Understanding the effect of customer relationship management effors on customer retention and customer share development

Peter C. Verhoef

Understanding the Effect of
Customer Relationship Management
Efforts on Customer Retention and
Customer Share Development
Scholars have questioned the effectiveness of several customer relationship management strategies. The author
investigates the differential effects of customer relationship perceptions and relationship marketing instruments on
customer retention and customer share development over time. Customer relationship perceptions are considered
evaluations of relationship strength and a supplier’s offerings, and customer share development is the change in
customer share between two periods. The results show that affective commitment and loyalty programs that pro-
vide economic incentives positively affect both customer retention and customer share development, whereas
direct mailings influence customer share development. However, the effect of these variables is rather small. The
results also indicate that firms can use the same strategies to affect both customer retention and customer share
development.

ustomer relationships have been increasingly studied tics can succeed in developing customer share in consumer

C in the academic marketing literature (Berry 1995;
Dwyer, Schurr, and Oh 1987; Morgan and Hunt
1994; Sheth and Parvatiyar 1995). An intense interest in cus-
markets (Dowling 2002; Dowling and Uncles 1997).
Several studies have considered the impact of CRP on
either customer retention or customer share, but not on both
tomer relationships is also apparent in marketing practice (e.g., Anderson and Sullivan 1993; Bolton 1998; Bowman
and is most evident in firms’ significant investments in cus- and Narayandas 2001; De Wulf, Odekerken-Schröder, and
tomer relationship management (CRM) systems (Kerstetter Iacobucci 2001). A few studies have considered the effect of
2001; Reinartz and Kumar 2002; Winer 2001). Customer RMIs on customer retention (e.g., Bolton, Kannan, and
retention rates and customer share are important metrics in Bramlett 2000). In contrast, the effect of RMIs on customer
CRM (Hoekstra, Leeflang, and Wittink 1999; Reichheld share has been overlooked. Furthermore, most studies focus
1996). Customer share is defined as the ratio of a customer’s on customer share in a particular product category (e.g.,
purchases of a particular category of products or services Bowman and Narayandas 2001). Higher sales of more of the
from supplier X to the customer’s total purchases of that cat- same product or brand can increase this share; however,
egory of products or services from all suppliers (Peppers and firms that sell multiple products or services achieve share
Rogers 1999). increases by cross-selling other products. Moreover, no
To maximize these metrics, firms use relationship mar- study has considered the effect of CRPs and RMIs on both
keting instruments (RMIs), such as loyalty programs and customer retention and customer share. It is often assumed
direct mailings (Hart et al. 1999; Roberts and Berger 1999). in the literature that the same strategies used for maximizing
Firms also aim to build close relationships with customers to customer share can be used to retain customers; however,
enhance customers’ relationship perceptions (CRPs). recent studies indicate that increasing customer share might
Although the impact of these tactics on customer retention require different strategies than retaining customers (Blat-
has been reported (e.g., Bolton 1998; Bolton, Kannan, and tberg, Getz, and Thomas 2001; Bolton, Lemon, and Verhoef
Bramlett 2000), there is skepticism about whether such tac- 2002; Reinartz and Kumar 2003).
Prior studies have used self-reported, cross-sectional
data that describe both CRPs and customer share (e.g., De
Peter C. Verhoef is Assistant Professor of Marketing, Department of Mar- Wulf, Odekerken-Schröder, and Iacobucci 2001). The use of
keting and Organization, Rotterdam School of Economics, Erasmus Uni-
versity, Rotterdam. The author gratefully acknowledges the financial and such data may have led to overestimation of the considered
data support of a Dutch financial services company. The author thanks associations because of methodological problems such as
Bas Donkers, Fred Langerak, Peter Leeflang, Loren Lemon, Peeter Ver- carryover and backfire effects and common method variance
legh, Dick Wittink, and the four anonymous JM reviewers for their helpful (Bickart 1993). Such data cannot establish a causal relation-
suggestions. The author also acknowledges the comments of research ship; indeed, the argument could be made that causality
seminar participants at the University of Groningen, Yale School of Man- works the other way (i.e., I am loyal, therefore I like the
agement, Tilburg University, and the University of Maryland. Finally, he
company) (Ehrenberg 1997). Longitudinal data rather than
acknowledges his two dissertation advisers, Philip Hans Franses and
Janny Hoekstra, for their enduring support. cross-sectional data should be used to establish the causal
relationship between customer share and its antecedents.

Journal of Marketing
30 / Journal of Marketing, October 2003 Vol. 67 (October 2003), 30–45

I have the following research objectives: First, I aim to tomer relationship largely depends on the applied RMIs
understand the effect of CRPs and RMIs on customer reten- (Bhattacharya and Bolton 2000; Christy, Oliver, and Penn
tion and customer share development over time. Second, I 1996; De Wulf, Odekerken-Schröder, and Iacobucci 2001).
examine whether the effect of CRPs and RMIs on customer Moreover, because of the increasing popularity of CRM
retention and customer share development is different. My among businesses, an increasing number of firms are using
study analyzes questionnaire data on CRPs, operational data RMIs.
on the applied RMIs, and longitudinal data on customer In the model, I also include customers’ past behavior in
retention and customer share of a (multiservice) financial the relationship as control variables, which might capture
service provider. inertia effects that are considered important determinants of
customer loyalty in business-to-consumer markets (Dowling
Literature Review and Uncles 1997; Rust, Zeithaml, and Lemon 2000). Past
customer behavioral variables (e.g., relationship age, prior
CRPs and Customer Behavior customer share) can also be indicators of past behavioral
Table 1 provides an overview of studies that report the effect loyalty, which often translates into future loyalty. Prior
of CRPs on customer behavior, and it describes the depen- research suggests that the type of product purchased in the
dent variables, the design and context of the study, the CRPs past is an indicator of future cross-selling potential (e.g.,
studied, and the effect of CRPs on behavioral customer loy- Kamakura, Ramaswami, and Srivastava 1991).
alty measures (which can be self-reported or actual observed
loyalty measures). Table 1 shows that the results of studies Hypotheses
that relate CRPs to actual customer behavior are mixed.
CRPs
RMIs and Customer Behavior Relationship marketing theory and customer equity theory
Table 2 provides an overview of the limited number of aca- posit that customers’ perceptions of the intrinsic quality of
demic studies that consider the effect of RMIs. The majority the relationship (i.e., strength of the relationship) and cus-
of the studies have focused on loyalty or preferential treat- tomers’ evaluations of a supplier’s offerings shape cus-
ment programs, and the results show mixed effects of these tomers’ behavior in the relationship (Garbarino and Johnson
programs on customer loyalty. Despite the intensive use of 1999; Rust, Zeithaml, and Lemon 2000; Woodruff 1997).
direct mailings in practice, their effect on customer loyalty The most prominent perception representing the strength of
has almost been ignored. More important, the effect of RMIs the relationship is (affective) commitment (Moorman, Zalt-
on customer share development over time has not been man, and Desphandé 1992; Morgan and Hunt 1994).
investigated. Because satisfaction and payment equity are important con-
structs with respect to the evaluation of a supplier’s offerings
Conceptual Model (Bolton and Lemon 1999), I included these three constructs
in the model. The two categories of constructs differ in
Figure 1 shows the conceptual model. In this model, I con-
terms of both content and time orientation: Affective com-
sider customer retention and customer share development
between two periods (T1 and T0) as the dependent variables, mitment is forward looking, whereas satisfaction and pay-
which are affected by CRPs and RMIs. Because I consider ment equity are retrospective evaluations.
customer retention and customer share development as two In the customer equity and relationship marketing liter-
separate processes, relationship maintenance and relation- ature, other CRPs that are not included in my model are
ship development, the underlying hypotheses of the model often studied. Trust and brand perceptions are the most
explicitly predict that different constructs of CRPs, and dif- prominent of these variables (Morgan and Hunt 1994; Rust,
ferent RMIs influence customer retention and customer Zeithaml, and Lemon 2000). I did not include brand percep-
share development. The rationale for this distinction is that tions because the focus is on current customers. My con-
a customer’s decision to stay in a relationship with a firm tention is that the brand is especially significant in attracting
may be different from his or her incremental decision to add new customers. During the relationship, the brand probably
or drop existing products. Consistent with this notion, Blat- influences affective commitment (Bolton, Lemon, and Ver-
tberg, Getz, and Thomas (2001) argue that customer reten- hoef 2002). I did not include trust, because trust should be
tion is not the same as customer share, because two firms considered merely an antecedent of satisfaction and com-
could retain the same customer. Reinartz and Kumar (2003) mitment (Geyskens, Steenkamp, and Kumar 1998). No
suggest that relationship duration and customer share should direct effect on customer behavior should be expected.
be considered as two separate dimensions of the customer
Affective Commitment
relationship. Bolton, Lemon, and Verhoef (2002) propose
that the antecedents of customer retention might be different Commitment is usually defined as the extent to which an
from the antecedents of cross-buying behavior. I explicitly exchange partner desires to continue a valued relationship
address these differences in the hypotheses. (Moorman, Zaltman, and Desphandé 1992). I focus on the
The inclusion of CRPs as antecedents of retention and affective component of commitment, that is, the psycholog-
customer share development is based on relationship mar- ical attachment, based on loyalty and affiliation, of one
keting theory, which suggests that CRPs affect behavioral exchange partner to the other (Bhattacharya, Rao, and
customer loyalty. I included RMIs because a successful cus- Glynn 1995; Gundlach, Achrol, and Mentzer 1995).

