The current issue and full text archive of this journal is available at www.emeraldinsight.com/1746-5265.

htm

BJM 4,1

How to avoid customers leaving: the case of the Estonian telecommunication industry
Andres Kuusik and Urmas Varblane
Faculty of Economics and Business Administration, University of Tartu, Tartu, Estonia
Abstract
Purpose The purpose of this paper is to show that the major factors affecting loyalty are dependant on the level of loyalty of customers. Design/methodology/approach A model of relationship between factors of loyalty and loyalty levels of customers was constructed and tested on the empirical data about 1,000 private customers from the Customer Satisfaction Survey of Elion, the biggest telecommunication company in Estonia. Logit model was used in order to examine which factors inuence the probability of the customers remaining on their loyalty level or to moving to another loyalty level. Findings The ndings of the study revealed that it is not accurate to treat all customers equally in terms of methods of increasing their loyalty. The results reveal that four analysed factors affecting customer loyalty (satisfaction, trustworthiness, image and importance of relationship) are playing different roles in the different levels of customer loyalty. The overall satisfaction and importance of a relationship build the foundation of any kind of loyalty. The reliability of products or trustworthiness of the vendor is most critical for behavioural loyalists and the image creation is the main tool for getting committed customers. Research limitations/implications The method for collecting the source data set certain constraints on the adequacy of the model. Practical implications The results presented in this paper could be used by rms developing targeted approaches as part of the aim of increasing customers loyalty. Originality/value This paper introduces an original approach combining factors affecting the customers loyalty with the customers different loyalty levels. Also, the ladder of customer loyalty levels could be used in other research areas. Keywords Customer loyalty, Customer relations, Market segmentation, Consumer behaviour, Telecommunications, Estonia Paper type Research paper

66
Received April 2008 Revised August 2008 Accepted September 2008

Introduction Recent years have shown a growing interest in customer loyalty. Actually all industries suffer from voluntary churn the decision by the customer to switch to another company or service provider. This is particularly true for telecommunication companies. Annual churn rates for telecommunications companies average between 10 and 67 per cent (Hughes, 2007).
Baltic Journal of Management Vol. 4 No. 1, 2009 pp. 66-79 q Emerald Group Publishing Limited 1746-5265 DOI 10.1108/17465260910930458

This study has been prepared with nancial support received from the Estonian Science Foundation (Grant 7405) and from the Estonian Ministry of Education and Research (Target Financing SF0180037s08).

The globalisation of competition, saturation of markets, and development of information technology have enhanced customer awareness and created a situation where long-term success is no longer achieved through optimised product price and qualities. Instead, companies build their success on a long-term customer relationship. According to former studies, it can cost as much as six times more to win a new customer than it does to keep an existing one (Rosenberg and Czepiel, 1984). Depending on the particular industry, it is possible to increase prot by up to 60 per cent after reducing potential migration by 5 per cent (Reichheld, 1993). Hence, we can see that the increase and retention of loyal customers has become a key factor for long-term success of the companies. The main emphasis in marketing has shifted from winning new customers to the retention of existing ones. Traditionally there are two approaches to treating customer loyalty. Some researchers have investigated the nature of different levels of loyalty; others have explored the inuence of individual factors on loyalty. In this paper, both treatments are combined to investigate which specic factors in the telecommunication sector inuence the loyalty rate of the various customers segmented by loyalty. The potential for establishing loyalty depends on the object (i.e. product or vendor), on the subject (customer) or on the environment (market, other suppliers, etc.). This paper focuses on the analysis of object-related factors that are subject to direct impact by companies. The rst section presents an overview of previous research on customer loyalty. The second section is devoted to the construction of the model and description of the research object and used data. The last section presents the results and provides the discussion. The impact of the various factors on loyalty is explored by the example of Elion the leading telecommunication provider in Estonia. The data used for the analysis originate from an Elion Customer Satisfaction Survey, which contains information about 1,000 private customers. LOGIT method was used for testing the hypothesis. Literature overview Segmentation based on customer loyalty There are multiple approaches to customer loyalty. Theories of behavioural loyalty were dominating until 1970, considering loyalty as the function of the share of total purchases (Cunningham, 1956; Farley, 1964), function of buying frequency or buying pattern (Tucker, 1964; Sheth, 1968) or function of buying probability (Harary and Lipstein, 1962; McConnell, 1968; Wernerfelt, 1991). These approaches looked at brand loyalty in terms of outcomes (repeat purchase behaviour) rather than reasons, until Day (2003) introduced the two-dimensional concept of brand loyalty, which stated that loyalty should be evaluated with both behavioural and attitudinal criteria. Contemporary researches consider and accent the psychological (mostly attitudinal and emotional) factor of loyalty (Jacoby and Kyner, 1973; Oliver, 1999; Chaudhuri, 1995; Djupe, 2000; Reichheld, 2003). There are also approaches comparing loyalty with marriage (Hofmeyr and Rice, 2000; Lewitt, 1983; Dwyer et al., 1987). These different approaches allow customers to be distinguished as either behaviourally or emotionally loyal. Behaviourally loyal customers act loyal but have no emotional bond with the brand or the supplier, whereas emotionally loyal customers do. Jones and Sasser call these two kinds of loyalty accordingly false or true long-term loyalty (Jones and Sasser, 1995). Hofmeyr and Rice (2000) divide customers into loyal

