be easily perceived that after deleting the questions with little communalities, the powers of explanation of the questions have risen. This result would be very useful in confirmatory factor analysis. Later we will deal with factor loads and naming each component. Table 5.22 shows the Rotated Component Matrix. Table 5.22: Rotated Component Matrix Rotated Component Matrix a Component Organizational HR HR HR HR org org org org org Extraction Method: Principal Component Analysis. Rotation Method: Varimax with Kaiser Normalization. a. Rotation converged in 3 iterations The results of Exploratory Factor Analysis for the variable of "CRM processes" Table 5.23 shows the result of KMO and Bartlett's test for the questions related to the variable of 'CRM processes'. 83

85 Table 5.23: KMO and Bartlett's Test Kaiser-Meyer-Olkin Measure of Sampling Adequacy..801 Bartlett's Test of Sphericity Approx. Chi-Square df 45 Sig..000 Since KMO is greater than 0.7 and Bartlett's test significance number is less than 0.05 (sig<0.05), we can say that the data is suitable for doing the factor analysis. The communalities of the questions for the variable of 'CRM processes' are represented in table Table 5.24: Communalities Initial Extraction okp okp okp oip oip mp mp mp mp mp Extraction Method: Principal Component Analysis. The above table shows the inadequacy of the first question among the above questions because of a lack of harmony with other questions within the process of factor analysis. As it can be clearly seen, the communality of this question is less than Table 5.25 represents the total variance explained before deleting the mentioned question. 84

86 Table 5.25: Total variance explained before deleting the question Initial Eigenvalues Component % of Cumulative Total Variance % Extraction Sums of Squared Loadings % of Cumulative Total Variance % Rotation Sums of Squared Loadings % of Cumulative Total Variance % Extraction Method: Principal Component Analysis. The total variance explained in table 5.25 (before deleting the questions with little communalities) shows that these questions totally form three factors and these three factors explain and cover about 66.03% of the variance of 'CRM processes'. After deleting the questions which there communalities are less than 0.5, we will deal with the mentioned issues again. Table 5.26 shows the communalities of the questions after deleting the question with little communalities. 85

87 Table 5.26: Communalities of the questions after deleting the question with little communalities Communalities Initial Extraction okp okp oip oip mp mp mp mp mp Extraction Method: Principal Component Analysis. Since the numbers of communalities of the questions in table 5.26 are acceptable (higher than 0.5), these questions are suitable in the process of factor analysis. Table 5.27 presents the total variance explained after deleting the mentioned question. Table 5.27: Total Variance Explained after deleting the question with little communality Initial Eigenvalues Component % of Cumulative Total Variance % Extraction Sums of Squared Loadings % of Cumulative Total Variance % Rotation Sums of Squared Loadings % of Cumulative Total Variance % Extraction Method: Principal Component Analysis. 86

88 The total variance explained in table 5.27 (after deleting the question with little communality) shows that these questions totally form three factors and these three factors explain and cover about 70.77% of the variance of 'CRM processes'. Comparing the two 'total variance explained' tables, it can be easily perceived that after deleting the question with little communality, the powers of explanations of the questions have been raised. This result would be very useful in confirmatory factor analysis. Later, we will deal with factor loads and naming each component. Table 5.28 shows the component matrix. Table 5.28: Component Matrix a Component Management processes Operational knowledge processes Operational interaction processes okp okp oip oip mp mp mp mp mp Extraction Method: Principal Component Analysis. a. 3 components extracted The results of Exploratory Factor Analysis for the variable of "customer retention programs of CRM" Table 5.29 presents the results of KMO and Bartlett's test for the data related to the variable of 'customer retention programs of CRM'. 87

89 Table 5.29: KMO and Bartlett's test Kaiser-Meyer-Olkin Measure of Sampling Adequacy..896 Bartlett's Test of Sphericity Approx. Chi-Square 3.475E3 df 171 Sig..000 Since KMO is greater than 0.7 and Bartlett's test significance number is less than 0.05 (sig<0.05), it can be said that the data is acceptable to do the factor analysis. The communalities of the questions for the mentioned research variable are presented in table Table 5.30: Communalities Initial Extraction cust cust cust loy loy loy loy cs cs cs cs cs cs cs cs cs com com com Extraction Method: Principal Component Analysis. 88