Customer Relationship Management Efforts / 31

TABLE 1
Overview of Studies on Effect of CRPs on Behavioral Loyalty

Behavioral Loyalty Included Perceptions Additional
Measurement Examples of Studies Study Design Study Context (Effect) Results/Comments

Self-Reported
Purchase intentions Anderson and Sullivan Experiment Various industries Satisfaction (+)
(1993)
Morgan and Hunt (1994) Cross-sectional Channels Benefits (+),
commitment (+)
Zeithaml, Berry, and Cross-sectional Various industries Service quality (+)
Parasuraman (1996)
Garbarino and Johnson Cross-sectional Theater visitors Satisfaction (+), Effect depends on relationship

32 / Journal of Marketing, October 2003
(1999) commitment (–) orientation of customer
Mittal, Kumar, and Longitudinal Car market Satisfaction (+)
Tsiros (1999)
Customer share Macintosch and Cross-sectional Retailing Commitment (+)
Lockshin (1997)
De Wulf, Odekerken- Cross-sectional Retailing Relationship
Schröder, and quality (+)
Iacobucci (2001)
Bowman and Cross-sectional Grocery brands Satisfaction (+) Quadratic effect of satisfaction
Narayandas (2001)

Observed
Customer retention Gruen, Summers, and Cross-sectional Professional Commitment (0)
and/or relationship Acito (2000) association
duration
Bolton (1998) Longitudinal Telecommunications Satisfaction (+) Effect of satisfaction enhanced
by relationship age
Bolton, Kannan, and Longitudinal Credit card Satisfaction (+), Performance differences with
Bramlett (2000) payment equity (+) other firms are important
Mittal and Kamakura Longitudinal Car market Satisfaction (+) Effect of satisfaction
(2001) moderated by consumer
characteristics
Lemon, White, and Longitudinal Entertainment Satisfaction (0) Effect of satisfaction mediated
Winer (2002) by future expected service
usage
Service usage Bolton and Lemon Longitudinal Telecommunications, Satisfaction (+) Payment equity positively
(1999) entertainment affects satisfaction
Bolton, Kannan, and Longitudinal Credit card Satisfaction (+), Performance differences with
Bramlett (2000) payment equity (+) other firms are important
Cross-buying Verhoef, Franses, Longitudinal Financial services Satisfaction (0), Effect of satisfaction and
and Hoekstra (2001) payment equity (0) payment equity enhanced by
relationship age

Effect on customer retention. Given the previous defini- from a particular supplier (Oliver and Winer 1987). This
tion of affective commitment, it might be expected that this depends on, among other things, the customer’s satisfaction
type of commitment affects customer retention positively. In level. As a consequence, customers who are more satisfied
line with this, researchers who relate commitment to self- are more likely to remain customers. Thus:
reported behavior, such as purchase intentions, usually find
H3: Satisfaction positively affects customer retention.
that commitment positively affects customer loyalty (e.g.,
Garbarino and Johnson 1999; Morgan and Hunt 1994). Effect on customer share development. Although a posi-
However, the appearance of such an effect has recently been tive relationship between satisfaction and customer share
questioned (Gruen, Summers, and Acito 2000; MacKenzie, has been demonstrated in a single product category (Bow-
Podsakoff, and Ahearne 1998). Despite this, I hypothesize man and Narayandas 2001), this does not necessarily imply
the following: that satisfaction also positively affects customer share devel-
H1: Affective commitment positively affects customer opment for a multiservice provider. A theoretical explana-
retention. tion for the absence of such an effect could be that positive
evaluations of currently consumed products or services do
Effect on customer share development. Relationship not necessarily transfer to other offered products or services.
marketing theory posits that because affectively committed In other words, satisfied customers are not necessarily more
customers believe they are connected to the firm, they dis- likely to purchase additional products or services (Verhoef,
play positive behavior toward the firm. As a consequence, Franses, and Hoekstra 2001). Another explanation is that
affectively committed customers are less likely to patronize though customer retention relates to the focal supplier alone,
other firms (Dick and Basu 1994; Morgan and Hunt 1994; customer share development also involves competing sup-
Sheth and Parvatiyar 1995). In other words, committed cus- pliers. As a result, development of a customer’s share might
tomers are more (less) likely to increase (decrease) their cus- be affected more by the actions of competing suppliers than
tomer share for the focal supplier over a period of time. by the focal firm’s prior performance. Thus, I do not expect
H2: Affective commitment positively affects customer share satisfaction to have a positive effect on customer share
development over time. development.

Satisfaction Payment Equity
I define satisfaction in this study as the emotional state that Payment equity is defined as a customer’s perceived fairness
occurs as a result of a customer’s interactions with the firm of the price paid for the firm’s products or services (Bolton
over time (Anderson, Fornell, and Lehmann 1994; Crosby, and Lemon 1999, p. 173) and is closely related to the cus-
Evans, and Cowles 1990). Szymanski and Henard’s (2001) tomer’s price perceptions. Payment equity is mainly affected
meta-analysis shows that satisfaction has a positive impact by the firm’s pricing policy. As a result of its grounding in
on self-reported customer loyalty. fairness, a firm’s payment equity also depends on competi-
Despite such positive results in the literature, the link tors’ pricing policies and the relative quality of the offered
between satisfaction and actual customer loyalty has been services or products.
questioned (e.g., Jones and Sasser 1995). Researchers have Effect on customer retention. Higher payment equity
searched for a better understanding of this link and have pro- (i.e., price perceptions) leads to greater perceived utility of
posed a nonlinear relationship between satisfaction and cus- the purchased products or services (Bolton and Lemon
tomer behavior (e.g., Anderson and Mittal 2000; Bowman 1999). As a result of this greater perceived utility, customers
and Narayandas 2001). Other studies have shown that rela- should be more likely to remain with the firm. Conse-
tionship age, product usage, variety seeking, switching quently, payment equity should have a positive effect on
costs, consumer knowledge, and sociodemographics (e.g., customer retention. This is consistent with empirical studies
age, income, gender) moderate the link between satisfaction that show that payment equity positively affects customer
and customer loyalty (Bolton 1998; Bowman and Narayan- retention (Bolton, Kannan, and Bramlett 2000; Varuki and
das 2001; Capraro, Broniarczyck, and Srivastava 2003; Colgate 2001). Thus:
Homburg and Giering 2001; Jones, Mothersbaugh, and
Beatty 2001; Mittal and Kamakura 2001). Finally, dynamics H4: Payment equity positively affects customer retention.
during the relationship may also affect this link. Customers
Effect on customer share development. Although I
update their satisfaction levels using information gathered
expect payment equity to have a positive effect on customer
during new interaction experiences with the firm, and this
retention, I do not necessarily expect this to be true for cus-
new information may diminish the effect of prior satisfac-
tomer share development. There are two reasons payment
tion levels (Mazursky and Geva 1989; Mittal, Kumar, and
equity may have no effect on customer share development.
Tsiros 1999).
First, literature on price perceptions suggests that customers
Effect on customer retention. Despite the apparent with higher price perceptions are more likely to search for
absence of an empirical link between satisfaction and behav- better prices (Lichtenstein, Ridgway, and Netemeyer 1993).
ioral customer loyalty, several studies show that satisfaction Intuitively, the suggestion that such customers are less loyal
affects customer retention (Bolton 1998; Bolton, Kannan, makes sense. For example, customers of discounters (with
and Bramlett 2000). The underlying rationale is that cus- high scores on price perceptions) are known to visit the
tomers aim to maximize the subjective utility they obtain greatest number of stores in their search for the best bargain.