How to avoid customers leaving 67

BJM 4,1

68

(behavioural) or committed (emotional). Emotional loyalty is much stronger and longer lasting than behavioural loyalty. It is an enduring desire to maintain a valued relationship. The relationship is so important for the customer that he or she makes maximum efforts to maintain it (Morgan and Hunt, 1995; Reichheld, 2003; Moorman et al., 1992). Highly bonded customers will buy repeatedly from a provider to which they are bonded, recommending that provider to others, and strongly defending these choices to others insisting that they have chosen the best product or service (Butz and Goodstein, 1996). Behaviourally loyal customers could be divided to sub-segments by the reason for so acting: forced to be loyal, loyal due to inertia and functionally loyal. Customers are forced to be loyal when they have to be clients even if they do not want to be clients. Customers may be forced to consume certain products or products/services offered by certain vendor, e.g. when the company acts as a monopoly nholdt, or the poor nancial status of the customer is limiting his selection of goods. Gro Martensen and Kristensen have found that companies with low price strategy had a much higher loyalty than expected from their customer satisfaction. On the other hand, companies that had used a lot of energy on branding indeed had a high customer nholdt et al., satisfaction but they did not have a correspondingly high loyalty (Gro 2000). Also, forced loyalty could be established through creating exit barriers. Loyal behaviour may also result from inertia customer does not move to another vendor due to comfort or the relatively low importance of operation if the choice has low importance, there is no point spending time and effort on searching for alternatives. Thus, based on his faith in the suitability of the current product, the customer continues to use it without checking alternatives. It is in accordance with Olivers approach of cognitive loyalty: the loyalty that is based on brand belief only:
Cognition can be based on prior or vicarious knowledge or on recent experience-based information. If the transaction is routine, so that satisfaction is not processed (e.g. trash pickup, utility provision), the depth of loyalty is no deeper than mere performance (Oliver, 1999).

Hofmeyr and Rice (2000) say that one of the reasons that customers dont switch brands when they are dissatised is that they feel that the alternatives are just as bad as the brand they are using, or even worse. Inertia may be caused also by lack of information about attractive characteristics of the brands (Wernerfelt, 1991). Functionally loyal customers are loyal because they have an objective reason to be. Wernerfelt points out cost-based brand loyalty where brand utilities have a positive inuence on brand choice (Wernerfelt, 1991). Functional loyalty can be created by functional values using price, quality, distribution, usage convenience of a product or through different loyalty programs (points, coupons, games, draws, etc.) giving a concrete reason to prefer a certain supplier. Unfortunately competitors can most easily copy functional values. Thus, creating functional value offers a eeting competitive advantage: functional loyalty cannot be very long lasting (Barnes, 2004). Jones and Sasser (1995) propose three measures of loyalty that could be used in segmentation by loyalty: (1) Customers primary behaviour recency, frequency and amount of purchase. (2) Customers secondary behaviour customer referrals, endorsements and spreading the word.