90 Table 5.30 shows the inadequacy of the five (colored) questions among the above questions because of a lack of harmony with the other questions within the process of factor analysis. As we see, the communalities of these questions are less than Table 5.31 presents the total variance explained before deleting the question. Table 5.31: Total variance explained before deleting the question Initial Eigenvalues Component % of Cumulative Total Variance % Extraction Sums of Squared Loadings % of Cumulative Total Variance % Rotation Sums of Squared Loadings % of Cumulative Total Variance % Extraction Method: Principal Component Analysis. The total variance explained in table 5.31 (before deleting the questions with little communalities) shows that these questions totally form four factors and which explain about % of the variance of 'customer retention programs of CRM'. After 89

91 deleting the questions with little communalities (less than 0.5) we will deal with the mentioned issues again. Communalities of the questions after deleting the questions with little communalities are shown in table Table 5.32: Communalities of the questions after deleting the questions with little communalities Communalities Initial Extraction cust cust loy loy loy loy cs cs cs cs cs com com com Extraction Method: Principal Component Analysis. Since the numbers of communalities of the questions in table 5.32 is higher than 0.5, these questions are acceptable for the process of factor analysis. Table 5.33 presents the total variance explained after deleting the questions with little communalities. 91

92 Table 5.33: Total Variance Explained after deleting the questions with little communalities Initial Eigenvalues Component % of Cumulative Total Variance % Extraction Sums of Squared Loadings % of Cumulative Total Variance % Rotation Sums of Squared Loadings % of Cumulative Total Variance % Extraction Method: Principal Component Analysis. The total variance explained in table 5.33 (after deleting the question with little communality) shows that these questions totally form four factors and these four factors explain and cover about % of the variance of 'customer retention programs of CRM'. Comparing the two 'total variance explained' tables, we can conclude that after deleting the question with little communality, the powers of explanation of the questions have been raised. Later, we will deal with factor loads and naming each component. The results of Rotated Component Matrix are presented in table

93 Table 5.34: Rotated Component Matrix a Component Customization Loyalty programs Community Customer service cust cust loy loy loy loy cs cs cs cs cs com com com Extraction Method: Principal Component Analysis. Rotation Method: Varimax with Kaiser Normalization. a. Rotation converged in 5 iterations The results of Exploratory Factor Analysis for the variable of "customer retention" Table 5.35 presents the results of KMO and Bartlett's test for the data related to the variable of 'customer retention'. Table 5.35: KMO and Bartlett's Test Kaiser-Meyer-Olkin Measure of Sampling Adequacy..675 Bartlett's Test of Sphericity Approx. Chi-Square df 6 Sig

94 Since KMO is between 0.5 and 0.7 and Bartlett's test significance number is less than 0.05 (sig<0.05), it can be said that the data is partially suitable for doing the factor analysis. The communalities of the questions for the mentioned research variable are presented in table Table 5.36: Communalities Communalities Initial Extraction CR CR CR CR Extraction Method: Principal Component Analysis. Table 5.36 shows the inadequacy of the first question among the above questions because of a lack of harmony with other questions within the process of factor analysis. As we can vividly see, the communality of this question is less than Table 5.37 presents the total variance explained before deleting the mentioned question. Table 5.37: Total Variance Explained before deleting the question with little communality Compo nent Initial Eigenvalues Extraction Sums of Squared Loadings Total % of Variance Cumulative % Total % of Variance Cumulative % Extraction Method: Principal Component Analysis. The total variance explained in table 5.37 (before deleting the question with little communality) shows that these questions totally form one factor which explains and covers about % of the variance of 'customer retention'. After deleting the 93

