A cost, benefit approach to understanding service loyalty

CunninghamProfessor of Marketing, College of Business and Administration, University of Colorado at Denver, Denver, Colorado, USA Keywords Service quality, Customer loyalty, Marketing st

A cost/benefit approach to understanding service loyalty

Moonkyu LeeAssociate Professor of Marketing, College of Business and Economics, Yonsei University, Seoul, Korea

Lawrence F CunninghamProfessor of Marketing, College of Business and Administration, University of Colorado at Denver, Denver, Colorado, USA

Keywords Service quality, Customer loyalty, Marketing strategy, Cost/benefit analysis Abstract Examines determinants of service loyalty under the assumption that consumers perform a cost/benefit analysis when deciding whether or not they want to be ``regular customers'' It develops potential determinants of service loyalty based on the service quality, transaction cost, andswitching cost literature, andestimates their relative influences with survey data from customers currently using banks and travel agencies The results indicate that, in addition to service quality perceptions, transaction/switching cost factors have a significant impact on service loyalty Implications of the results are discussed.

IntroductionCustomer loyalty is one of the major sources of sustainable competitiveadvantage for service firms (Bharadwaj et al., 1993) Consistently highlevels of customer loyalty can not only create tremendous competitiveadvantage, but also boost employee morale and productivity On the otherhand, persistent customer defection has a devastating impact on a servicecompany's performance Reichheld and Sasser (1990), for instance,demonstrated that a 5 percent decrease in customer attrition could translateinto a 25 to 85 percent increase in profits, depending on the service industry.Therefore, developing and maintaining customer loyalty, or creating long-term relationship with customers is the key to the survival and growth ofservice firms (e.g Duffy, 1998; Griffin, 1995; Kandampully, 1998;

Reichheld, 1996)

This paper identifies potential determinants of customer loyalty to serviceproviders (called ``service loyalty'' hereafter) and examines their relativeinfluences While a considerable amount of literature has been devoted to thecustomer loyalty issues in the product domain (e.g Jacoby and Chestnut,1978; LaBarbera and Mazursky, 1983; Mazursky et al., 1987; Newman andWerbel, 1973), little empirical research has been conducted to deal withthese issues in the services area to date The present study takes a cost/benefit approach to understanding service loyalty The basic assumptionunderlying the approach is that customers' intention to repatronize theircurrent service provider is determined by the comparison analysis betweencosts and benefits associated with the customer/service provider relationship

The research register for this journal is available at http://www.mcbup.com/research_registers The current issue and full text archive of this journal is available at http://www.emerald-library.com/ft

The authors wish to express their appreciation for the support of the Graduate School

of Business and Administration and the Institute for International Business at theUniversity of Colorado at Denver

Sustainable competitiveadvantage

Potential determinants ofcustomer loyalty

An executive summary formanagers and executivereaders can be found at theend of this article

Although the measurement and management issues of the benefit component

in this assumption have been extensively studied over the past years under therubric of service quality (Bitner, 1990; Bolton and Drew, 1991a, 1991b;Cronin and Taylor, 1992; Brown and Swartz, 1989; Parasuraman et al., 1985,1988; 1991; also see Fisk et al., 1993 for a review), those of the costcomponent have been relatively overlooked This study conceptualizes severaltypes of costs related to service loyalty based on existing literature includingthe transaction cost analysis (Williamson, 1975, 1979, 1981, 1985) and theswitching cost approach (Heide and John, 1988; Porter, 1980, 1985; Weiss andAnderson, 1992) It is believed that by examining the relative importance ofthese cost/benefit variables in determining service loyalty, this research wouldcontribute not only to better understanding how service loyalty is formed, butalso to effectively managing it from a marketer's standpoint

Research hypotheses

In this section, the major constructs of the study including service loyalty andvarious cost/benefit factors are discussed In addition, specific hypothesesare developed based on the existing literature

Service loyalty (LOYALTY)Service loyalty is the dependent variable of this study Czepiel and Gilmore(1987) define service loyalty as a specific attitude to continue in an exchangerelationship based on past experiences Their definition implies that levels ofservice loyalty can be assessed by attitudinal measures such as the ones based

on intentions to repatronize a service provider Such attitudinal measures have

an advantage over behavioral measures (e.g repeat patronage) in that theycan provide greater understanding of the factors associated with thedevelopment and modification of loyalty (Oliva et al., 1992); (also see Dickand Basu (1994) and Jacoby and Chestnut (1978) for the comparison betweenbehavioral and attitudinal measures of customer loyalty)

