Marketing service relationships:

the role of commitment

Martin Wetzels
Assistant Professor of Marketing and Marketing Research, Faculty of
Economics and Business Administration, Maastricht University,
Maastricht, The Netherlands
Ko de Ruyter
Associate Professor of Marketing and Marketing Research, Faculty of
Economics and Business Administration, Maastricht University,
Maastricht, The Netherlands
Marcel van Birgelen
Faculty of Economics and Business Administration, Maastricht
University, Masstricht, The Netherlands

Introduction
As competition intensifies, products and services become more
homogeneous, and markets become mature, it is becoming increasingly
harder for companies in both manufacturing and service industries to
differentiate themselves from other organizations. Merely providing
customers with technical solutions to problems does not suffice anymore to
be competitive and gain and retain market share. Various value-added
services, which start before the actual transaction and go far beyond it, have
to be delivered in order to stay competitive and create customer loyalty.
Research and business practices have shown that keeping customers through
value-added services costs less than acquiring new ones (Anderson and
Fornell, 1994; Reichheld and Sasser, 1990). In addition, loyal customers will
generate more revenues in the long run. Therefore, the concept of
relationship marketing was introduced in the (services) marketing literature
and incorporated in models of the service delivery processes.
Central role of In the relationship marketing literature the concept of commitment plays a
commitment central role, as it is a major characteristic of relationship marketing models
(Scanzoni, 1979). Commitment refers to an implicit or explicit pledge of the
continuity of a relationship between exchange partners. This paper should be
viewed as a preliminary attempt to integrate concepts from the service
quality/satisfaction area with concepts from the relationship marketing area.
Service quality and satisfaction will be treated as antecedents of
commitment in addition to the concepts of trust and dependence. Moreover,
the consequences of commitment will also be examined. A model which
incorporates antecedents and consequences of commitment in a
comprehensive framework will be tested.
Our paper is structured as follows. First, we will discuss the concept of
relationship marketing in services and pay attention to the relationship life
cycle, a process through which relationships develop. Subsequently, the
concept of commitment in business relationships will be discussed in more
detail, because of its importance in relationship marketing. Several
antecedents and consequences of commitment will be identified and
discussed and hypotheses regarding these constructs will be formulated.
These hypotheses will be graphically summarized in a conceptual model in

We would like to thank Klaas Fleischmann for his efforts during data collection for
this study.

406 JOURNAL OF BUSINESS & INDUSTRIAL MARKETING, VOL. 13 NO. 4/5 1998, pp. 406-423 © MCB UNIVERSITY PRESS, 0885-8624
 which service quality, satisfaction, commitment and other related concepts
will be incorporated. Next, we will discuss the results of a study designed to
investigate the applicability of this conceptual model. This study was
conducted in a business-to-business setting, in which the relationships of
business customers with service engineers of a major Dutch office equipment
manufacturer were being studied. We conclude the paper with a discussion of
our results and of a number of theoretical and managerial implications
following from these results.

Relationship marketing in services

Close relationships with The service business originally evolved around keeping close relationships
customers with customers. However, because of the increasing number of customers
and the stressing of the marketing mix paradigm, the relationship with
customers moved to the background during the 1960s and the 1970s
(Grönroos, 1995). It was not until the early 1980s that relationship marketing
reemerged again in service practices (Berry, 1995). This was due to the
maturing of service marketing with an emphasis on quality, increased
recognition of potential benefits for companies and the customer, and
technological advances. Another factor was that the marketing mix paradigm
did not fit service company’s customer relations very well, which led to the
development of service marketing as a discipline in its own right (Grönroos,
1990).
How can relationships in services marketing be conceptualized? Berry
(1983, p. 25) defined relationship marketing as “attracting, maintaining and
– in multi-service organizations – enhancing customer relationships”.
Morgan and Hunt (1994) propose that relationship marketing refers to all
marketing activities directed towards establishing, developing and
maintaining successful relational exchanges. They argue that in fact there are
no buyers and sellers or providers and customers in relational exchanges but
partners exchanging resources. Relational exchange differs from discrete
transaction on several dimensions (Dwyer et al., 1987). First, relational
exchanges may decrease over time as this may be beneficial to one or both
partners. Second, relational exchange partners can be expected to derive
complex, personal, non-economic satisfactions and engage in social
exchange. Third, because obligations and activities may be relatively
complex and take place during an extended time period, partners may
engage in carefully planning and controlling items of exchange. Finally,
third parties may be called in to arbitrate and additional mechanisms for
collaboration and resolving conflict may have to be designed. Relational
exchange can provide a competitive advantage to the extent that it
contributes to product and/or service differentiation and creates barriers to
switching (Dwyer et al., 1987).
Five phases Relationships between partners do not just emerge or exist, they evolve
through a process over time. This process is characterized by five general
phases (Scanzoni, 1979). The first phase in this relationship life-cycle is
called awareness and refers to the recognition that some second partner is a
feasible exchange partner. There is no interaction between the parties, they
only try to position themselves in the best possible way increase the
attraction toward other companies. After awareness comes exploration,
which is the search and trial phase in relational exchange. In this phase
potential exchange partners first consider obligations, benefits and burdens
and, the possibility of exchange. The third phase is called expansion, which
refers to the continual increase in benefits obtained by exchange partners and