TABLE 2
Studies on Effect of RMIs on Behavioral Loyalty

Study
RMI Study Loyalty Measure Study Design Context Results

Direct mail Bawa and Shoemaker (1987) Aggregated purchase Panel design Grocery Short-term positive effect
shares brands on purchase rates

De Wulf, Odekerken-Schröder, Customer share Cross-sectional survey, Retailing No effect
and Iacobucci (2001) perceptions on direct
mail use

Loyalty programs Dowling and Uncles (1997) No empirical data — — —

Sharp and Sharp (1997) Aggregated penetration, Aggregated panel data Retailing No convincing effect of
average purchase loyalty programs
frequency, customer share,
sole buyers

Rust, Zeithaml, and Lemon Purchase intentions Cross-sectional survey Airlines Positive effect
(2000) data, perceptions on
loyalty program use

Bolton, Kannan, and Bramlett Customer retention, service Longitudinal Credit card Positive effect on retention
(2000) usage and service usage

De Wulf, Odekerken-Schröder, Customer share Cross-sectional survey Retailing No effect
and Iacobucci (2001) data, perceptions on
preferential treatment
programs

FIGURE 1
Conceptual Model

CRPs RMIs

Satisfaction H3
H6a
H4 Customer retention T 1 T0 Loyalty program
H6b
Payment equity H1
H5
H2 ∆ Customer share T 1 T0 Direct mailings

Affective commitment

Control Variables
Customer share T0
Relationship age T0
Type of service purchased T0

According to this reasoning, customers with better price per- Direct Mailings
ceptions are more likely to decrease customer share over Direct mailings have some unique characteristics: enable-
time. Second, as is satisfaction, a customer’s payment equity ment of personalized offers, no direct competition for the
is based on the customer’s awareness of the prices of ser- attention of the customer from other advertisements, and a
vices or products purchased from the focal firm in the past capacity to involve the respondent (Roberts and Berger
(Bolton, Lemon, and Verhoef 2002). However, the prices of 1999). Because direct mailings focus on creating additional
additional services or products from the focal supplier might sales, I do not expect them to influence customer retention.
be different from the currently purchased services or prod- Moreover, the data do not enable me to relate direct mailings
ucts. Therefore, a high payment equity score may not indi- to customer retention.
cate that the customer will purchase other products or
Effect on customer share development. There are several
services from the same supplier. As a consequence, I do
theoretical reasons direct mailings should positively influ-
not expect payment equity to affect customer share
ence customer share development. First, direct mailings can
development. create interest in a (new) service and thereby lead to a final
RMIs purchase (Roberts and Berger 1999). Second, the personal-
ization afforded by direct mailings may increase perceived
Bhattacharya and Bolton (2000) suggest that RMIs are a relationship quality, because customers are approached with
subset of other marketing instruments that are specifically individualized communications that appeal to their specific
aimed at facilitating the relationship, and they distinguish needs and desired manner of fulfilling them (De Wulf,
between loyalty or reward programs and tailored promo- Odekerken-Schröder, and Iacobucci 2001; Hoekstra,
tions. In addition, RMIs can be classified according to Leeflang, and Wittink 1999). Third, according to the sales
Berry’s (1995) first two levels of relationship marketing. At promotions literature, the short-term rewards (i.e., price dis-
the first level (Type I), firms use economic incentives, such counts) offered by direct mailings may motivate customers
as rewards and pricing discounts, to develop the relation- to purchase additional services and thus increase customer
ship. At the second level (Type II), instruments include more share. In support of this claim, Bawa and Shoemaker (1987)
social attributes. By using Type II instruments, firms attempt report short-term gains in redemption rates of direct mail
to give the customer relationship a personal touch. coupons. I hypothesize the following:
In this study, I focus on two specific Type I RMIs: direct H5: Direct mailings positively affect customer share develop-
mailings and loyalty programs. Direct mailings usually are ment over time.
personally customized offers on products or services that the
customer currently does not purchase. In most cases, price Loyalty Programs
discounts or other sales promotions (e.g., gadgets) are used Effect on customer retention and customer share devel-
to entice the customer to buy. I focus on direct mailings that opment. There are several theoretical reasons the reward-
are a “call to action” rather than only a reinforcing mecha- based loyalty program being studied should positively affect
nism for the relationship (e.g., thank-you letters). The loy- both customer retention and customer share development.
alty program I include in the study is a reward program that First, psychological investigations show that rewards can be
provides price discounts based on the number of products or highly motivating (Latham and Locke 1991). Research also
services purchased and the length of the relationship. shows that people possess a strong drive to behave in what-


ever manner necessary to achieve future rewards (Nicholls products, and customer characteristics. In the second (T1)
1989). According to Roehm, Pullins, and Roehm (2002, p. survey, I collected data on customer ownership of various
203), it is reasonable to assume that during participation in insurance products.
a loyalty program, a customer might be motivated by pro- Although the company whose data I used offers other
gram incentives to purchase the program sponsor’s brand products, such as loans, I limited the study to the category of
repeatedly. insurance products. The rationale for this limitation is that
Second, because the program’s reward structure usually customers usually buy each type of insurance product from
depends on prior customer behavior, loyalty programs can a single insurance carrier (i.e., insurance type X [life insur-
provide barriers to customers’ switching to another supplier. ance] from insurance carrier Y [i.e., Allianz Life Com-
For example, when the reward structure depends on the pany]), but this does not necessarily hold for other financial
length of the relationship, customers are less likely to switch products or services. For example, it is well known that
(because of a time lag before the same level of rewards can many customers have savings accounts at several financial
be received from another supplier). It is well known that institutions. Moreover, the insurance market is the most
switching costs are an important antecedent of customer important market for this company in terms of the number
loyalty (Dick and Basu 1994; Klemperer 1995). of customers and customer turnover (approximately 90%).
Despite the theoretical arguments in favor of the positive As a result of this choice, the sample is restricted to those
effect of loyalty programs on customer retention and cus- customers who purchase insurance products only from the
tomer share development, several researchers have ques- company. This resulted in a usable sample size of 1677 cus-
tioned this effect (e.g., Dowling and Uncles 1997; Sharp and tomers for the first measurement (T0) and 918 for the second
Sharp 1997). In contrast, Bolton, Kannan, and Bramlett measurement (T1).
(2000) and Rust, Zeithaml, and Lemon (2000) show that
loyalty programs have a significant, positive effect on cus- Contents of the Company Customer Database
tomer retention and/or service usage. In this study, I build on The company’s customer database provided data on the past
the theoretical argument in favor of the positive effect that behavior of individual customers and the company RMIs
loyalty programs have on customer retention and customer directed at individual customers. The past customer behav-
share development. ior data cover two periods. The first period starts at the
H6: Loyalty program membership positively affects (a) cus- beginning of a relationship between the company and the
tomer retention and (b) customer share development. customer and ends at T0 (this period differs among cus-
tomers). The data on past customer behavior included vari-
ables such as number of insurance policies purchased, type
Research Methodology of insurance policies purchased, and relationship length.
The second period covers the interval between T0 and T1.
Research Design For this period, the database provided data about which cus-
I combined survey data from customers of a Dutch financial tomers left the company and the number of company insur-
services company with data from that company’s customer ance policies a customer owned at T1.
database. I used a panel design, displayed in Figure 2, to col- The company’s customer database contains the follow-
lect the data. I collected the survey data at two points in ing information on RMIs: loyalty program membership at
time: T0 and T1. I used the first (T0) survey to measure CRPs T0 and the number of direct mailings sent between T0 and
of the company, customer ownership of various insurance T1. Every customer who purchases one or more financial

FIGURE 2
Panel Design

Data from Customer Database

Start of T T
0 1
Relationship
(Survey 1 Among (Survey 2 Among
Customers) Customers Interviewed
in Survey 1)