While Czepiel and Gilmore (1987) propose that service loyalty is formedbased solely on past experience, research evidence shows that consumerstend to adopt loyalty as a tool to reduce risk and uncertainty when finding anew service provider (Guseman, 1981) In other words, future expectationscan be another important basis for the loyalty formation Thus, serviceloyalty in this study is defined as:

customers' intention to repatronize their current service provider (or company)based on past experiences and future expectations

Antecedents of service loyalty are identified below based on the literature

As mentioned earlier, there are two sets of factors, one concerned withservice quality as the benefit component of the model and another related tothe present and future costs of services, which are further categorized intoeconomic, transaction, and switching costs

The benefit component: overall service quality (QUALITY)One of the critical determinants of service loyalty would be the overallquality of services as perceived by customers Perceived service quality hasbeen conceptualized and measured as a form of attitude (Bolton and Drew,1991a; Parasuraman et al., 1988) Recently, researchers have demonstratedthat service quality is determined by customer satisfaction/dissatisfactionwith service experiences (Bitner, 1990; Bolton and Drew, 1991a, 1991b) andthat service quality, in turn, affects service loyalty (Bitner, 1990) Therefore,

H1: The better the perceived quality of services, the higher customers'intention to repatronize their service provider.

An important issue in testing this hypothesis is how to measure perceivedservice quality Parasuraman et al (1985, 1988, 1991) have developed aservice quality measure, SERVQUAL, on the basis of their gap theory,which states that customers' assessment of overall service quality isdetermined by the degree and direction of the gap between their expectationsand perceptions of actual performance levels They have also identified fivedimensions underlying overall service quality, namely tangibles, reliability,responsiveness, assurance, and empathy In addition, they have proposed thatperceived service quality could be estimated by calculating the differencebetween expectations and perceptions of actual service performance Thismeasurement paradigm is adopted in the present study In the followingsections, the cost component of the model is dimensionalized and discussed.Economic costs

Economic costs are basically what consumers have to give up or sacrifice toobtain a product or a service Two types of economic costs are

conceptualized, monetary and nonmonetary costs

Service cost (SERVCOST) The effect of monetary cost or price on consumerbehavior has been extensively dealt with in the product domain Theliterature, applied in the service context, suggests that the cost incombination with the benefit of using a service determines overall servicevalue, which influences customers' purchase intention and behavior (Doddsand Monroe, 1985; Dodds et al., 1991; Monroe, 1990; Zeithaml, 1988).Since the cost has a negative impact on customers' budgets, it would have anegative impact on their intention to patronize or repatronize the serviceprovider, ceteris paribus Therefore,

H2: The higher the perceived service cost, the lower customers' intention

to repatronize their service provider

An important assumption underlying this hypothesis is that it is not actual orobjective, but rather, it is perceived cost that would affect consumerjudgment and behavior (Jacoby and Olson, 1977; Monroe, 1990; Zeithaml,1988) This assumption underlies all the cost-related hypotheses in the paper.Service time (SERVTIME) The concept of economic costs can be extended

to include nonmonetary costs such as service time, i.e the amount of timeduring which a service is provided (see Murphy and Enis, 1986 for a review

of time prices) Service time would affect repatronage intention in the sameway as service cost would Many customers would like to have fasterservices, although there may be some individual or situational differences inthe value of time Fast-food services and express check-out services ingrocery stores are a few of the many examples of firms' recognition thatfaster services are important to large segments in their market Thus,H3: The longer the service time, the lower customers' intention torepatronize their service provider

Transaction costsAnother type of nonmonetary costs is considered based on the transactioncost analysis (Williamson, 1975, 1979, 1981, 1985) Williamson postulatesthat transaction difficulties or ``transaction costs'' are present in exchangeprocesses as a consequence of the interaction among various factors Heposits that the organizational structures and boundaries are determined by the