(3) Customers intent to repurchase is the customer ready to repurchase in the future. Based on the theoretical literature presented above, the customers of a certain telecommunication provider could be segmented by their loyalty as follows (Figure 1): . Committed or emotionally loyal customers active customers who use only the certain providers services, declare that they will use only this provider in the future and recommend this provider to others. . Behaviourally loyal customers active customers who use only the certain providers services, declare that they will use only this provider in the future but do not agree to recommend this provider to others (inert or functionally loyal). . Ambivalent or dubious customers active customers who use only the certain providers services but dont know which provider they will use in the future. . Disloyal reducers customers who have reduced or will reduce the percentage of the providers services in their usage. . Leavers customers who declare, that they will certainly leave this provider. Factors affecting customer loyalty The impact of satisfaction on loyalty has been the most popular subject of studies. Several studies have revealed that there exists a direct connection between satisfaction and loyalty: satised customers become loyal and dissatised customers move to another vendor (Heskett et al., 1994). The primary objective of creating American Customer Satisfaction Index (ACSI) in 1984 was to explain the development of customer loyalty. In the ACSI model, customer satisfaction has three antecedents: (1) perceived quality; (2) perceived value; and (3) customer expectations (Andreson and Fornell, 2000). In the European Customer Satisfaction Index model, perceived quality is divided into two elements: hard ware, which consists of the quality of the product or service attributes, and human ware, which represents the associated customer interactive
Customers

How to avoid customers leaving 69

Loyal Dubious Behaviorally loyal

Disloyal

Committed

Reducers

Leavers

Figure 1. General segmentation of customers by loyalty

BJM 4,1

70

elements in service, i.e. the personal behaviour and atmosphere of the service nholdt et al., 2000). In both models increased satisfaction should environment (Gro increase customer loyalty. In the case of low satisfaction, customers have the option to exit (e.g. going to a competitor) or express their complaints. Researches have shown that 60-80 per cent of customers who defect to a competitor said they were satised or very satised on the survey just prior to their defection (Reichheld et al., 2000). It leads to the conclusion about the existence of other factors beside satisfaction that have a certain impact on customer loyalty (Mascarenhas et al., 2006). Image of brand or supplier is one of the most complex factors. It affects loyalty at least in two ways. First, a customer may use his preferences to present his own image. That may occur both in conscious and subconscious levels. According to Belks theory of extended self, people dene themselves by the possessions they have, manage or create (Belk, 1988). Aaker has shown how consumers prefer brands with personality traits that are congruent with the personality traits that constitute their (malleable) self-schemas (Aaker, 1999). Kim, Han and Park have researched the link between brand personality and loyalty. They did get positive support to the hypothesis that the attractiveness of the brand personality indirectly affects brand loyalty (Kim et al., 2001). Tidwell and Horgan (1992/1993) have shown that people use products to enhance self-image. Second, according to social identity theory, people tend to classify themselves into different social categories. That leads to evaluation of objectives and values in various groups and organisations in comparison with the customers own values and objectives. They prefer partners who share similar objectives and values (Ashforth and Mael, 1989). Fournier (1998) states that consumer-brand relationships are more a matter of perceived goal compatibility. Brands cohere into systems that consumers create not only to aid living but also to give meanings to their lives. Oliver (1999) argues that for fully bonded loyalty the consumable must be part of the consumers self-identity and his or her social-identity. Trustworthiness of the partner is a factor that has a certain impact on the establishment of loyalty nobody expects a long-term relation with a partner that cannot be trusted. Trustworthiness is one criterion for measuring the value of the partner (Doney and Cannon, 1997). Spekman (1988) calls trust a cornerstone of the strategic partnership. Morgan and Hunt (1995) posit that trust is a major determinant of relationship commitment: brand trust leads to brand loyalty because trust creates exchange relationships that are highly valued. Chaudhuri and Holbrook (2001) have shown that brand trust is directly related to both purchase and attitudinal loyalty. Many authors have accented that trust is important in conditions of uncertainty (Moorman et al., 1992; Doney and Cannon, 1997; Dwyer et al., 1987; Morgan and Hunt, 1994). Uncertainty may be caused by dependence or large choice: people tend then to prefer popular or familiar brands or partners. Many denitions describe loyalty as a desire to retain a valuable or important relationship (Morgan and Hunt, 1994; Moorman et al., 1992). That way the establishment of loyalty is predetermined by the importance of a relevant relationship or selection. Weiss (2001) points out three aspects that may increase the importance of the relationship: strategic importance of a product, high risks involved in the transaction or costs incurred by cancellation of contracts. Hofmeyr and Rice point out that the more important the relationship is to a person, the more willing that person is to tolerate dissatisfaction in favour of trying to x it. By contrast, when a relationship