services from the company can become a member of the Validation of CRPs
loyalty program (an opt-in program). At the end of each The final measures are reported in the Appendix. The scales
year, the program gives customers a monetary reward based for commitment and satisfaction have reasonable coefficient
on the number of services purchased and the age of the rela- alphas. For payment equity, I report a correlation coefficient
tionship. Because the company uses regression-type models of .49, which is not considerably high.2 However, note that
to select the customers with the highest probability of the reported composite reliabilities of all scales are suffi-
responding to direct mailings, the number of direct mailings cient (Bagozzi and Yi 1988). I applied confirmatory factor
sent differs among customers. analysis in Lisrel 83 to further assess the quality of the mea-
Customer Survey Data Collection sures (Jöreskog and Sörbom 1993), and I achieved the fol-
lowing model fit: χ2 = 217.4 (degrees of freedom [d.f.]) =
At T0, customer survey data were collected by telephone 51, p <.01), χ2/d.f. = 4.26 (d.f. = 1, p < .05), goodness-of-fit
from a random sample of 6525 customers of the company. A index = .98, adjusted goodness-of-fit index = .97, compara-
quota sampling approach was used to obtain a representative tive fit index = .98, and root mean square error of approxi-
sample. I received data from 2300 customers (35% response mation = .04. These fit indexes satisfy the criteria for a good
rate). After those responses with too many missing values model fit (Bagozzi and Yi 1988; Baumgartner and Homburg
were deleted, a sample size of 1986 customers remained. At 1996). A series of χ2 difference tests on the respective fac-
T1, I again collected data from those customers, except for tor correlations provided further evidence for discriminant
those who left the company between T0 and T1. In the sec- validity (Anderson and Gerbing 1988). On the basis of these
ond data collection effort, 1128 customers were willing to results, I summed the scores on the items of each construct.
cooperate (65% response rate). To assess nonresponse bias The means, standard deviations, and correlation matrix are
at T1, I tested whether respondents and nonrespondents dif- shown in Table 3.
fered significantly with respect to customer share at T0. A t-
test does not reveal a significant difference (p = .36). Thus, Measurement of Dependent Variable
I conclude that there is no nonresponse bias.
An often-used method of measuring customer share is ask-
Measurement of CRPs ing customers to report the number of purchases of the focal
For the measurement of CRPs (i.e., affective commitment, brand they normally make (Bowman and Narayandas 2001;
satisfaction, and payment equity), I adapted existing scales De Wulf, Odekerken-Schröder, and Iacobucci 2001). In this
to fit the context of financial services. For the affective com- study, I sought a more objective measure. In line with the
mitment scale, I adapted items from the studies of Anderson conceptualization of customer share, I define customer share
and Weitz (1992), Garbarino and Johnson (1999), and of customer i for supplier j in category k at time t as
Kumar, Scheer, and Steenkamp (1995). To measure satisfac-
tion, I adapted Singh’s (1990) scale and added four new
items. Finally, I based the payment equity scale on items
adapted from Bolton and Lemon’s (1999) and Singh’s
(1990) studies. 2I report correlation coefficient rather than Cronbach’s alpha
To assess construct validity and clarify wording, the because I used only two items. Cronbach’s alpha is designed to test
original scales were tested by a group of 12 marketing aca- the interitem reliability of a scale by comparing every combination
demics and 3 marketing practitioners familiar with customer of each item with all other items in the scale as a group. Because
relationships. Subsequently, the scales were tested by a ran- there is no group with which each item can be compared in a two-
item scale (only the other item), Cronbach’s alpha is meaningless
dom sample of 200 customers of the company. On the basis for two-item scales. It might also be argued that one of the single
of interitem correlations, item-to-total correlations, coeffi- items would be better suited for measuring the construct from a
cient alpha, and exploratory and confirmatory factor analy- content validity perspective. To check this, I also estimated the
sis, I reduced the set of items in each scale.1 models (see Tables 4 and 5) with a single item as an antecedent.
For both items, the effect of payment equity remained insignificant
1I follow Steenkamp and van Trijp’s (1991) proposed method, in the two models. Because in general multiple-item measurement
using exploratory factor analysis and then confirmatory factor is preferred over single-item measurement, I report the model
analysis to validate marketing constructs. results of the summated two-item scores.

TABLE 3
Means (Standard Deviation) and Correlation Matrix Independent Variables

Mean X1 X2 X3 X4 X5 X6

X1 Commitment 2.960 (.77) 1.00
X2 Satisfaction 3.750 (.44) .37** 1.00
X3 Payment equity 3.410 (.56) .14** .21** 1.00
X4 Direct mail 3.510 (2.12) .01 .02 .01 1.00
X5 Loyalty program .300 (.46) .09** .14** .03 .56** 1.00
X6 Log customer share T0 –.152 (.66) .12** .09** .06* .48** .53** 1.00
*p < .05.
**p < .01.


Number of services covariates (past behavior) as independent variables and the
purchased in category k mean-centered composites of the items in the relationship
at supplier j at time t perception scales (perceptions; e.g., affective commitment,
(1) Customer share i, j,k,t = .
Number of services satisfaction, payment equity). Finally, I include RMIs. For
purchased in category k the loyalty program, I constructed a dummy variable that
from all suppliers at time t indicated whether the customer was a member of the loyalty
Data for the numerator were available from the company program at T0. I dealt with the number of direct mailings
customer database; however, data for the denominator were sent to a customer as follows: Because the company stops
generally not stored in the company customer database. direct mailing customers when they defect, the number of
Therefore, I asked customers in the survey which insurance direct mailings was not included in the probit model for cus-
products (of both the company and competitors) they owned tomer retention. Because customers leave during the period
at T0 and at T1. covered in the study, the number of mailings could be cor-
related with defection. However, this correlation is not due
Analysis to the positive effect of direct mailings on customer reten-
tion; rather, it is the result of the company’s mailing policy.
The theoretical distinction between customer retention and The foregoing results in the following two equations:
customer share development has implications for my analy-
sis. As a result of this distinction, I use a dual approach. I (2) P(retention = 1) = α0 + α1past behavior0 + α2perceptions0
first estimate a probit model to explain customer retention or + α3RMIs0 – 1, and
defection for the remaining sample after T0 (N = 1677). Sec-
ond, I use a regression model to explain customer share (3) Log(CS1) – log(CS0) = β0 + β1past behavior0
development over time for the customers who remain with
the company. A serious issue with this type of approach is + β2perceptions0 + β3RMIs0 – 1 + β4Heckman correction.
that the explanatory variables explaining customer retention In Equations 2 and 3, I provide the formulation of the model
also explain customer share development. As a conse- in the form of matrices in which each α or β may comprise
quence, the regression parameters may be biased (Franses several separate parameters. For example, in the case of β2,
and Paap 2001). I apply the Heckman (1976) two-step pro- there are three different parameters for the effect of com-
cedure to correct for this bias. Using this procedure, I mitment, satisfaction, and payment equity.
include the so-called Heckman correction term (or inverse
Mills ratio) in the regression model for customer share
development. This correction term is calculated by means of Hypothesis Testing
outcomes of the probit model for customer retention. This
modeling approach is also known as the Tobit2 model Customer Retention
(Franses and Paap 2001). Because the inclusion of this cor- Approximately 6.4% of the 1677 customers in the sample
rection term may cause heteroskedasticity, I apply White’s defected during the period of the study.3 I report the estima-
(1980) method to adjust for heteroskedasticity. Another tion results of Equation 3 in Table 4. The first model (which
issue with the approach is that restricting the sample in the only includes control variables with respect to past customer
customer share development regression model to remaining behavior) explains approximately 17% of the variance and is
customers might restrict the potential variance in the depen- significant (p < .01). The coefficients of the included control
dent variable, thus affecting the estimation results. To assess variables intuitively have the expected signs. Customers
whether this is true, I calculated the standard deviations for with high prior customer shares and lengthy relationships
the restricted and unrestricted sample. The differences are less likely to defect. Furthermore, the ownership of a
between standard deviations in customer share development coinsurance, damage insurance, car insurance, and/or life
are small: .10 for the unrestricted sample, including defec- insurance product has a positive effect (p < .05). In the sec-
tors, and .09 for the restricted sample. In the empirical mod- ond model, including CRPs, McFadden R2 increases by
eling, I further assess this issue by estimating the customer approximately 1% (p = .06). Only affective commitment has
share development model for the unrestricted sample and a significant, positive effect (p < .01) on customer retention,
comparing the results with those of the restricted sample. in support of H1. I found no effect for either satisfaction or
Because I am interested in the changes in customer share payment equity. These results do not support H3 or H4. Fol-
over time, I use a difference model to test the hypotheses lowing Bolton (1998), I also explored whether relationship
(Bowman and Narayandas 2001). In line with the literature age moderates the effect of satisfaction. The estimation
on market share models, the difference between the logs of results indicate that the interaction term between satisfaction
customer share at T1 and T0 (CS0, CS1) is the dependent
variable in the regression model. This variable can be inter-
preted as the percentage change in customer share over the
3The sample of 1677 for the analysis of the antecedents of cus-
measured period.
tomer retention is much larger than the sample used in the cus-
In both the probit model for customer retention and the tomer share development model, because behavioral data about
regression model for customer share development of the customers’ past purchase behavior were unnecessary in the cus-
customers who remain with the company, I use a hierarchi- tomer retention analysis. Consequently, customers who did not
cal modeling approach. I include the past customer behavior respond in the second survey can be included in this analysis.