95 question with little communality (less than 0.5), we will deal with the mentioned issues again. Table 5.38 represents the values of communalities of the questions after deleting the question with little communality. Table 5.38: Communalities of the questions after deleting the question with little communality Communalities Initial Extraction CR CR CR Extraction Method: Principal Component Analysis. According to table 5.38, the numbers of communalities of the questions are more than 0.5; therefore, these questions are proper for the process of factor analysis. Table 5.39 presents the total variance explained after deleting the question with little communality. Table 5.39: Total Variance Explained after deleting the question with little communality Compo nent Initial Eigenvalues Extraction Sums of Squared Loadings Total % of Variance Cumulative % Total % of Variance Cumulative % Extraction Method: Principal Component Analysis. The total variance explained in table 5.39 (after deleting the question with little communality) shows that these questions totally form one factor and this factor explains and covers about 67.28% of the variance of 'customer retention'. Table 5.40 presents the component matrix for the questions related to the variable of 'customer retention'. 94

96 Table 5.40: Component Matrix Component Matrix a Component 1 CR2.847 CR3.756 CR4.854 Extraction Method: Principal Component Analysis. a. 1components extracted Structural Equation Model for Research Hypotheses Tests One of the most suitable and the most powerful methods of analysis in the field of social sciences is the multi-variable analysis. This is because the nature of these researches is multi-variable, and cannot be solved by two-variable methods in which an independent variable is considered with one dependent variable in each time. Multi-variable analysis is attributed to a series of analysis methods whose main characteristic is the ability to analyze K independent variables and N dependent variables simultaneously. It should be said that the covariance structures analysis or Structural Equation Model is one of the main methods of analyzing complex data structures Structural Equation Modeling Procedure: The process of covariance structures analysis includes a series of steps which have to be followed. These steps include: 1. Model expression 2. Model estimation 3. Correction of the model (if needed) 4. Hypothesis test 95

97 Figure 5.7 shows the general form of the structural equation model. Measurement component Construct component Measurement component X1 X2 Ksi1 Eta1 Y1 X3 Y2 X4 Ksi2 X5 Eta2 Y3 X6 Ksi3 Y4 X7 Figure 5.7: The general form of structural equation model In this model Y1 to Y4 and X1 to X7 are observed variables and Ksi1 to Ksi3 and Eta1 to Eta2 are Latent variables. Ksi1 to ksi3 are exogenous and Eta1 to Eta2 are endogenous. In the structural model it is tried to make clear whether the relationships between the latent traits which are taken from the theory are confirmed by data gathered from the sample or not. The relationships between the variables in structural equation model are divided into two general areas: 1) the relationships between hidden and evident variables and 2) the relationships between hidden variables and hidden variables. The first group is called measuring model and the second is called structural model (Kalantari, 2009). The symbolization of the presented path gained by Lisrel software is also shown in figure

98 Figure 5.8: Symbolization According to what has been said above, the sign oval indicates the main variables (latent variables) and rectangular signs indicate the research dimensions (evident variables). In this figure, the one-way arrow from the latent variable (oval) to evident variable (rectangular) indicates a correlation and two-way arrow between latent variables shows the relationship between variables which is known as (φ) in the structural equation model (Kalantari, 2009) In the conceptual model of the research there are four latent variables (oval) including: CRM resources, CRM processes, customer retention programs of CRM, and customer retention. The evident variables are the research dimensions which have been gained after doing the factor analysis. The research conceptual model is presented in figure

99 Figure 5.9: The research conceptual model In this section the following steps will be followed: 1. Studying the accuracy of the measuring model by its indices 2. Correction of the model if it is needed (Kalantari, 2009) 3. Studying the significance of the relationships with t-value (if they are not significant, a special number will be determined) 4. Studying the correlation The first basic question which is raised after modeling is whether the model is a proper one (accuracy of the measuring model) or not. To check this, some indices should be used including chi-square relative to its degree of freedom which should be less than 3, root mean square error of approximation (RMSEA) whose value should be less than 0.08, p-value which should be less than 0.05, goodness of fit index (GFI) and also adjusted goodness of fit index (AGFI) which should be more than 0.9. The standard model presented in figure 5.10 investigates the mentioned issues to find the accuracy. 98

100 Figure 5.10: Standard model The mentioned indices indicate the suitability of the measuring model of the related variables. Indeed, chi-square relative to its degree of freedom is 2.79 (less than 3), RMSEA is less than 0.08, and the p-value is less than The other fitness indices of the model are presented in table