How to measure perceivedservice quality

Recognition that fasterservices are important tolarge market segments

efforts to minimize transaction costs The transaction cost analysis has beenwidely employed to understand many marketing phenomena such asdistribution channel relationships (e.g Anderson and Coughlan, 1987;Dwyer and Oh, 1988; Heide and John, 1988; Klein et al., 1990) andcustomer/service provider relationships (e.g Bowen and Jones, 1986; Oliva

et al., 1992) To the service context, the following factors have beenparticularly pertinent

Difficulty of assessing service performance (DIFFICULT) From acustomer's standpoint in the exchange process, one major source oftransaction costs is the difficulty of evaluating service performance (Alchianand Demsetz, 1972; Bowen and Jones, 1986; Williamson, 1975, 1981) Suchdifficulty is prevalent with services primarily because they are generallyintangible Zeithaml (1981), based on the classification system developed byDarby and Karni (1973), proposes that most services are high in experiencequalities or credence qualities since their quality can hardly be assessed until

or even after they are consumed It should be noted here that the intangibility

of services is a matter of degree (McDougall and Snetsinger, 1990) Thus,varying degrees of intangibility would give rise to differences in thetransaction costs

The difficulty of assessing a service provider's performance puts a burden oncustomers because they would have to closely monitor the services Even ifthey would do so, they would not often be sure if they were paying a fairprice (Bowen and Jones, 1986) Therefore, such transaction difficulty wouldeventually have a negative impact on service loyalty This leads to thefollowing hypothesis:

H4: The greater the difficulty of assessing the performance of the currentservice provider, the lower consumers' intention to repatronize theirservice provider

Service provider's specific knowledge about customers (SPECIFIC).Another factor related to transaction costs is a service provider's specificknowledge about individual customers' idiosyncrasies and their needs andwants (Williamson, 1979, 1985) This concept is different from theambiguity of service performance described above in that it reducestransaction difficulty and works positively on service loyalty Services aretypically produced and consumed in the same time period, involving a highlevel of client/provider interaction (Lovelock, 1981; Shostack, 1977;Zeithaml et al., 1985) Thus, the service provider's intimate understanding ofcustomers' tastes and preferences would not only speed up the transactionprocess, but also bring about customer satisfaction and loyalty throughservice customization Therefore:

H5: As the service provider's knowledge about customers' needs andwants increases, customer loyalty increases

Switching costsSwitching costs have been examined in the context of micro-economics (e.g.Ferrell and Shapiro, 1988; Klemperer, 1987a, 1987b), interfirm relationships(e.g Caves and Porter, 1977; Porter, 1980, 1985), and distribution channelrelationships (e.g Heide and John, 1988; Weiss and Anderson, 1992).Although they have been dealt with in different areas, the basic idea that hasinspired the research is essentially similar That is, once a transactionrelationship is established, one party becomes more dependent on the other

as the cost of switching transaction partners gets higher Applied in the

Evaluating serviceperformance

Specific knowledge aboutindividual customers'idiosyncrasies

current research context, this means that customers often become ``lockedinto'' their current service provider due to high switching costs.

Switching costs are the costs it is anticipated will be incurred in the future,whereas economic and transaction costs are those incurred in the presenttransaction As mentioned earlier, service loyalty is formed based on futureanticipations as well as past experiences Thus, switching costs should beexplored here as an important determinant of customer loyalty The cost ofswitching service providers should depend on the levels of the followingfactors

Information search cost (INFOCOST) When customers consider switching

to a new service provider, one of the first difficulties they face involvesgathering information about substitutes (Porter, 1985) Such difficulties aretermed ``information search cost.'' Although the cost can be high or lowdepending on the service, it would affect the overall level of switching costs.Specifically, the higher the information search cost, the higher the overallswitching costs, and thus, the less likely customers are to switch to a newprovider This leads to:

H6: As the cost of searching for information about a substitute increases,service loyalty increases

Perceivedrisk (RISK) Another factor related to service provider switching iscustomers' perceptions of risk in selecting a new provider Perceived risk hasbeen extensively dealt with in the literature since it accompanies allpurchases to some extent and affects various aspects of purchasing behavior(e.g Bettman, 1973; Cox, 1967) It has been suggested that consumers try tohandle the risks by deliberately searching for prepurchase information(Locander and Hermann, 1979; Lutz and Reilly, 1973; Murray, 1991),resulting in increased brand/store loyalty once a good choice is made(Cunningham, 1967; Guseman, 1981; Roselius, 1971; Sheth and Venkatesan,1968) Thus, applied in the service context, this implies that perceived riskfunctions as an important deterrent to service switching Hence,