TABLE 4
Probit Model Results for Customer Retention (N = 1677)

Hypothesis Model 1 Model 2 Model 3
Variable (Sign) (z-Value) (z-Value) (z-Value)

Constant 1.66 (5.10)** 1.68 (5.03)** 1.58 (4.58)**
Log customer share T0 .34 (2.23)** .33 (2.13)** .30 (1.89)**
Log relationship age .11 (2.32)** .11 (2.14)** .09 (1.79)**
Coinsurance .12 (1.21)** .12 (1.22)** .11 (1.04)**
Damage insurance .78 (4.13)** .78 (4.06)** .74 (3.92)**
Car insurance .36 (2.97)** .33 (2.70)** .33 (2.72)**
Life insurance 1.02 (4.00)** 1.01 (3.99)** 1.00 (3.95)**

Perceptions
Commitment H1 (+) .21 (2.66)** .20 (2.58)**
Satisfaction H3 (+) –.21 (1.52)** –.22 (1.63)**
Payment equity H4 (+) –.03 (.26)** –.03 (.30)**

RMIs
Loyalty program H6a (+) .38 (2.02)**

McFadden R2 .168 .178 .184
Adjusted McFadden R2 .165 .173 .179
Likelihood ratio statistic 127.64** 135.20** 139.68**
(d.f.) (6) (9) (10)
Akaike information criterion .384 .383 .382
*p < .05.
**p < .01.

and relationship age is significant (α = .28; p = .01), in sup- FIGURE 3
port of the idea that relationship age enhances the effect of Customer Share Development (N = 918)
satisfaction. In the third model, with the loyalty program
included, McFadden R2 increases by approximately 1% (p < 500
.05). I found the loyalty program to have a significant, pos-
itive effect (p < .05), in support of H6a.
400
Customer Share Development for Remaining
Customers
300
Figure 3 shows the changes in customer share for the cus-
tomers who did not defect. Although on average changes in
customer share are almost zero, I observed changes in cus- 200
tomer share for approximately 68% of the customers in the
sample (N = 918). The distribution in Figure 3 is symmetri-
cal. For 34% of customers in the sample, I observed nega- 100
tive changes, and for approximately 34%, their customer
shares increased. As a logical consequence, the average for
changes in customer share is zero (i.e., the mean values for 0
customer share at T0 and T1 have approximately the same – .25 .00 .25 .50 .75
value of .285).
The regression results of Equation 3 are reported in
Table 5. The first model (including past customer behavior)
explains approximately 10% of the variance in customer
share changes. The log of customer share at T0 has a nega- nificant, positive effect on customer share development (p <
tive effect on changes in customer share (p < .01). Thus, cus- .05). Thus, I find support for H2. However, I found no sig-
tomers with large (small) customer shares are more likely to nificant effect for either satisfaction or payment equity.
decrease (increase) their customer share in the next period. These results are in line with my expectations that such
Customers who own damage insurance, car insurance, or CRPs do not directly affect customer share development. In
coinsurance are more likely to increase their customer share the third model (which includes RMIs), the loyalty program
(p < .01). The estimation results of the second model (which has a significant, positive effect on customer share develop-
includes CRPs) show that affective commitment has a sig- ment (p < .05). Direct mailings also positively affect cus-


TABLE 5
Regression Model Results of Changes in Customer Share (N = 918)

Hypothesis Model 1 Model 2 Model 3
Variable (Sign) (t-Value) (t-Value) (t-Value)

Constant –.44 (6.84)** –.46 (7.09)** –.52 (7.80)**
Heckman correction .06 (.90) .07 (1.27) .10 (1.48)
Log customer share T0 –.17 (9.97)** –.19 (10.3)** –.20 (11.0)**
Coinsurance .02 (3.83)** .02 (3.92)** .02 (3.26)**
Damage insurance .14 (6.35)** .15 (6.52)** .14 (6.09)**
Car insurance .04 (2.69)** .01 (2.29)* .04 (2.52)**
Legal insurance .03 (1.15) .03 (1.16) .03 (1.16)

Perceptions
Commitment H2 (+) .03 (2.55)* .03 (2.58)**
Satisfaction H3 (+) .00 (.01) –.00 (.21)
Payment equity H4 (+) –.01 (.85) –.01 (.66)

RMIs
Loyalty program H5 (+) .04 (2.22)*
Direct mailing H6 (+) .01 (2.31)*

R2 .10 .11 .13
Adjusted R2 .10 .10 .12
F-value 16.95** 12.21** 11.72**
*p < .05.
**p < .01.

tomer share development (p < .05).4 Thus, both H5 and H6b First, I cannot include direct mailings as an explanatory
are supported. variable because, as I noted previously, no mailings are sent
The Heckman (1976) correction term is not significant, to defectors. Second, because the log of 0 does not exist, the
which implies that selecting only the remaining customers differences in logs of customer share between T1 and T0 for
does not affect the estimation results (Franses and Paap defectors cannot be calculated. A solution to this problem is
2001). It might be argued that leaving out defectors would to impute a share value that is close to 0 (e.g., .001). I used
reduce variance in the customer share development measure, this approach and imputed several different values to assess
which in turn might affect the estimation results. To assess the stability of the results, and the results remained the same
this issue further, I also estimated a model that included the for the different imputations. The estimation results for an
defectors.5 However, there are two problems with the model. imputed value for customer share at T1 for defectors of .001
show that the coefficients of affective commitment and the
4An issue in estimating the effect of direct mailings is that the
loyalty program remain significant, but there is no effect of
company whose data are used does not randomly select customers satisfaction or payment equity. The R2 of the model is .09,
to receive such mailings; the company uses models to target the which is lower than the R2 of .12 of the model that includes
most receptive customers. These models are not known. The com-
pany’s use of such models might lead to an endogeneity problem, only the remaining customers reported in Table 5. Given
which could result in (upwardly biased) inconsistent parameter these results, I conclude that restricting the sample to
estimates for direct mailings. To test for possible endogeneity, I remaining customers does not affect the hypotheses-testing
used the Hausman test that Davidson and MacKinnon (1989) pro- results.
pose. This test does not reveal any evidence for endogeneity (p =
.88).
5Notwithstanding this result, I also used two approaches to cor- Additional Analysis
rect for possible endogeneity. The first approach applied instru-
mental variables using two-stage least squares in the estimation of Mediating Effect of Commitment 6
a system of two equations (Pindyck and Rubinfeld 1998). I used
two sociodemographic variables as instrumental variables: income In the relationship marketing literature, there has been a
and age. I selected these variables because they are often included debate about the mediating role of commitment (Garbarino
in CRM models (Verhoef et al. 2003). The estimation of this model and Johnson 1999; Morgan and Hunt 1994). In this study,
results in the same parameter estimate for direct mailings (.04); commitment may mediate the effect of payment equity and
however, this parameter is only marginally significant (p = .10). satisfaction on customer share development, which in turn
The second approach estimated a system of equations in which two
separate equations are estimated: one with customer share devel-
may explain the nonsignificant effects of both satisfaction
opment as a dependent variable and the other with the number of and payment equity. To test for this mediating effect, I used
direct mailings as a dependent variable. With this approach, the Baron and Kenny’s (1986) proposed mediation test. I reesti-
effect of direct mailings remained significant (p < .05); however, mated Model 2 (Column 3, Tables 4 and 5) in both the cus-
the parameter estimate decreased from .04 to .013. On the basis of tomer retention and the customer share development appli-
these analyses, I conclude that endogeneity of direct mailings does
not affect the hypothesis testing. 6A reviewer suggested this analysis.


cations, but I left out commitment. The parameter estimates Effect of CRPs and RMIs on Customer Retention
for satisfaction and payment equity remain insignificant in and Customer Share Development
both models (customer retention: α = –.10, p > .10; α = .03, The first notable finding of this research is that affective
p > .10; customer share development: β = .01, p > .10; β = commitment is an antecedent of both customer retention and
–.01, p > .10). In addition, I reestimated both models, leav- customer share development. This result is not in line with
ing out satisfaction and payment equity. The parameter esti- recent findings that commitment does not influence cus-
mates for commitment were significant in both models (customer retention (e.g., Gruen, Summers, and Acito 2000).
tomer retention: α = .17, p < .05; customer share However, it confirms previous claims in the relationship
development: β = .02, p < .01). Finally, I estimated a regres- marketing literature that commitment is a significant vari-
sion model in which I related satisfaction and payment able in customer relationships (Morgan and Hunt 1994;
equity to commitment. The parameters of both satisfaction Sheth and Parvatiyar 1995); more precisely, it affects both
and payment equity were positive and significant (γ = .61, relationship maintenance and relationship development. At
p < .01; γ = .09, p <.05). These results show that satisfaction the same time, the absence of an effect of satisfaction and
and payment equity should be considered antecedents of payment equity raises some notable issues. This result con-
affective commitment; however, affective commitment does tradicts previous findings in the literature (e.g., Bowman and
not function as a mediating variable. Narayandas 2001; Szymanski and Henard 2001); several
reasons may explain this. First, prior research has typically
Conclusions relied on survey measures for which self-reported dependent
variables are correlated as a result of common method of
Summary of Findings measures. This study uses behavioral data based (partially)
In this article, I contributed to the marketing literature by on internal company data. Second, unlike prior studies on
studying the effect of CRPs and RMIs on both customer customer share (e.g., Bowman and Narayandas 2001; De
retention and customer share development in a single study. Wulf, Odekerken-Schröder, and Iacobucci 2000) in which
The objectives of this article were twofold. First, I aimed to causality is problematic, this study focuses on the change in
understand the effect of CRPs and RMIs on customer reten- customer share. An understanding of customer share devel-
tion and customer share development. Second, I examined opment may require a deeper understanding of the role of
whether different variables of CRPs and RMIs influence CRPs and RMIs. Third, prior studies focus on customer
customer retention and customer share development. Using share of a single brand in a single product category (Bow-
a longitudinal research design, I related CRPs and RMIs to man and Narayandas 2001), but this study focuses on cus-
actual customer retention and customer share development. tomer share across multiple different services.
An overview of the hypotheses, those that were supported Customer share changes occur over time when cus-
and those that were not supported, is provided in Table 6. tomers add (or drop) new (current) products or services to
For the remainder of this discussion, I focus on the notable (from) their portfolio of purchased products or services at
findings. the focal supplier or at competing suppliers. In this underly-