JOURNAL OF BUSINESS & INDUSTRIAL MARKETING, VOL. 13 NO. 4/5 1998 407
 their increasing interdependence. The main distinction between this phase
and the previous phase of exploration is that now the partners trust each
other to a greater extent and are satisfied with the relationship which leads to
increased risk taking in the dyad. If relationships develop then commitment
is formed between the two business partners. Commitment is the most
desirable aspect of relationships and refers to an implicit or explicit pledge
of relational continuity between exchange partners (Dwyer et al., 1987).
Now the exchange partners have achieved a level of satisfaction from the
exchange process that virtually precludes other primary exchange partners
who could provide similar benefits. Commitment is established on the basis
of two important assumptions. The first is that partners provide relatively
high levels of inputs to the association. The second assumption pertains to
durability, i.e. the common belief of effectiveness of future exchange, also
referred to as continuation.
Dissolution The final phase of the relationship development process is dissolution. This
phase begins with an intrapsychic stage in which one partner privately
evaluates his or her dissatisfaction with the other party, concluding that the
costs of modification or continuation outweigh the disadvantages of the
relationship. Afterwards, the relationship enters an interactive phase in
which the parties negotiate their unbonding. Dissolution is then presented
publicly.
It should be clear that of these five phases, the fourth phase of commitment,
is the most desirable one in the development of an ongoing relationship
between a buyer and a seller. If the parties are not committed to their
relationship, it will end at a rapid pace. Therefore, the concept of
commitment should play a central role in the development of a model that
introduces relational exchanges to the marketing of services. However, as a
result of the paradigm shift toward relationships, researchers are facing the
challenge of conceptualizing and operationalizing heterogeneous,
overlapping and equivocal constructs. With respect to the central concept of
commitment, Morrow (1983, p. 486) accentuates this challenge by stating
that “the growth of commitment-related concepts has not been accompanied
by a careful segmentation of commitment’s theoretical domain”. Therefore,
we shall direct our attention to the conceptualization of commitment in the
next section.

Conceptualization of commitment
Commitment in more detail
Benefits As mentioned before, commitment plays a central role in relationships. As
Scanzoni (1979) stated “commitment is the most advanced phase of
partners’ interdependence”. In marketing-practice and research it is agreed
that mutual commitment among partners in business relationships produces
significant benefits for companies. Manufacturers can obtain improved
product developments, increased margins and market shares. Distributors
gain deeper market penetration and higher customer satisfaction. The
conceptualization of commitment stems from industrial/organizational
psychology and has been viewed as an intention to continue a course of
action or activity, such as maintaining a relationship with a business partner
(Fehr, 1988). As mentioned in the introduction, commitment in the buyer-
seller relationship literature is defined as an implicit or explicit pledge of
relational continuity between exchange partners (Dwyer et al., 1987). Parties
identify commitment among exchange partners as key to achieving valuable
outcomes for themselves, and they endeavor to develop and maintain this