does not matter, then even the perfectly satised consumer can switch on a whim (Hofmeyr and Rice, 2000). A relationship can also be made important by the personal approach. Various authors have compared loyalty with marriage (Lewitt, 1983; Dwyer et al., 1987; Gummesson, 1998; Hofmeyr and Rice, 2000). Marriage is one of the most personal and important relationships. That means that intimacy is one determinant for importance of a relationship. Lewitt (1983) has considered a role of the salesman in making a relationship more personal. Summarising the discussion above, Figure 2 presents the major groups of factors affecting customer loyalty. Other factors not shown in Figure 2 are, for example, price, price activity, distribution, existence or domination of alternatives (Farley, 1964); social class, demographic characteristics and other individual and environmental factors (Kanwar and Pagiavlas, 1992); advertising pressure, constraints on choice (budget limitations, time pressures), usage situation (Lattin, 1987); double jeopardy phenomenon (Ehrenberg et al., 1990), etc. Construction of the model, data description and methods Based on the literature review, the following simplied model of relationship between factors and loyalty levels of customers was constructed by the authors (Figure 3). It combines ve levels in customer loyalty and four specic factors which may inuence the customers loyalty on each level. It allows comparing of the customers on different loyalty levels and examining what kind of factors inuence the probability of the customers remaining on that level or moving to another loyalty level. P(ab) (a 1.4; b 2.5) in Figure 3 indicates the probability that the person will move from loyalty level a to the next level b. Based on the previous discussion, the following hypothesis was proposed: H1. The relevance of factors affecting loyalty depends on the levels of loyalty of customers. The data used for the analysis originated from an Elion Customer Satisfaction Survey carried out in November 2003. Elion is the largest telecommunications and IT provider
Satisfaction with products Expectations Satisfaction Satisfaction with service

How to avoid customers leaving 71

Supplier Trustworthines

Product Salesman

Loyalty

Other factors

Brand personality

Image Goals and values of supplier

Importance of relationship Importance of the product

Termination costs Intimacy of relationship

Figure 2. Major factors of loyalty

BJM 4,1

72

in Estonia. It is owned by AS Eesti Telekom, listed at the stock exchanges of Tallinn and London. Elion provides comprehensive household and business communication services from phone calls to integrated business solutions, from the sale of handsets to the provision of broadband, the installation and maintenance of large IT systems. Elion is the market leader in xed network calls, internet subscriptions and data communication solutions in Estonia and has made a powerful entry into the IT and digital television market. In 2006, the consolidated revenue of Elion Group was 187 million Euros and prot 29.5 million Euros. In 2003, the new brand name Elion was launched. Elion had ca. 360,000 customers and ca. 51,000 broadband internet subscriptions. Elion was providing only telephone and internet services for private customers at that time. There were three groups of competitors for Elion: (1) rms providing telephone services on Elions network; (2) cable-TV companies providing internet services on their own networks; and (3) mobile telephone companies. Therefore, most customers had the following options for using telecommunication services: . phone and internet services only from Elion; . phone from Elion and internet from cable-TV company; . phone from Elion; . phone from an alternative provider (customer pays the monthly fee to Elion and the call charges to the other vendor); . leave and use a mobile phone; or . any combination of the possibilities described above. The satisfaction survey contained information about 1,000 private customers. Data were collected by phone interviews. During the survey the customers were asked how they evaluate satisfaction with and importance of following factors: . service quality; . management of complaints;
Satisfaction Trustworthiness Importance P(34) P(23) P(12) 1. Leavers 2. Reducers 3. Dubious P(45) 4. Loyal 5. Committed

Figure 3. The model combining the loyalty levels of customers with the factors of loyalty. P(ab) is marking the probability of moving customers from the level a to the level b

Image

. . . . .

offering of various products; quality of products; price; accounting accuracy; and image.