TABLE 6
Summary of Hypothesis-Testing Results

Customer Retention Customer Share Development

Hypothesis Hypothesis
Antecedents (Sign) Effect Support (Sign) Effect Support

Affective H1 (+) + Yes H2 (+) + Yes
commitment

Satisfaction H3 (+) 0; positively No No effect 0 Yes
moderated by
relationship age

Payment H4 (+) 0 No No effect 0 Yes
equity

Direct No effect N.A. N.A. H5 (+) + Yes
mailings

Loyalty H6a (+) + Yes H6b (+) + Yes
program
Notes: N.A. = not available; this effect could not be estimated because of data limitations.


ing decision process, satisfaction and payment equity play increased relationship age, increased customer shares, and
only a marginal role for several reasons. First, satisfaction purchases of certain additional products or services (e.g., car
and payment equity are based on one’s current experiences insurance, life insurance). Some of these variables positively
with the focal supplier. These experiences do not necessar- affect customer retention and customer share development
ily transfer to other products or services of that supplier: in later stages of the customer relationship.
New events may occur during the relationship that could
change these perceptions (e.g., Mazursky and Geva 1989; Differences Between the Antecedents of
Mittal, Kumar, and Tsiros 1999), thereby limiting the Customer Retention and Customer Share
explanatory power current perceptions. Second, in a com- Development
petitive environment, firms attempt to maximize customer Another research objective was to examine whether the
share. Although customers may be satisfied with the focal antecedents of customer retention and customer share devel-
firm’s offering, they may be equally satisfied with compet- opment are different. Theoretically, there is a clear distinc-
ing offerings from other suppliers. This again limits the tion between relationship maintenance and relationship
explanatory power of satisfaction and payment equity. In development; however, this has not been empirically inves-
contrast, affective commitment seems less vulnerable to new tigated. Unfortunately, a statistical comparison of the coeffi-
experiences in the relationship; it is also unlikely that cus- cients in the customer retention model and customer share
tomers will consider themselves committed to multiple sup- development model is not possible (Franses and Paap 2001,
pliers. Instead of satisfaction and payment equity being Ch. 4). Thus, the only possible comparison is whether the
considered direct antecedents of customer retention and cus- significant predictors are different. The results show that the
tomer share development, they should be considered vari- significant variables (see Table 6) are remarkably consistent
ables that shape commitment (e.g., Morgan and Hunt 1994). across the two models (i.e., affective commitment and loy-
A second notable finding is that RMIs can influence cus- alty programs are significant predictors of both customer
tomer retention and customer share development. Direct retention and customer share development). The only excep-
mailings with a “call to action” are suitable to enhance cus- tion is the interaction effect between satisfaction and rela-
tomer share over time. Loyalty programs that provide eco- tionship age.
nomic rewards are useful both to lengthen customer rela-
However, with consideration of the effect of the past
tionships and to enhance customer share. Bolton, Kannan,
customer behavior control variables, there are some differ-
and Bramlett (2000) report that loyalty programs for credit
ences. For example, whereas high prior customer share has
card customers have a strong, positive effect on customer
a positive effect on customer retention, it has a negative
retention; however, no studies have yet considered the effect
effect on customer share development. Likewise, relation-
of loyalty programs and direct mailings on customer share
ship age has a positive effect on customer retention but no
development. The repeatedly reported positive effect of the
effect on customer share development. The latter results
loyalty program counters the contention of Dowling and
Uncles (1997, p. 75) that “it is difficult to increase brand confirm that different variables affect customer retention
loyalty above the market norms with an easy-to-replicate and customer share development. However, from a CRM
‘add on’ customer loyalty program.” perspective, this difference is not as important as it seems,
The third relevant finding pertains to the explanatory because the same CRM variables affect both customer reten-
power of both CRPs and RMIs. For both customer retention tion and customer share development.
and customer share development, past customer behavior Management Implications
explains the largest part of the variance (CRPs and RMIs are
responsible only for approximately 10% of the total This research provides implications for effective manage-
explained variances in both the customer retention and the ment of customer relationships. First, if managers strive to
customer share development models). This finding seems to affect customer retention, they should focus on creating
support the claims of skeptics of CRM that there is not much committed customers. In addition, a loyalty program with
a firm can do to affect customer loyalty in consumer markets economic incentives leads to greater customer retention.
(Dowling 2002). During reflection on the results of the cus- These results contrast with recent recommendations that
tomer share development model, it might also be perceived creating close ties with customers is a better strategy for
that Ehrenberg’s (1997, p. 19) remarks on the antecedents of enhancing customer loyalty than using economically ori-
market share also hold for the antecedents of customer share ented programs (Braum 2002); firms should do both. Both
development; in particular, his claim “that most markets are affective commitment and economically oriented RMI pro-
near stationary and that everybody has to run hard to stand grams (direct mailings and loyalty programs) enhance cus-
still” might also be applicable to customer share develop- tomer retention and customer share development. Enhanc-
ment. In the short run, my results point to the effect of RMIs ing satisfaction and using attractive pricing policies can also
as only marginal. For example, stopping direct mailings for increase affective commitment. Other Type II RMIs, such as
one year may not necessarily severely harm customer share affinity programs and other socially oriented programs, may
development in that year. In a long-term perspective, the help as well (Rust, Zeithaml, and Lemon 2000). If firms
effects might be different. The effect of both CRPs and strive for immediate results, economically based loyalty
RMIs on customer purchase behavior could result in programs and direct mailings are preferable.


Second, if firms strive to maximize customer share, cre- The last research limitation pertains to the measurement
ating affectively committed customers using a loyalty pro- of payment equity. In this research, I used only two items
gram and sending direct mailings that provide economic (see the Appendix), which could have undermined the relia-
incentives are recommended. However, the short-term posi- bility of the measurement. Further research could develop
tive effects of such approaches are rather small. This might more extensive scales.
support the claim of experts of CRM that trying to maximize
customer retention and customer share development is diffi- Further Research
cult. However, this does not mean that firms should not use Further research should focus on the following issues: First,
such strategies. In the long run, the positive effects of such the results show that the effect of CRPs and RMIs on cus-
strategies may be larger. The short-term small positive tomer retention and customer share is not large. Perhaps
effects of these strategies on customer retention and cus- other variables, such as service calls or sales visits, are
tomer share development could result in larger positive important antecedents. In addition, competing marketing
effects in the long run as a result of the positive effects of variables, such as competitive loyalty programs and direct
past customer behavioral variables, such as relationship age mailings, have not been included here. Further research
and prior customer share. could investigate the effect of these variables. A second
Third, my analysis suggests that, in general, firms can avenue for further research is the effect of RMIs on CRPs
use the same strategies to affect customer retention and cus- and in turn on customer behavior. A simultaneous equation
tomer share development. Fourth, a principle of CRM is to approach, with an appropriate test for mediating effects,
focus efforts on the most loyal customers. However, improv- would be necessary to address this issue. In this respect, the
ing share for loyal customers is much more difficult, interactions between CRPs and RMIs could also be investi-
because they have a greater tendency to reduce their shares gated. Finally, further research could develop decision sup-
in the future. port type models (using data available in customer databases
and data from questionnaires) that would demonstrate the
Research Limitations impact of various CRM strategies.
This study has the following limitations: First, it is con-
ducted for one company in the financial services market. I
chose the financial services market because it is an impor-
Appendix
tant segment of the economy and because there is a long tra- Description of Scales for
dition of customer data storage in this market, which makes Perceptions
it relatively easy to collect behavioral customer loyalty data.
However, the financial services market has some unique Commitment (Cronbach’s Alpha [CA] = .77;
characteristics. Customers purchase insurance products Composite Reliability [CR] = .78)
infrequently, and as a result changes in customer share are I am a loyal customer of XYZ.
not observed as frequently as in other industries. Because of Because I feel a strong attachment to XYZ, I remain a cus-
relatively high switching costs, switching behavior is not tomer of XYZ.
common. These characteristics may have limited the vari- Because I feel a strong sense of belonging with XYZ, I want
ance in the customer share development measure. These to remain a customer of XYZ.
characteristics may also explain some of the results and
may, to some extent, threaten the generalizability of the Satisfaction (CA = .83; CR = .83)
results. Thus, there is a need to extend this study to other How satisfied (1 = “very dissatisfied” and 5 = “very satis-
markets, especially markets in which more switching is fied”) are you about
observed. •the personal attention of XYZ.
Second, although the study applied a longitudinal •the willingness of XYZ to explain procedures.
research design, the causality question remains difficult. •the service quality of XYZ.
Because of the dynamic nature of customer relationships, •the responding by XYZ to claims.
multiple measurements in time (including changes in CRPs) •the expertise of the personnel of XYZ.
are needed in the model. •your relationship with XYZ.
Third, modeling the effect of RMIs is rather difficult, •the alertness of XYZ.
particularly if the RMIs are self-selected or based on cus-
tomers’ purchase behavior. In the loyalty program I studied, Payment Equity (r = .49; CR = .88)
customers can choose whether to become a member. It could How satisfied (1 = “very dissatisfied” and 5 = “very satis-
be argued that customers who expect to purchase new ser- fied”) are you about the insurance premium?
vices are more inclined to join. I chose not to correct for this Do you think the insurance premium of your insurance is
in the analysis at this time. Further research could develop too high, high, normal, low, or too low?
models to correct for possible endogeneity of the RMIs.