408 JOURNAL OF BUSINESS & INDUSTRIAL MARKETING, VOL. 13 NO. 4/5 1998
 precious attribute in their relationship (Morgan and Hunt, 1994).
Commitment is seen as a sentiment that is critically important in the
development of long-term channel relationships or as a favorable affective
reaction (Kumar et al., 1995). Therefore, commitment is a psychological
sentiment of the mind through which an attitude concerning continuation of
a relationship with a business partner is formed.
Commitment is an In the literature two views of organizational commitment have dominated.
affective state of mind According to one view, commitment is an affective state of mind an
individual or partner has toward a relationship with another individual or
partner. This kind of commitment is called affective commitment. Affective
commitment is brought about by a person sharing, identifying with, or
internalizing the values of the organization (Morgan and Hunt, 1994).
Affective commitment is based on a sense of liking and emotional
attachment to the partnership. The other view sees commitment as being
more behavioral than affective. This form is referred to as calculative
commitment and stems from a cognitive evaluation of the instrumental
worth of a continued relationship with the organization. All gains and losses,
plusses and minuses or rewards and punishments are added up (Morgan and
Hunt, 1994). Geyskens et al. (1996, p. 304) define commitment as the
perceived need to “maintain a relationship given the significant anticipated
termination or switching costs associated with leaving”. This view stems
from the side-bet theory of Becker (1960), which links extraneous interests
with a consistent line of activity. Calculative commitment is based on inputs
like investments and allocation of resources specifically for the relationship
between two business partners (Williamson, 1985). After these inputs have
been made, they may reduce the number of alternative alliances a company
may have with other partners, or lead to redirecting costs (Williamson, 1983,
1984). Williamson (1975, 1979) provides an economic argument for
cooperative interorganizational relationships between firms on the basis of
transaction cost analysis. On the one hand there is the internal cost of
carrying out an activity within the firm, while on the other hand there is the
cost of an external supplier or potential partner. Transaction cost then
represents the cost of establishing a relationship as well as costs associated
with balancing dependence in the relationship and prohibiting opportunistic
behavior by the exchange partner. Calculative commitment holds when
internal costs are less than the external costs.
Affective commitment Research suggests that of these two forms of commitment affective
commitment is the most effective for developing and maintaining mutually
beneficial relationships between partners (Kumar et al., 1994). Affective
commitment has strong positive influences on:
(1) intention to stay in a relationship;
(2) desire to stay in a relationship;
(3) performance; and
(4) willingness to invest in a relationship.
Furthermore, it was found that it has strong negative influences on:
(5) development of alternatives for a relationship; and
(6) opportunistic behavior.
Calculative commitment in contrast has positive influences on (5) develop-
ment of alternatives, and (6) opportunism. Therefore, this latter form of
commitment ultimately has a negative impact on relationships.

JOURNAL OF BUSINESS & INDUSTRIAL MARKETING, VOL. 13 NO. 4/5 1998 409
 Antecedents of commitment
Owing to the importance of commitment in the development of
relationships, information on the factors that influence commitment is
essential. First, service quality and satisfaction, the two key constructs in the
services marketing area, will be related to commitment. Subsequently, two
additional factors will be related to commitment: trust and dependence.
Various studies in the relationship marketing area have shown that these two
factors seem to be crucial in influencing commitment (Anderson and Narus,
1990; Anderson and Weitz, 1989; Ganasan, 1994; Geyskens and Steenkamp,
1995; Kumar et al., 1995; Morgan and Hunt, 1994). As we discuss these
constructs, hypotheses for our study will be formulated.

Service quality and satisfaction

Crucial importance of During the past decade substantial research has been devoted to the quality
service quality of services and the measurement of service quality. Similarly, in marketing
practice, service quality has been recognized as being of crucial importance.
Satisfaction has been defined as a consumer’s evaluative judgment related to
the pleasurable level of consumption-related fulfillment (Oliver, 1996).
Similary, perceived service quality has been conceptualized as the
comparison of service expectations with actual performance expectations.
Especially in recent years, the relationship between service quality and
customer satisfaction has received a lot of attention. A focal point of
discussion is the causal ordering between service quality and satisfaction.
Also there is considerable debate about the conceptual distinctness of these
two concepts. The predominant view is that quality influences satisfaction.
“Quality is the logical predecessor to satisfaction” (Iacobucci et al., 1996).
Cronin and Taylor (1992) find empirical support that “service quality is an
antecedent of consumer satisfaction”.
According to Grönroos (1990), quality can be classified into two categories.
First, there is technical quality which relates to the quality of the result or
outcome of the service, i.e. what the customer is actually receiving from the
service. Second, functional quality relates to the manner in or process by
which a service is provided, i.e. how the service is delivered.
Positive relationships In literature, there is no conceptual or empirical evidence for the relationship
between service quality, satisfaction and commitment. However, as
commitment can be viewed and indeed has been operationalized as a proxy
for loyalty (Assael (1987) defines loyalty as commitment toward a certain
brand) and because strong positive associations between service quality,
service satisfaction and loyalty have been reported extensively in the
services marketing area, we posit that there will be positive relationships
between technical and functional service quality, satisfaction and affective
and calculative commitment:
H1: There will be a positive relation between satisfaction and affective
commitment.
H2: There will be a positive relation between satisfaction and calculative
commitment.
H3: There will be a positive relation between technical quality and
affective commitment.
H4: There will be a positive relation between functional quality and
affective commitment.
H5: There will be a positive relation between technical quality and
calculative commitment.

410 JOURNAL OF BUSINESS & INDUSTRIAL MARKETING, VOL. 13 NO. 4/5 1998
 H6: There will be a positive relation between functional quality and
calculative commitment.
H7: There will be a positive relation between technical quality and
satisfaction.
H8: There will be a positive relation between functional quality and
satisfaction.