H7: As perceived risk increases in selecting a new service provider, serviceloyalty increases

Substitutability of the service provider (SUBSTITUTE) Substitutability isdefined in this study as the extent to which the service activities the currentservice provider performs are available from alternative sources (Bagozziand Phillips, 1982) The literature in interorganizational relationshipssuggests that there is an inverse relationship between the substitutability anddependence of one organization on another (e.g Bagozzi and Phillips, 1982;Heide and John, 1988; Pfeffer and Salancik, 1978) Although substitutabilitymight vary across service industries, it is expected to lower switchingbarriers, and thus decrease service loyalty Therefore,

H8: The higher the substitutability of the current service provider, thelower consumers' intention to repatronize the provider

Geographical proximity to the service provider (PROXIMITY) Sinceservices are generally consumed in the same place where they are produced,

a service provider's location is a critical factor For most services, consumersprefer convenient locations When they move, they tend to switch to a newservice provider in their areas Thus, inconvenient location would encourageswitching tendency and discourage service loyalty This leads to:

Switching costs

Perceptions of risk inselecting a new provider

Inverse relationship

H9: The less convenient the location of the current service provider, thelower consumers' intention to repatronize the provider (or the highercustomers' intention to switch to a new provider).

Figure 1 shows the potential determinants of service loyalty discussed thus far.Research method

Selection of servicesTwo service categories, banks and travel agencies, were selected for thestudy based on the expectation that the characteristics of these serviceswould cover a wide range of variations on the taxonomic dimensionsproposed by Lovelock (1983) and Bowen (1990) Specifically, for manyconsumers, banking services are directed at intangible assets, delivered on acontinuing basis, and more importantly, involve formal ``membership''relations, thereby reducing customers' abilities to switch On the other hand,the services provided by travel agencies generally produce physical products(i.e airline tickets), involve discrete transactions with no formal

membership, and thus impose little restriction to service switching

Measure development

By the procedure suggested by Churchill (1979) and Nunnally (1978),multiple items were developed for the major constructs in the study on thebasis of relevant literature However, some straightforward constructs weremeasured by single items For instance, information search cost wasmeasured by the extent to which respondents agreed or disagreed with thestatement, ``It would be extremely costly to search for information aboutgood banks [travel agencies].''

Perceived service quality and risk were measured in the following way First,past studies indicate that overall service quality is composed of fivedimensions, which can be assessed by 22 service performance items pairedwith as many expectation items (Parasuraman et al., 1988, 1991) However,including all these items in this study would have made the survey task tooonerous for respondents, considering the number of questions used formeasuring the other constructs Thus, in addition to the overall servicequality measure, five service performance items paired with expectationitems were included to capture the corresponding dimensions extracted from

Figure 1 Potential determinants of service loyalty

Taxonomic dimensions

Past studies

factor analyses in previous studies (i.e tangibles, reliability, responsiveness,assurance, and empathy) Second, past literature also suggests that perceivedrisk is a multi-dimensional construct (Kaplan et al., 1974; Roselius, 1971).Therefore, in this study, overall perceived risk was measured as well asperceptions of five risk dimensions:

(1) financial,(2) performance,(3) physical,(4) psychological, and(5) social risks (Murray and Schlacter, 1990)

The purpose of including such measures of individual dimensions was todetermine their relative impact on the overall service quality and riskperceptions, in case these overall perceptions would prove to be significant

in determining service loyalty

QuestionnaireTwo sets of questionnaires were prepared, one for banks and the other fortravel agencies The constructs and their corresponding items covered by thequestionnaire were essentially the same Most items were measured on nine-point Likert scales ranging from ±4 = ``strongly disagree'' to +4 = ``stronglyagree.'' The face validity of these items was then established through apretest with several marketing professionals Some items were revised todeliver clear meanings; several others were deleted due to unclearinterpretations