How to avoid customers leaving 73

The importance was measured on a ve-point scale, where 1 is not important at all and 5 is very important. Also the satisfaction was measured on a ve-point scale, where 1 is not satised at all and 5 is very satised. Every factor listed above was divided into subcategories (e.g. accounting had subcategories like the accuracy and understandability of bills, dept management, and availability of different payment methods). Satisfaction with the factors was computed as average of the satisfaction with those subcategories weighted by the importance of every subcategory. In addition, customers answered about their general satisfaction with the rm Elion on ve-point scale, where 1 is not satised at all and 5 is very satised. Customer loyalty was measured by following questions: . What operators are the customers using right now? . What operators will the customers use in two years? . Does the customer recommend or is the customer ready to recommend Elion to others? For data processing and hypotheses testing, LOGIT model was used, which contained ten equations each describing the probability that the customer will move from a certain loyalty level to any other certain level of loyalty when the value of any four factors (satisfaction, trustworthiness, importance or image) changes (Appendix). The logic of the equation numbering system is the same as in Figure 3 for example equation L(23) is measuring the probability of moving from second level (reducer) to the third level (dubious) and vice versa. Because of the LOGIT equation is nonlinear in bs, these coefcients do not show exactly how much the probability will change. Therefore, only the direction of change of the probability was analysed. On the ladder of loyalty levels shown in Figure 3, all possible movements (i.e. also the moving from leavers level to the emotionally committed level and vice versa) were handled. Independent variables of the model were formed as follows: . Satisfaction (SATi) is a complex variable, aggregating many different types of satisfaction: overall satisfaction, satisfaction with products, service quality and prices. It was derived as the weighted average of customers opinions about the satisfaction and importance of the mentioned satisfaction types. . Image (IMi) showed how image of Elion ts with the customers values. Brand Elion was relatively young and so there were no surveys about the brand personality. Therefore, suitability of Elions image was measured by the customers opinions about satisfaction with and importance of Elions customer friendliness, domestic orientation and novelty (these are the values that Elion internally wants to own). It was expected by the authors that if the customer is satised, then he/she agrees with these values and feels that Elion operates in

BJM 4,1
.

74
.

harmony with these values. The variable was calculated as the average of above-mentioned three opinions weighted by importance of these values. Trustworthiness (TRSTi ) showed customers satisfaction with the trustworthiness of Elion. It was computed as an average of six opinions about satisfaction with overall credibility, competence of employees, abidance with agreements and contracts, accuracy of bills, safeness and reliability of products, and was weighted by the importance of those elements. Importance (IMPi) had to show the importance of the relationship for customers. Unfortunately there were no questions in the customer satisfaction survey for measuring the importance of relationship and therefore proxy was derived using two indicators: turnover (average sums on the bill) and number of products the customer is actively using. Variable was constructed so that importance is bigger when the customer is using more products and the bills are bigger.

The model contained also some dummy variables marking customers family type (FT1i; FT2i; FT3i) and purchase power (PP1i and PP2i). These variables were included for diminishing variability of ui and thereby to improve the models suitability. Results, discussion and conclusions The results reveal that four analysed factors affecting customer loyalty are playing different roles in the different levels of customer loyalty. Signicant factors affecting the movement of customers to another loyalty level at condence level a 0.05 are shown in Table I. The impact of each factor of loyalty (SAT, satisfaction; TRST, trustworthiness; IM, image and IMP, importance of relationship) is signicant in at least one level of loyalty. In equations L(12)-L(15) and L(45) the satisfaction and importance of a relationship had a signicant inuence on loyalty. Trustworthiness inuences loyalty signicantly in equations L(34) and L(35). The image is signicant only in equation L(45).
Equation L(12)i L(13)i L(14)i L(15)i L(23)i L(24)i L(25)i L(34)i L(35)i L(45)i Factors of loyalty SAT IMP SAT IMP SAT IMP SAT IMP Coefcient 1.71 0.18 1.41 0.33 1.59 0.25 2.57 0.30 Z 4.04 2.04 3.03 2.71 3.35 2.29 4.27 2.4 p 0.00 0.04 0.00 0.01 0.00 0.02 0.00 0.02