REFERENCES
Anderson, Erin and Barton Weitz (1992), “The Use of Pledges to Behavior,” Journal of Marketing Research, 38 (August),
Build and Sustain Commitment in Distribution Channels,” 281–97.
Journal of Marketing Research, 29 (February), 18–34. Braum, Lutz (2002), “Involved Customers Lead to Loyalty Gains,”
Anderson, Eugene W., Claes Fornell, and Donald R. Lehmann Marketing News, 36 (2), 16.
(1994), “Customer Satisfaction, Market Share, and Profitabil- Capraro, Anthony J., Susan Broniarczyk, and Rajendra K. Sriva-
ity: Findings from Sweden,” Journal of Marketing, 58 (July), stava (2003), “Factors Influencing the Likelihood of Customer
53–66. Defection: The Role of Consumer Knowledge,” Journal of the
——— and Vikas Mittal (2000), “Strengthening the Satisfaction- Academy of Marketing Science, 31 (Spring), 164–75.
Profit Chain,” Journal of Service Research, 3 (2), 107–120. Christy, Richard, Gordon Oliver, and Joe Penn (1996), “Relation-
——— and Mary W. Sullivan (1993), “The Antecedents and Con- ship Marketing in Consumer Markets,” Journal of Marketing
sequences of Customer Satisfaction for Firms,” Marketing Sci- Management, 12 (1), 175–88.
ence, 12 (Spring), 125–43. Crosby, Lawrence A., Kenneth R. Evans, and Deborah Cowles
Anderson, James C. and David W. Gerbing (1988), “Structural (1990), “Relationship Quality in Services Selling: An Interper-
Equation Modeling Practice: A Review and Recommended sonal Influence Perspective,” Journal of Marketing, 54 (July),
Two-Step Approach,” Psychological Bulletin, 103 (3), 411–23. 68–81.
Bagozzi, Richard P. and Youjae Yi (1988), “On the Evaluation of Davidson, Russell and James G. MacKinnon (1989), “Testing for
Structural Equation Models,” Journal of the Academy of Mar- Consistency Using Artificial Regressions,” Econometric The-
keting Science, 16 (Winter), 74–94. ory, 5 (3), 363–84.
Baron, Reuben M. and David A. Kenny (1986), “The Moderator- De Wulf, Kristof, Gaby Odekerken-Schröder, and Dawn Iacobucci
Mediator Variable Distinction in Social Psychological (2001), “Investments in Consumer Relationships: A Cross-
Research,” Journal of Personality and Social Psychology, 51 Country and Cross-Industry Exploration,” Journal of Market-
(6), 1173–82. ing, 65 (October), 33–50.
Baumgartner, Hans and Christian Homburg (1996), “Applications Dick, Alan S. and Kunal Basu (1994), “Customer Loyalty: Toward
of Structural Equation Modeling in Marketing and Consumer an Integrated Conceptual Framework,” Journal of the Academy
Research,” International Journal of Research in Marketing, 13 of Marketing Science, 22 (Spring), 99–113.
(2), 139–61. Dowling, Grahame W. (2002), “Customer Relationship Manage-
Bawa, Kapil and Robert W. Shoemaker (1987), “The Effects of a ment: In B2C Markets, Often Less Is More,” California Man-
Direct Mail Coupon on Brand Choice Behavior,” Journal of agement Review, 44 (Spring), 87–104.
Marketing Research, 24 (November), 370–76. ——— and Mark Uncles (1997), “Do Customer Loyalty Programs
Berry, Leonard L. (1995), “Relationship Marketing of Services: Really Work?” Sloan Management Review, 38 (Fall), 71–82.
Growing Interest, Emerging Perspectives,” Journal of the Acad- Dwyer, Robert F., Paul H. Schurr, and Sejo Oh (1987), “Develop-
emy of Marketing Science, 23 (Fall), 236–45. ing Buyer–Seller Relationships,” Journal of Marketing, 51
Bhattacharya, C.B. and Ruth N. Bolton (2000), “Relationship Mar- (April), 11–27.
keting in Mass Markets,” in Handbook of Relationship Market- Ehrenberg, Andrew (1997), “Description and Prescription,” Jour-
ing, Jagdish N. Sheth and Atul Parvatiyar, eds. Thousand Oaks, nal of Advertising Research, 37 (6), 17–22.
CA: Sage Publications, 327–54. Franses, Philip Hans and Richard Paap (2001), Quantitative Mod-
———, Hayagreeva Rao, and Mary Ann Glynn (1995), “Under- els in Marketing Research. Cambridge, UK: Cambridge Uni-
standing the Bond of Identification: An Investigation of Its Cor- versity Press.
relates Among Art Museum Members,” Journal of Marketing, Garbarino, Ellen and Mark S. Johnson (1999), “The Different
59 (October), 46–57. Roles of Satisfaction, Trust, and Commitment in Customer
Bickart, Barbara A. (1993), “Carryover and Backfire Effects in Relationships,” Journal of Marketing, 63 (April), 70–87.
Marketing Research,” Journal of Marketing Research, 30 (Feb- Geyskens, Inge, Jan-Benedict E.M. Steenkamp, and Nirmalya
ruary), 52–62. Kumar (1998), “Generalizations About Trust in Marketing
Blattberg, Robert C., Gary Getz, and Jacquelyn S. Thomas (2001), Channel Relationships Using Meta-Analysis,” International
Customer Equity: Building and Managing Relationships as Journal of Research in Marketing, 15 (3), 223–48.
Valuable Assets. Boston: Harvard Business School Press. Gruen, Thomas W., John O. Summers, and Frank Acito (2000),
Bolton, Ruth N. (1998), “A Dynamic Model of the Duration of the “Relationship Marketing Activities, Commitment, and Mem-
Customer’s Relationship with a Continuous Service Provider: bership Behaviors in Professional Associations,” Journal of
The Role of Satisfaction,” Marketing Science, 17 (Winter), Marketing, 64 (July), 34–49.
45–65. Gundlach, Gregory T., Ravi Achrol, and John T. Mentzer (1995),
———, P.K. Kannan, and Mathew D. Bramlett (2000), “Implica- “The Structure of Commitment in Exchange,” Journal of Mar-
tions of Loyalty Program Membership and Service Experiences keting, 59 (January), 78–92.
for Customer Retention and Value,” Journal of the Academy of Hart, Susan, Andrew Smith, Leigh Sparks, and Nikos Tzokas
Marketing Science, 28 (Winter), 95–108. (1999), “Are Loyalty Schemes a Manifestation of Relationship
——— and Katherine N. Lemon (1999), “A Dynamic Model of Marketing?” Journal of Marketing Management, 15 (6),
Customers’ Usage of Services: Usage as an Antecedent and 541–62.
Consequence of Satisfaction,” Journal of Marketing Research, Heckman, James J. (1976), “The Common Structure of Statistical
36 (May), 171–86. Models of Truncation, Sample Selection, and Limited Depen-
———, ———, and Peter C. Verhoef (2002), “The Theoretical dent Variables and a Simple Estimator for Such Models,”
Underpinnings of Customer Asset Management: A Framework Annals of Economic and Social Measurement, 5, 475–92.
and Propositions for Future Research,” Erasmus Research Insti- Hoekstra, Janny C., Peter S.H. Leeflang, and Dick R. Wittink
tute in Management Working Paper No. ERS-2002-80-MKT, (1999), “The Customer Concept: The Basis for a New Market-
Erasmus University, Rotterdam. ing Paradigm,” Journal of Market-Focused Management, 4 (1),
Bowman, Douglas and Das Narayandas (2001), “Managing 43–76.
Customer-Initiated Contacts with Manufacturers: The Impact Homburg, Christian and Annette Giering (2001), “Personal Char-
on Share of Category Requirements and Word-of-Mouth acteristics as Moderators of the Relationship Between Cus-