Data collectionData were collected through personal interviews with people working inoffices in a large metropolitan area Interviewers were graduate students whowere trained about the survey procedure They visited the offices and askedpeople whether they were currently using either banking or travel agencyservices They contacted only those who were using either one of theservices They presented them with the purpose of the study and thequestionnaire titled ``Customer evaluations of service quality'' Eachrespondent filled out either the bank or travel agency questionnaire At thebeginning of each questionnaire, respondents were asked to consider theservice provider (bank or travel agency) that they were using If they usedmultiple providers, they were asked to consider the one they used most often.They then filled in measures of service quality, transaction and switchingcosts, and service loyalty They also provided their demographic andbackground information A total of 165 usable questionnaires were collected,which consisted of 84 for banks and 81 for travel agencies

Measure validationThe multiple items measuring single constructs were factor-analyzed todetermine unidimensionality (Churchill, 1979) As expected, all themeasures showed unidimensionality The items were then submitted toreliability analyses via Cronbach alpha The reliability values of all theconstructs were either close to or greater than 0.7, the threshold Nunnally(1978) recommended for basic research The items and their Cronbachalphas are shown in the Appendix The correlation matrix, shown in Table I,indicates that few of the constructs are correlated at high levels, suggestingsome evidence of discriminant validity

Face validity

Unidimensionality

ResultsSample characteristicsTable II shows the demographic profile of the respondents In addition, itindicates the respondents' usage, familiarity, and involvement differencesbetween banks and travel agencies, e.g the respondents used banks morefrequently and knew better about banks compared to travel agencies.

Variable (N = 165)Overall (n = 84)Bank Travel agency(n = 81) Sex

Notes: a standard deviation; b measured on a nine-point bipolar scale ranging from ±4

to +4, where +4 = very familiar with the service; c simple average ratings on three nine-point bipolar scales ranging from ±4 to +4, where +4 = very important, very useful, and very relevant to me and my life

Table II Sample characteristics

Regression analysisMultiple regression analysis was performed to determine differential effects

of the service quality, transaction cost, and switching cost variables on serviceloyalty The results are summarized in Table III As can be seen, the overallregression model was highly significant (F = 8.55, p < 0.001), with 35 percent

of the variance in service loyalty explained by those independent variables

As expected, overall service quality (QUALITY) had a strong impact onservice loyalty (coefficient = 0.24), supporting H1; its impact wasconsistently strong across the two service categories A further analysis wasconducted to determine the relative importance of the five SERVQUALdimensions in influencing customers' overall quality perceptions This wasdone by regressing the overall QUALITY ratings on the scores for theindividual dimensions, which was calculated by the procedure suggested bythe gap theory (Parasuraman et al., 1985, 1988, 1991; see Appendix for thosemeasures) As shown in Table IV, the results indicate that QUALITY wasinfluenced by all the five dimensions except tangibles (±0.04), implying thatcustomers did not consider up-to-date facilities and equipments to be asimportant as the actual service outcomes from these tangibles The resultsalso point out the industry-specific nature of the relative importance of

Independent variable coeff t-value coeff t-value coeff t-value Overall service quality

coeff t-value coeff t-value coeff t-value

Effects on service loyalty

SERVQUAL dimensions (see Parasuraman et al., 1988, 1991, for similarfindings); while reliability and responsiveness were the most important in thebank case (0.23 and 0.17, respectively), empathy was the most critical in thetravel agency case (0.25).

Two economic cost variables, SERVCOST and SERVTIME, were notsignificant in either cases One possible reason for this finding would be thatthe economic cost factors involved in receiving banking or travel agencyservices are generally not so high or well-differentiated across serviceproviders (overall means = ±1.43 for SERVCOST and ±1.73 forSERVTIME, respectively) that customers do not consider them to beextremely important in determining service loyalty Consequently, H2 andH3 were not supported The impact of the transaction difficulty involved inassessing the service provider's performance, DIFFICULT, did not reachsignificance, either, contrary to H4 An explanation for this would be thatbanking or travel agency services are not so complicated (overall mean =

±0.46) that customers can monitor service performance relatively easily, andthus do not consider DIFFICULT to be critical However, another transactioncost variable, SPECIFIC, or the service provider's specific knowledge aboutcustomers, proved to be an important factor contributing to service loyalty(0.23), supporting H5 Its importance was greater for banks than for travelagencies (0.30 and 0.19, respectively)