Trust
Reliability is key to trust Trust is defined as one party’s belief that its needs will be fulfilled in the
future by actions undertaken by the other party (Anderson and Weitz, 1989).
Trust exists when one party has confidence in the reliability and integrity of
the exchange partner (Morgan and Hunt, 1994). In social psychology a
consensus emerges that trust consists of two essential elements, trust in the
partner’s honesty and trust in the partner’s benevolence. Honesty is the
belief that one’s partner stands by its word. Benevolence is the belief that
one’s partner is interested in the company’s welfare and will not take
unexpected actions which will negatively impact the company (Geyskens
and Steenkamp, 1995). When a distinction is made between affective and
calculative commitment, it was found that trust, consisting of honesty as
well as benevolence, has a positive influence on affective commitment
(Anderson and Narus, 1990; Anderson and Weitz, 1989; Morgan and Hunt,
1994; Geyskens and Steenkamp, 1995). Calculative commitment was
negatively influenced by trust (Geyskens and Steenkamp, 1995). It seems
that if partners in a relationship trust each other more they are more
emotionally involved and less consciously weighing the benefits against the
costs of that relationship. Commitment in this sense is conceptualized as the
amount of effort put in the process of calculating costs and benefits of a
relationship. Therefore, we expect that both types of trust will be positively
related to affective commitment and negatively to calculative commitment:
H9: There will be a positive relation between trust (benevolence) and
affective commitment.
H10: There will be a negative relation between trust (benevolence) and
calculative commitment.
H11: There will be a positive relation between trust (honesty) and affective
commitment.
H12: There will be a negative relation between trust (honesty) and
calculative commitment.

Dependence
Dependence to achieve Dependence is referred to as asymmetry in a relationship between two
goals? partners. A company’s dependence on a partner traditionally has been defined
in channels as the company’s need to maintain a relationship with the partner
to achieve its goals (Kumar et al., 1995). In other words, the higher the
dependence between two parties, the more one party needs the other party to
benefit from the relationship. It was found that dependence in relationships
has a negative impact on affective commitment (Anderson and Weitz, 1989;
Kumar et al., 1995) and a positive influence on calculative commitment
(Ganasan, 1994; Geyskens and Steenkamp, 1995). If parties feel that they are
more dependent on some other party they are less emotionally involved in the
relationship and more in comparing the costs with the benefits. It has been
argued that dependence and the potential consequent fear of exploitation,
reduces the motivation to continue the relationship on affective grounds

JOURNAL OF BUSINESS & INDUSTRIAL MARKETING, VOL. 13 NO. 4/5 1998 411
 (Anderson and Weitz, 1989). If a potential relationship with another party
seems more favorable in terms of more equality between the partners, then
the current relationship is easily terminated. A different picture emerges for
calculative commitment. As dependence increases, the dependent partner will
find itself in an increasingly vulnerable position. Hence, it could be argued
that the dependent firm is more likely to continue the relationship because it
seems necessary given the switching costs involved in terminating the
partnership. Alternatively, increased dependence will force a partner to weigh
the benefits more heavily. Therefore, we expect that higher dependence
between partners will lead to less affective commitment and more calculative
commitment:
H13: There will be a negative relation between dependence and affective
commitment.
H14: There will be a positive relation between dependence and calculative
commitment.

Consequences of commitment
Continuance of Ultimately, any firm interested in exchange relationships with its customers
commitment would have an interest in the continuance of commitment. This has been
called the temporal dimension of commitment (Gundlach et al., 1995). The
manifestation of this temporal dimension can be found in partners’
intentions to continue the relationship in the future. In modeling terms,
intention to stay would be the variable that one would like to be able to
explain. We expect positive relationships between the two types of
commitment and the intention to stay in the relationship. Furthermore, we
expect that customer satisfaction is positively related to the intention to stay
in the relationship:
H15: There will be a positive relation between satisfaction and intention to
stay in the relationship.
H16: There will be a positive relation between affective commitment and
intention to stay in the relationship.
H17: There will be a positive relation between calculative commitment and
intention to stay in the relationship.
In the next section a framework for relationship marketing in the services
sector will be developed.

Development of a conceptual framework

A conceptual framework To study relational commitment and its relationship with various antecedents
and consequences, it is necessary to develop a conceptual framework. The
theoretical as well as managerial relevance of an investigation of relationship
marketing, in which commitment plays a central role, is based on the
identification of a number of antecedents of this type of marketing. These
antecedents were discussed in the previous sections of this paper. With
affective commitment as primary focal construct and calculative commitment
as secondary focal construct five independent antecedents are identified:
(1) technical quality;
(2) functional quality;
(3) trust (benevolence);
(4) trust (honesty); and
(5) “dependence”.