Table I. Results of LOGIT analysis

IM TRST TRST SAT IM

2 0.45 0.77 0.92 0.82 0.75

2 1.96 2.31 2.43 2.29 3.36

0.05 0.02 0.02 0.02 0.00

Note: Includes only signicant results on condence level a 0.05

While equations in the model (1) describe the probability to move from one loyalty segment to another, it is possible to interpret regression coefcients both ways. The positive coefcient for example in equation L(12) shows that growth of a particular factor score will increase the probability that the leaver will move to the disloyal segment; but it also shows that growth of a particular factor score will decrease the probability that the disloyal customer moves to the leavers segment. According to the results (Table I) the IMP has a major impact on making the decision to terminate a contract with the current vendor if the relationship is considered important, the chances of moving to another vendor are less probable. In this model the importance of relationship depended primarily on the number of products used by the customer. Hence, Elion has to think of providing, free of charge, additional and comfort services in order to increase the loyalty of customers who potentially could leave. According to the data, the behavioural loyalists valued the TRST of Elion. In case of reduction of the trustworthiness of the vendor, the probability that the customer starts to consider alternatives, i.e. becomes doubtful, will increase. In order to turn a behaviourally loyal customer into an emotional loyalist, one has to ensure maximum satisfaction of the customer and it is also important to match the image and values. The results reveal that loyalty factors included into the model do not affect the probability of disloyal customers increasing their loyalty or inuencing loyal or dubious customers shift to the disloyal segment. Maybe the disloyal segment is not correctly dened; but it could also be the case that the border line between doubtful and disloyal customers is obscure. It is possible that doubtful customers have developed a predisposition to disloyalty and their actual realisation of disloyalty depends on factors that were not covered by the model established in this paper. In equation L(34) the image had a negative coefcient which seems illogical. The authors opinion is that there is no problem in the model or dissonance in the hypothesis. The explanation could be hidden partly in the scale used in the questionnaire and partly in the nature of the behavioural loyalist segment. The ve-point scale was used in the questionnaire, where 1 not satised at all was a negative opinion about image and 5 very satised was a positive opinion. In the middle of the scale was 3, marking a neutral attitude toward the image of Elion. It is very probable that the customers who do not pay attention to the vendor and its image, will answer neutrally they have neither positive nor negative feelings about Elions image. They are probably loyal due to inertia. As said before, customers who are loyal due to inertia start to think about alternatives when something does not work, i.e. trustworthiness of the vendor decreases. Until this point they do not think about the image of their current vendor. They need to evaluate it when they start to compare the current vendor with others, or they are moving to the committed customers segment. Figure 4 summarises the ndings of the study and shows that overall satisfaction and importance of products (or relationship) build the foundation of any kind of loyalty. It shows also that reliability of products or trustworthiness of the vendor is most critical for behavioural loyalists and the image creation is the main tool for getting committed customers. In general, the ndings revealed that it is not accurate to treat all customers equally in terms of methods of increasing their loyalty. The research supported the hypothesis that the list of most important factors affecting loyalty is dependant on the customers loyalty levels.

How to avoid customers leaving 75

BJM 4,1
???

Image Trustworthiness P(34) P(23) P(12) 1. Leavers 2. Reducers 3. Dubious P(45) 4. Loyal 5. Committed

76

Importance Satisfaction

Figure 4. Factors having statistically signicant inuence on the customers loyalty on particular loyalty levels

Importance Satisfaction

Results could be integrated into the customer relationship management strategies (CRM) of rms. Many telecommunication rms are using the CRM software for churn analysis and could predict the potential leavers. In order to keep potential leavers, rms should provide additional services with the aim of increasing the importance of the relationship for customers. With the help of usage history it is possible to nd behaviourally loyal customers and through the imago-campaigns increase their emotional commitment. The current research has also certain limitations. The method for collecting the source data set constraints on the adequacy of the model. Although the quality of the source data was good (high number and trustworthiness of respondents), the questionnaire was drawn for measuring customer satisfaction, i.e. for purposes other than required by this given model. Therefore, some factors were derived indirectly and with limitations, and they may lack some qualities considered in the theoretical part. There are multiple ways for further development of the model, but rst there is a need to elaborate reliable questionnaires for colleting the source data.
References Aaker, J.L. (1999), The malleable self: the role of self-expression in persuasion, Journal of Marketing Research, Vol. 36 No. 1, pp. 45-58. Andreson, E.W. and Fornell, C. (2000), Foundations of the American Customer Satisfaction Index, Total Quality Management, Vol. 11 No. 7, pp. 869-82. Ashforth, B.E. and Mael, F. (1989), Social identity theory and the organization, Academy of Management Review, Vol. 14 No. 1, pp. 20-39. Barnes, J. (2004), Loyalty: a multidimensional concept, available at: www.CRMguru.com (accessed 2004). Belk, R.W. (1988), Possessions and the extended self, Journal of Consumer Research, Vol. 15, pp. 139-68. Butz, H.E. Jr and Goodstein, L.D. (1996), Measuring customer value: gaining a strategic advantage, Organizational Dynamics, Vol. 24 No. 3, pp. 63-77.