The effects of three switching cost variables on service loyalty were highlysignificant overall ± information search cost (INFOCOST; 0.29),

substitutability (SUBSTITUTE; ±0.21), and geographic proximity(PROXIMITY; 0.26), supporting H6, H8, and H9 A close examination,however, revealed some industry differences in the magnitude, if not in thedirection, of the effects of these variables Specifically, SUBSTITUTE in thebank case (±0.11) was not as important as that in the travel agency case (±0.41), and the importance of PROXIMITY in the travel agency case (0.07)was substantially lower than that in the bank case (0.27) Finally, perceivedrisk, or RISK, did not reach significance, contrary to H7 Since thishypothesis was not supported, no further analysis was conducted for theindividual dimensions of perceived risk One possible interpretation of thiswould be that the level of risk generally perceived by consumers in selecting

a bank or a travel agency is not so high (mean = ±1.00) that they do notregard risk as a critical determinant of service loyalty

DiscussionThe overall pattern of the study results supports the notion that serviceloyalty is determined not only by perceived service quality, but also by costconsiderations that arise from present transactions and future switchingpossibilities The implications and limitations of the study are discussed inthis section

Research implicationsSince Shostack (1977) proposed the idea of ``breaking free from productmarketing,'' a considerable amount of research has been accumulated in thedomain of services marketing A significant portion of the research has beendevoted to developing and improving methodological tools to measure theconcept of service quality (Fisk et al., 1993) For consumers, however,service quality evaluations represent only one stage in the process ofdetermining patronage or repatronage of the service provider; service quality

is compared with service cost, and on the basis of the comparison, servicevalue and patronage intentions are determined (Bolton and Drew, 1991b;

Two economic costvariables

Idiosyncrasies

Zeithaml, 1988) For marketers, it is the outcomes of the comparison processthat count, i.e patronage or repatronage intentions In this sense, this studyextends the service quality research framework and sheds light on the areasthat have been relatively neglected In particular, the study demonstrates thatconsumers do factor service-related costs into their patronage decisions, and

it identifies several types of such costs It is believed that the conceptual andmethodological framework suggested in the study contributes to

understanding service provider/customer relationships, and can be used forinvestigating the idiosyncrasies of these relationships in a wider spectrum ofservice industries

Managerial implicationsThe study results have several implications for service marketers First, theresearch shows that the service provider's specific knowledge aboutcustomers (SPECIFIC) has a significant positive impact on customer loyalty.This finding underscores the importance of service customization as a viablestrategy at least for the two service categories examined in the study In fact,some travel agencies are very successful through service customization Theyassign frequent travelers to a specific travel agent to coordinate all travelarrangements The travel agent develops a personal file for each traveler withsuch information as seating preference and preferred form of payment, anduses the information to custom-design services Travelers, in this situation,would like to become ``regular customers,'' expecting that they can get betterservices and higher satisfaction from these types of travel agencies

Second, the results show that the cost involved in finding a new serviceprovider (INFOCOST) and the geographic proximity to the current serviceprovider (PROXIMITY) have significant positive influences on customerloyalty These results imply that not all customers are loyal to their serviceprovider because they are truly satisfied with the service; instead some might

be loyal simply because they do not want to go through the inconvenience ofswitching, or they might prefer the convenient locations Such customers areready to switch whenever possible, e.g when they find a better alternativethat can eliminate or reduce switching costs

Finally, it is revealed that the substitutability of the current service provider(SUBSTITUTE) has a significant negative impact on customer loyalty,suggesting that customers may switch if offered a more convenient alternateservice Service firms, especially those operating in marginally differentiatedindustries such as banking and travel, should be cognizant that the quality ofthe services they provide may be easily duplicated by competitors Since thesuccess and profitability of service companies depend upon satisfied repeatcustomers, marketers must identify dissatisfied customers, listen to theirproblems, and develop effective programs which satisfy their needs andwants and eliminate expectation gaps Only in doing so can marketers beproactive toward future competitive threats

Customer loyalty is a vital element for survival and growth for many servicefirms To develop customer loyalty and maintain a long-term relationshipwith customers, service firms should continually improve and differentiateservice quality, thus positively influencing customers' perceptions andensuring positive outcomes to their cost/benefit analysis

Limitations andfuture researchInterpretation of the results should be tempered by the recognition of themethodological limitation of the study The limitation pertains to the use of

Implications for servicemarketers

Significant negative impact

on customer loyalty