412 JOURNAL OF BUSINESS & INDUSTRIAL MARKETING, VOL. 13 NO. 4/5 1998
 Satisfaction is dependent on service quality but also influences affective and
calculative commitment. Therefore, this concept also will be a direct
antecedent of commitment. Finally, we expect that a relationship with high
levels of satisfaction and commitment will lead to a high level of loyalty of
the customer toward the selling organization. This behavioral consequence,
an intention to stay in the relationship, will therefore be the outcome of our
framework of relationship marketing. The hypotheses we formulated can be
summarized graphically by the following conceptual framework, rendered in
Figure 1.
Tested in three phases The conceptual model of the service delivery process, as shown in Figure 1,
is tested empirically in three phases. The first phase relates the concepts of
perceived service quality (divided into technical and functional quality),
satisfaction and affective and calculative commitment with each other. In
this phase hypotheses 1 through 8 will be tested. The second phase tests the
relationships between commitment, trust and dependence. This phase
involves the testing of hypotheses 9 through 14. Finally, the behavioral
consequence of the service delivery process is investigated. This will be
done by testing hypotheses 15 through 17. In the next section we will report
on a test of aforementioned hypotheses.

An empirical study
Research setting and sampling
Mail survey The relationships that were at central focus in this study were the
relationships between a major Dutch office equipment manufacturer and
its industrial customers. The hypotheses were tested empirically using a
stratified random sample of customers of this office equipment
manufacturer. A mail questionnaire was designed that consisted of
multiple items that formed an operationalization of our conceptual
framework. A total of 1,988 questionnaires, accompanied with
personalized cover letters, were mailed. Addresses were randomly selected
from the customer database of the office equipment manufacturer. In total
572 usable questionnaires were received from customers, which is a
response rate of 28.8 percent. Customers who were not willing to
participate in the mail survey were asked to answer an abbreviated

Technical
Quality
+
+ +
Satisfaction

Functional +
Quality
+ +
+

Trust + +
Benevolence +
– Affective
Commitment +
Trust + Intention
Honesty – to Stay
Calculative +
– Commitment
+
Dependence

Figure 1. Conceptual model

JOURNAL OF BUSINESS & INDUSTRIAL MARKETING, VOL. 13 NO. 4/5 1998 413
 A time trend telephone survey in order to obtain insight into non-response bias. No
extrapolation test significant differences between respondents and non-respondents regarding
characteristics and attitudes were discovered. Furthermore, a time trend
extrapolation test was carried out. The assumption of such a test is that
respondents who respond less readily, are more like non-respondents
(Armstrong and Overton, 1977). Only a negligible number of variables
(less than 5 percent) used in the questionnaire showed a significant
difference between early and late respondents. However, this number of
variables was distributed among all items of the questionnaire and no
consistent pattern could be discerned. Therefore, we may conclude that our
data set did not exhibit response-related problems. In comparison with the
internal customer database, our sample could be considered representative
of the total population. Customers from various industries were included in
the sample: educational institutions (31 percent), services in general (22
percent), reprographic industries (16 percent), financial services (10
percent), trade (8 percent), health care (5 percent), manufacturing
industries in general (5 percent) and governmental institutions (3 percent).
Our respondents had the following positions within their companies:
manager reprographic unit (22 percent), facility manager (22 percent),
employee reprographic unit (20 percent), buying manager (11 percent),
general manager (9 percent) and various (15 percent). Most relationships
with customers have been established between five and ten years (46
percent of the respondents).

Questionnaire development
Established scales To measure the various constructs in the conceptual model of the service
delivery process established scales were used. All items were measured on
nine-point Likert-type scales, except for the measurement of satisfaction
with the manufacturer’s service, for which we used a ten-point scale. All
items were “double-back” translated (Brislin, 1980) into Dutch by a
qualified translator. Affective and calculative commitment were each
measured by four items based on the construct definitions and scales of
Kumar et al. (1994). As mentioned before, trust in a partner consists of
honesty and benevolence of that partner. Trust in the honesty of the
manufacturer was measured by five items based on the construct definitions
and scales of Kumar et al. (1995). Trust in benevolence of the partner was
also measured by five items based on construct definitions and scales of
Kumar et al. (1995). The dependence on the manufacturer was measured by
three items based on the definitions also by Kumar et al. (1995). Overall
perceived service quality, divided into technical and functional quality, was
measured using the disconfirmation method, that is customers were asked to
give a direct comparison of their perceptions to their expectations on 21
service items (Cronin and Taylor, 1992). The reference point for their
expectations was the service an excellent company should offer. These 21
items were customized in wording and subject to the specific service setting
of this study, as Carman (1990) proposes. For measuring the overall
satisfaction with the manufacturer’s service a single item was used. As
already mentioned, customers were asked to grade the service of the partner
on a ten-point scale. Finally, the behavioral consequence, the intention to
stay in the relationship was measured by three items. These items are based
on the “expectation of continuity” construct provided by Kumar et al.
(1995).
The questionnaire consisted of the items mentioned above and was pre-
tested in two stages. First, marketing research students were asked to fill in

414 JOURNAL OF BUSINESS & INDUSTRIAL MARKETING, VOL. 13 NO. 4/5 1998
 the questionnaire and to detect biases. Second, employees of the office
equipment manufacturer were asked to do the same. After each stage the
questionnaire was modified and refined. Sample items are included in the
Appendix.