Chaudhuri, A. (1995), Brand equity or double jeopardy, Journal of Product & Brand Management, Vol. 4 No. 1, pp. 26-32. Chaudhuri, A. and Holbrook, M.B. (2001), The chain of effects from brand trust and brand affect to brand performance: the role of brand loyalty, Journal of Marketing, Vol. 65 No. 2, pp. 81-94. Cunningham, R.M. (1956), Brand loyalty what, where, how much?, Harward Business Review, Vol. 34, pp. 116-28. Day, G.S. (2003), Creating a superior customer-relating capability, MIT Sloan Management Preview, Vol. 44 No. 3, pp. 77-83. Djupe, P.A. (2000), Religious brand loyalty and political loyalties, Journal for the Scientic Study of Religion, Vol. 39 No. 1, pp. 78-90. Doney, P.M. and Cannon, J.P. (1997), An examination of the nature of trust in buyer-seller relationships, Journal of Marketing, Vol. 61, pp. 35-51. Dwyer, F.R., Schurr, P.H. and Oh, S. (1987), Developing buyer-seller relationships, Journal of Marketing, Vol. 51, pp. 11-27. Ehrenberg, A.S.C., Goodhardt, G.J. and Barwise, T.P. (1990), Double jeopardy revisited, Journal of marketing, Vol. 54, pp. 82-91. Farley, JohnU. (1964), Why does brand loyalty vary over products?, Journal of Marketing Research, November, pp. 9-14. Fournier, S. (1998), Consumers and their brands: developing relationship theory in consumer research, Journal of Consumer Research, Vol. 24, pp. 343-73. nholdt, L., Martensen, A. and Kristensen, K. (2000), The relationship between customer Gro satisfaction and loyalty: cross-industry differences, Total Quality Management, Vol. 11 Nos 4/5&6, pp. 509-14. Gummesson, E. (1998), Productivity, quality and relationship marketing in service operations, International Journal of Contemporary Hospitality Management, Vol. 10 No. 1, pp. 4-15. Harary, F. and Lipstein, B. (1962), The dynamics of brand loyalty: a Markovian approach, Operations Research, Vol. 10 No. 1, pp. 19-40. Heskett, J.L., Jones, T.O., Lovemann, G.W., Sasser, W.E. Jr and Schlesinger, L.A. (1994), Putting the service-prot chain to work, Harward Business Review, Vol. 72 No. 2, pp. 164-74. Hofmeyr, J. and Rice, B. (2000), Commitment-led Marketing, Wiley, Chichester, p. 299. Hughes, A.M. (2007), Churn reduction in the telecom industry, available at: www.telecommarketing.com/index.php?q node/29 (accessed 1 March 2008). Jacoby, J. and Kyner, D.B. (1973), Brand loyalty vs. repeat purchasing behavior, Journal of Marketing Research, Vol. 10, pp. 1-9. Jones, T.O. and Sasser, W.E. Jr (1995), Why satised costumers defect, Harward Busines Review, Vol. 73 No. 6, pp. 88-99. Kanwar, N. and Pagiavlas, N. (1992), When are higher social class consumers more and less brand loyal than lower social class consumers?: the role of mediating variables, Advances in Consumer Research, Vol. 19, pp. 589-95. Kim, C.K., Han, D. and Park, S. (2001), The effect of brand personality and brand identication on brand loyalty: applying the theory of social identication, Japanese Psychological Research, Vol. 43 No. 4, pp. 195-206. Lattin, J.M. (1987), A model of balances choice behavior, Marketing Science, Vol. 6 No. 1, pp. 48-65. Lewitt, T. (1983), After the sale is over . . ., Harvard Business Review, September/October, pp. 87-93.