tomer Satisfaction and Loyalty—An Empirical Analysis,” Psy- and Propositions,” Journal of Economic Psychology, 8 (4),
chology and Marketing, 18 (1), 43–66. 469–99.
Jones, Michael A., David L. Mothersbaugh, and Sharon E. Beatty Peppers, Don and Martha Rogers (1999), Enterprise One-to-One:
(2001), “Switching Barriers and Repurchase Intentions in Ser- Tools for Competing in the Interactive Age. New York:
vices,” Journal of Retailing, 76 (2), 259–74. Doubleday.
Jones, Thomas O. and W. Earl Sasser Jr. (1995), “Why Satisfied Pindyck, Robert S. and Daniel L. Rubinfeld (1998), Econometric
Customers Defect,” Harvard Business Review, 73 Models and Economic Forecasts. Boston: Irwin/McGraw-Hill.
(November–December), 88–100. Reichheld, Frederick F. (1996), The Loyalty Effect. Boston: Har-
Jöreskög, Karl and Dag Sörbom (1993), LISREL 83 Structural vard Business School Press.
Equation Modeling with the SIMPLIS Command Language. Reinartz, Werner J. and V. Kumar (2002), “The Mismanagement of
Chicago: Scientific Software International. Customer Loyalty,” Harvard Business Review, 80 (July),
Kamakura, Wagner A., Sridhar N. Ramaswami, and Rajendra K. 86–94.
Srivastava (1991), “Applying Latent Trait Analysis in the Eval- ——— and ——— (2003), “The Impact of Customer Relationship
uation of Prospects for Cross Selling of Financial Services,” Characteristics on Profitable Lifetime Duration,” Journal of
International Journal of Research in Marketing, 8 (4), 329–49. Marketing, 67 (January), 77–99.
Kerstetter, Jim (2001), “Software Highfliers,” BusinessWeek, (June Roberts, Mary Lou and Paul D. Berger (1999), Direct Marketing
18), 108–109. Management. Englewood Cliffs, NJ: Prentice Hall.
Klemperer, Paul (1995), “Competition When Consumers Have Roehm, Michelle L., Ellen Bolman Pullins, and Harper A. Roehm
Switching Costs: An Overview with Applications to Industrial Jr. (2002), “Designing Loyalty-Building Programs for Pack-
Organization, Macro Economics, and International Trade,” The aged Goods Brands,” Journal of Marketing Research, 39
Review of Economic Studies, 62 (4), 515–39. (May), 202–213.
Kumar, Nirmalya, Lisa K. Scheer, and Jan-Benedict E.M. Rust, Roland T., Valarie A. Zeithaml, and Katherine N. Lemon
Steenkamp (1995), “The Effects of Supplier Fairness on Vul- (2000), Driving Customer Equity: How Customer Lifetime
nerable Resellers,” Journal of Marketing Research, 32 (Febru- Value Is Reshaping Corporate Strategy. New York: The Free
ary), 54–65. Press.
Latham, Garry P. and Edwin A. Locke (1991), “Self Regulation Sharp, Byron and Anne Sharp (1997), “Loyalty Programs and
Through Goal Setting,” Organizational Behavior and Human Their Impact on Repeat-Purchase Loyalty Patterns,” Interna-
Decision Processes, 50 (2), 212–47. tional Journal of Research in Marketing, 14 (5), 473–86.
Lemon, Katherine N., Tiffany Barnett White, and Russell S. Winer Sheth, Jagdish N. and Atul Parvatiyar (1995), “Relationship Mar-
(2002), “Dynamic Customer Relationship Management: Incor- keting in Consumer Markets: Antecedents and Consequences,”
porating Future Considerations into the Service Retention Deci- Journal of the Academy of Marketing Science, 23 (Fall),
sion,” Journal of Marketing, 66 (January), 1–14. 255–71.
Lichtenstein, Donald R., Nancy M. Ridgway, and Richard G. Nete- Singh, Jagdip (1990), “Voice, Exit, and Negative Word-of-Mouth
meyer (1993), “Price Perceptions and Consumer Shopping Behaviors: An Investigation Across Three Service Categories,”
Behavior: A Field Study,” Journal of Marketing Research, 30 Journal of the Academy of Marketing Science, 18 (Winter),
(May), 234–45. 1–15.
Macintosh, Gerrard and Lawrence S. Lockshin (1997), “Retail Steenkamp, Jan-Benedict E.M. and Hans C.M. van Trijp (1991),
Relationships and Store Loyalty: A Multi-Level Perspective,” “The Use of LISREL in Validating Marketing Constructs,”
International Journal of Research in Marketing, 14 (5), International Journal of Research in Marketing, 8 (4), 283–99.
487–97. Szymanski, David M. and David H. Henard (2001), “Customer
MacKenzie, Scott B., Philip M. Podsakoff, and Michael J. Ahearne Satisfaction: A Meta-Analysis of the Empirical Evidence,”
(1998), “Some Possible Antecedents and Consequences of In- Journal of the Academy of Marketing Science, 29 (Winter),
Role and Extra-Role Salesperson Performance,” Journal of 16–35.
Marketing, 62 (July), 87–98. Varuki, Sajeev and Mark Colgate (2001), “The Role of Price Per-
Mazursky, David and Aviva Geva (1989), “Temporal Decay in Sat- ceptions in an Integrated Model of Behavioral Intentions,”
isfaction–Purchase Intention Relationship,” Psychology and Journal of Service Research, 3 (3), 232–40.
Marketing, 6 (3), 211–27. Verhoef, Peter C., Philip Hans Franses, and Janny C. Hoekstra
Mittal, Vikas and Wagner A. Kamakura (2001), “Satisfaction, (2001), “The Impact of Satisfaction and Payment Equity on
Repurchase Intent, and Repurchase Behavior: Investigating the Cross Buying: A Dynamic Model for a Multi-Service Provider,”
Moderating Effect of Customer Characteristics,” Journal of Journal of Retailing, 77 (3), 359–78.
Marketing Research, 38 (February), 131–42. ———, Penny N. Spring, Janny C. Hoekstra, and Peter S.H.
———, Pankaj Kumar, and Michael Tsiros (1999), “Attribute- Leeflang (2003), “The Commercial Use of Segmentation and
Level Performance, Satisfaction, and Behavioral Intentions Predictive Modeling Techniques for Database Marketing in the
over Time: A Consumption-System Approach,” Journal of Mar- Netherlands,” Decision Support Systems, 34 (4), 471–81.
keting, 63 (April), 88–101. White, Halbert (1980), “A Heteroskedasticity-Consistent Covari-
Moorman, Christine, Gerald Zaltman, and Rohit Deshpandé ance Matrix and a Direct Test for Heteroskedasticity,” Econo-
(1992), “Relationships Between Providers and Users of Mar- metrica, 48 (4), 817–38.
keting Research: The Dynamics of Trust Within and Between Winer, Russell S. (2001), “A Framework for Customer Relation-
Organizations,” Journal of Marketing Research, 29 (August), ship Management,” California Management Review, 43 (Sum-
314–28. mer), 89–108.
Morgan, Robert M. and Shelby D. Hunt (1994), “The Woodruff, Robert B. (1997), “Customer Value: The Next Source
Commitment–Trust Theory of Relationship Marketing,” Jour- for Competitive Advantage,” Journal of the Academy of Mar-
nal of Marketing, 58 (July), 20–38. keting Science, 25 (Spring), 139–53.
Nicholls, John G. (1989), The Competitive Ethos and Democratic Zeithaml, Valarie A., Leonard L. Berry, and A. Parasuraman
Education. Cambridge, MA: Harvard University Press. (1996), “The Behavioral Consequences of Service Quality,”
Oliver, Rick and Russel S. Winer (1987), “A Framework for the Journal of Marketing, 60 (April), 31–46.
Formation and Structure of Consumer Expectations: Review


Understanding the effect of customer relationship management effors on customer retention and customer share development

Understanding the effect of customer relationship management effors on customer retention and customer share development

More Related Content

What's hot (19)

Similar to Understanding the effect of customer relationship management effors on customer retention and customer share development (20)

More from Wesley Pinheiro (14)

Recently uploaded (20)

Understanding the effect of customer relationship management effors on customer retention and customer share development