Construct validation
Structural equation To validate the constructs in our study we used structural equation modeling
modeling (Bagozzi and Philips, 1991; Bollen, 1989; Gerbing and Anderson, 1988;
Steenkamp and van Trijp, 1991). As the numbers of items adversely affects
the acceptance of structural equation models, we split the constructs in our
study into two subsets. The first model contains technical and functional
quality. The second model is concerned with affective commitment,
calculative commitment, trust (benevolence), trust (honesty), dependence
and intention to stay. This method has recently been recommended for
construct validation. PRELIS and LISREL 7 were used to obtain maximum
likelihood estimates for the parameters in the confirmatory factor analysis
(Jöreskog and Sörbom, 1989)[1].
In model 1 we initially obtained an unacceptable fit. However, after an
iterative process, in which we inspected the normalized residuals and the
modification indices, ten quality items were deleted and we obtained a
good fit: χ2 = 160.41 (p < 0.001); GFI = 0.94; AGFI = 0.90; RMSR = 0.05;
NFI = 0.95; TLI = 0.95 (Bagozzi and Yi, 1988; Bollen, 1989; Hoelter,
1983; Steenkamp and van Trijp, 1991). Although the significant χ2
indicates that the model does not fit the data adequately, we have to realize
that the large sample size affects the χ2-test (Bearden et al., 1982; Bentler
and Bonnett, 1980; Hoelter, 1983; Marsh et al., 1988). Therefore, we used
additional measures such as GFI, AGFI, NFI and TLI to assess the fit of
our model. For further analyses there remained three items for technical
quality and eight items for functional quality. The results for model 1 are
summarized in Table I. Initially, model 2 showed an inadequate model fit
as well. Therefore we omitted two items from the model using normalized
residuals and modification indices. The items that did not fit the structure
were the third item of the dependence construct and the first of trust in the
partner’s honesty. These two items were subsequently left out in further
analyses. Consequently model fit proved to be good: χ2 = 594.91
(p < 0.001); GFI = 0.90; AGFI = 0.86; RMSR = 0.05; NFI = 0.92; TLI =
0.93 (see Table I).
A high degree of Additionally, the reliability of the constructs, using composite reliability and
reliability variance extracted measures was evaluated (Fornell and Larcker, 1981).
From Table I it can be concluded that the constructs exhibited a high degree
of reliability in terms of composite reliability, all values exceeded the
recommended value of 0.7, and variance extracted, for which all values
exceeded 0.5.

Hypothesis testing
In order to test the hypotheses we formulated we used path analysis with
observable variables using PRELIS and LISREL 7 (Jöreskog and Sörbom,
1989). We used LISREL to obtain maximum likelihood estimates of the
path coefficients. It can be concluded that the fit of this model is good: χ2 =
12.82, df = 10 (p = 0.23); GFI = 0.99; AGFI = 0.97; RMSR = 0.04; NFI =
0.99; TLI = 0.99; R2 = 0.72. All measures well exceed the recommended
cut-off values (Bagozzi and Yi, 1988; Jöreskog and Sörbom, 1989).
Furthermore, the values of the χ2/df ratio also indicates a good fit of the

JOURNAL OF BUSINESS & INDUSTRIAL MARKETING, VOL. 13 NO. 4/5 1998 415
 Model fit χ2 df GFI AGFI RMSR NFIa TLIa R2

Model 1 160.41 (p < 0.001) 43 0.94 0.90 0.05 0.95 0.95 0.98
Model 2 594.91 (p < 0.001) 194 0.90 0.86 0.05 0.92 0.93 0.99

Reliability
of measures Construct Reliability Variance extracted

Model 1 Technical quality 0.84 0.64

Functional quality 0.90 0.53
Model 2 Affective commitment 0.95 0.83
Calculative commitment 0.87 0.64
Trust benevolence 0.92 0.70
Trust honesty 0.80 0.51
Dependence 0.80 0.53
Intention to stay 0.77 0.53