How to avoid customers leaving 77

BJM 4,1

78

McConnell, D.J. (1968), The development of brand loyalty: an experimental study, Journal of Marketing Research, Vol. 5, pp. 13-19. Mascarenhas, O.A., Kesavan, R. and Bernacchi, M. (2006), Lasting customer loyalty: a total customer experience approach, Journal of Consumer Marketing, Vol. 23 No. 7, pp. 397-405. Moorman, C., Zaltman, G. and Deshpande, R. (1992), Relationships between providers and users of market research: the dynamics of trust within and between organizations, Journal of Marketing Research, Vol. 29, pp. 314-28. Morgan, R.M. and Hunt, S.D. (1994), The commitment-trust theory of relationship marketing, Journal of Marketing, Vol. 58 No. 3, pp. 20-3. Morgan, R.M. and Hunt, S.D. (1995), The commitment-trust theory of relationship marketing, Journal of Marketing, Vol. 58 No. 3, pp. 20-39. Oliver, R.L. (1999), Whence consumer loyalty?, Journal of Marketing, Vol. 63, pp. 33-44. Reichheld, F.F. (1993), Loyalty-based management, Harvard Business Review, Vol. 71 No. 2, pp. 64-72. Reichheld, F.F. (2003), The one number you need to grow, Harward Business Review, Vol. 81 No. 12, pp. 46-55. Reichheld, F.F., Markey, R.G. and Hopton, C. (2000), The loyalty effect the relationship between loyalty and prots, European Business Journal, pp. 134-9. Rosenberg, L. and Czepiel, J.A. (1984), A marketing approach to customer retention, Journal of Consumer Marketing, Vol. 1, pp. 45-51. Sheth, J.N. (1968), A factor analytical model of brand loyalty, Journal of Marketing Research, Vol. 5, pp. 395-404. Spekman, R.E. (1988), Strategic supplier selection: understanding long-term buyer relationships, Business Horizons, July-August, pp. 75-81. Tidwell, P.M. and Horgan, D.D. (1992/1993), Brand character as a function of brand loyalty, Current Psychology, Vol. 11 No. 4, pp. 346-53. Tucker, W.T. (1964), The development of brand loyalty, Journal of Marketing Research, August, pp. 32-5. Weiss, A. (2001), When do B2B customers want a close relationship?, available at: www. MarketingProfs.com/tutorials (accessed 2004). Wernerfelt, B. (1991), Brand loyalty and market equilibrium, Marketing Science, Vol. 10 No. 3, pp. 229-45. Appendix. A LOGIT model used for testing hypothesis

L12i 1=1 e2Z L13i 1=1 e2Z L15i 1=1 e2Z L23i 1=1 e2Z L25i 1=1 e2Z L34i 1=1 e2Z L45i 1=1 e2Z

L14i 1=1 e2Z L24i 1=1 e2Z 1 L35i 1=1 e2Z

where Z b1 b2 SATi b3 IMi b4 TRSTi b5 IMPi b6 FT1i b7 FT2i b8 FT3i b9 PP1i b10 PP2 ui and:

. . . . . . . . . . . . . . . . . . . . .

L(12)i 0, if customer is leaver and 1, if customer is reducer. L(13)i 0, if customer is leaver and 1, if customer is dubious. L(14)i 0, if customer is leaver and 1, if customer is slightly loyal. L(15)i 0, if customer is leaver and 1, if customer is emotionally committed. L(23)i 0, if customer is reducer and 1, if customer is dubious. L(24)i 0, if customer is reducer and 1, if customer is slightly loyal. L(25)i 0, if customer is reducer and 1, if customer is emotionally committed. L(34)i 0, if customer is dubious and 1, if customer is slightly loyal. L(35)i 0, if customer is dubious and 1, if customer is emotionally committed. L(45)i 0, if customer is slightly loyal and 1, if customer is emotionally committed. SATi satisfaction. IMi image. TRSTi trustworthiness. IMPi importance of relation. FT1i 1, if it is a family with children and 0, if any other family type. FT2i 1, if it is a small family without children and 0, if any other family type. FT3i 1, If it is seniors family and 0, if any other family type. FT1i FT2i FT3i 0, it is not standard family for Elion. PP1i 1, if it is a family with medium income and 0, in any other. PP2i 1, if it is a family with high income and 0, in any other. PP1i PP2i 0, it is a family with low income.

How to avoid customers leaving 79

About the authors Andres Kuusik is currently working as Lecturer of Marketing in the University of Tartu (Estonia). His main research topics include CRM customer behaviour, loyalty and branding. His latest research is concerned with customer loyalty problems in the telecommunication sector. Urmas Varblane is working as Professor of International Business and Innovation and vice-dean of research at the Faculty of Economics and Business Administration in Tartu University (Estonia). His main research topics include innovation process and policy, knowledge and technology transfer and impact of foreign direct investments on host country. Urmas Varblane is the corresponding author and can be contacted at: varblane@mtk.ut.ee

To purchase reprints of this article please e-mail: reprints@emeraldinsight.com Or visit our web site for further details: www.emeraldinsight.com/reprints