Note: aThe null model assumes no underlying factors

Table I. Model reliability of the constructs

data to the hypothesized model (Bagozzi and Yi, 1988; Hoelter, 1983).
Inspection of the standardized residuals showed that none of these exceeded
the absolute value of 2.58, the cut-off value suggested by Jöreskog and
Sörbom (1989). Finally the Q-plot clearly showed a linear trend through the
plotted values indicating a good fit (Bagozzi and Yi, 1988; Jöreskog and
Sörbom, 1989).
Positive relationship Inspection of the path coefficients allows us to test our hypothesis (see
Table II). To begin with, the relationships between service quality,
satisfaction and affective and calculative commitment were tested. We fail to
reject hypothesis 1 on the basis of a significant positive relationship between
satisfaction and affective commitment (standardized path coefficient = 0.19;
t-value = 4.95). More satisfied customers will be more affectively committed
to the supplier. Furthermore, we also found a significant relationship
between satisfaction and calculative commitment, and therefore fail to reject
hypothesis 2 commitment (standardized path coefficient = 0.09; t-value =
1.94). More satisfied customers will be more calculatively committed to the
supplier. A significant positive relationship between technical quality and
affective commitment means we fail to reject hypothesis 3 (standardized
path coefficient = 0.16; t-value = 4.31). Higher quality of the service
outcome will result in more affective commitment of customers to the
supplier. In our study we found no support for hypothesis 4; there is no
significant relationship between the functional quality of a service, that is the
quality of the service process, and affective commitment of customers. Also
hypotheses 5 and 6 are not supported by our study: both technical and
functional quality are not significantly related to calculative commitment.
Hypothesis 7, however, a positive relation between technical quality and
satisfaction, is supported by our findings (standardized path coefficient =
0.33; t-value = 8.07). In other words, higher levels of outcome quality will
lead to higher levels of customer satisfaction. The same applies to functional
quality; higher quality of the service process will result in more satisfaction
(standardized path coefficient = 0.16; t-value = 4.05). This is in support of
hypothesis 8.

416 JOURNAL OF BUSINESS & INDUSTRIAL MARKETING, VOL. 13 NO. 4/5 1998
 Relationships Hypothesis Standardized path coefficientb Hypothesis testing

SAT → AC H1 0.19 (4.95) Failed to reject

SAT → CC H2 0.09 (1.94) Failed to reject
TQ → AC H3 0.16 (4.31) Failed to reject
FQ → AC H4 n.s. Rejected
TQ → CC H5 n.s. Rejected
FQ → CC H6 n.s. Rejected
TQ → SAT H7 0.33 (8.07) Failed to reject
FQ → SAT H8 0.16 (4.05) Failed to reject
TB → AC H9 0.28 (6.99) Failed to reject
TB → CC H10 0.40 (7.65) Rejected
TH → AC H11 0.33 (8.38) Failed to reject
TH → CC H12 –0.23 (–4.26) Failed to reject
DEP → AC H13 n.s. Rejected
DEP→ CC H14 0.30 (7.29) Failed to reject
SAT → ItS H15 n.s. Rejected
AC → ItS H16 0.39 (8.51) Failed to reject
CC → ItS H17 0.14 (3.68) Failed to reject

a SAT = satisfaction; TQ = technical quality; FQ = functional quality; TB = trust

benevolence; TH = trust honesty; DEP = dependence; AC = affective commitment;
CC = calculative commitment; ItS = intention to stay
b Corresponding t-value in parentheses

Table II. Estimated path coefficientsa

Commitment, trust and Next, the relationships between commitment, trust and dependence were
dependence studied. We fail to reject hypothesis 9 on the basis of a significant positive
relationship between trust (benevolence) and affective commitment (standard-
ized path coefficient = 0.28; t-value = 6.99). The more a customer believes that
its supplier is interested in its welfare and that the supplier will not take any
unexpected actions which negatively impacts the customer, the higher the
level of affective commitment of the customer in the relationship. The
negative relation between trust in the supplier’s benevolence and calculative
commitment, as stated by hypothesis 10, is not supported. A significant
positive relationship between these two constructs is found (standardized path
coefficient = 0.40; t-value = 7.65). Apparently, more trust in the partner’s
benevolence leads to increased calculative commitment of a customer. Both
hypotheses 11 and 12 on trust in the partner’s honesty are supported (i.e. fail to
reject). The higher the belief of the customer that the supplier will stand by its
word, the higher the affective commitment of the customer in the relationship
(standardized path coefficient = 0.33; t-value = 8.38) and the lower the
calculative commitment (standardized path coefficient = –0.23; t-value =
–4.26). Dependence on the supplier was not found to be significantly related to
affective commitment, which is in conflict with hypothesis 13. However, it is
positively related to calculative commitment (standardized path coefficient =
0.30, t-value = 7.29). The more a customer depends on its supplier, the higher
its calculative commitment in the relationship with that partner and therefore,
the more it will be balancing gains and losses of that relationship. This is in
accordance with hypothesis 14.
Finally, the consequences of satisfaction and commitment in a relationship
marketing context were tested. First, no significant relation between
customer satisfaction and the intention to stay in the relationship was found.
This leads to rejection of hypothesis 15. However, it can be concluded that
both types of commitment do influence the intention to stay. More

JOURNAL OF BUSINESS & INDUSTRIAL MARKETING, VOL. 13 NO. 4/5 1998 417
 affectively committed customers intend to stay longer in the relationship
with the supplier (standardized path coefficient = 0.39; t-value = 8.51) and
also more calculatively committed customers have a higher intention to
continue that relationship (standardized path coefficient = 0.14; t-value =
3.68). Therefore, we cannot reject hypotheses 16 and 17. Table II
summarizes the results of our hypotheses testing.
Trust in a partner’s Inspection of the modification indices suggested three additional
benevolence and honesty relationships in our model (Jöreskog and Sörbom, 1989). In addition to the
relations mentioned above, it could be concluded that trust in the partner’s
benevolence and honesty also have a direct positive effect on customer
satisfaction with the service (standardized path coefficient = 0.1; t-value =
2.15 and standardized path coefficient = 0.28, t-value = 6.34 respectively).
Moreover, trust in benevolence was also found to be directly influencing the
intention to stay in a positive way (standardized path coefficient = 0.26; t-
value = 5.49). Since we left the realm of confirmatory analyses by adding
these three paths to our model, our final model needs to be cross-validated
using additional samples (Cudeck and Brown, 1983; MacCallum, 1986).

Conclusion
Discussion
Our study indicates that for the establishment of affective commitment in a
supplier-customer relationship, it is very important that a customer has a lot
of trust in the supplier’s benevolence and honesty. If a customer believes that
the supplier is benevolent, which means that he/she is interested in the
customer’s welfare and will not take any unexpected actions which will
negatively influence the company, then also more affective commitment will
emerge. This similarly applies for honesty. Also by paying attention to
service outcome and constantly improving this technical quality, customers
will be not only more affectively committed in the relationship, but also
more satisfied. In turn, this higher satisfaction also contributes to more
affectively committed customers. Other ways of increasing customer
satisfaction is by improving the functional quality of the service, that is the
quality of the service process and developing trust in the supplier’s
benevolence and honesty.
Importance of service Our study also shows that affective commitment is not significantly
outcome influenced by functional quality and by the level of dependence between two
parties. Apparently, the quality of the service outcome is more important in
establishing affective commitment than the quality of the service process.
Furthermore, it has to be kept in mind that in this specific research setting
(office equipment) technical service is often considered most important by
people who frequently have a technical background. However, as mentioned,
functional quality results in higher levels of satisfaction, which in turn leads
to affective commitment. Therefore, functional quality is an indirect
antecedent of affective commitment. Furthermore, asymmetry in a
relationship between supplier and customer does not automatically imply
that a customer will be less affectively committed to the relationship.
Calculative commitment is not affected directly by quality. However, as both
technical and functional quality are related to satisfaction, quality indirectly
affects calculative commitment. Trust in the supplier’s benevolence
positively affects calculative commitment, whereas trust in the supplier’s
honesty reduces calculative commitment. Furthermore, we find that
dependence shows a positive impact on calculative commitment.

418 JOURNAL OF BUSINESS & INDUSTRIAL MARKETING, VOL. 13 NO. 4/5 1998
 Customers’ intentions to The importance of affective commitment in business relationships is stressed
continue the relationship by its strong relationship with customers’ intentions to continue a
relationship. More affectively committed partners show a stronger intention
to stay than customers who feel more calculative commitment. This latter
type of commitment is weakly positively related to the intention to stay.
Moreover, trust in benevolence significantly influences the intention to stay
in the relationship. More trust in benevolence of a supplier leads to a
stronger intention of customers to continue the relationship with that
supplier. Surprisingly, this intention of continuation is not significantly
influenced by customer satisfaction. Satisfaction is apparently not a
differentiator in this specific research setting. This might be a consequence
of the fact that all major players in the market for office equipment put a
high emphasis on customer satisfaction or because the company has reported
relatively high satisfaction ratings that have been fairly consistent over the
past five years. However, because satisfaction leads to higher affective
commitment, which in turn has the strongest influence on the intention to
stay, it has an indirect effect on intention to stay in the relationship.

Theoretical implications
Preliminary findings Certain limitations should be kept in mind when interpreting the preliminary
findings reported in this paper. First, the constructs that were studied repre-
sent very complex phenomena. Future studies will have to determine which
conceptualizations will be needed to yield most reliable and best explana-
tions. Second, important questions remain as to the complexity of the
antecedents of commitment. In this paper only a limited number were
reported. Additional research is needed as to other variables that lead to
relationship commitment. Potential variables include the nature and fre-
quency of interaction between exchange partners, the quality of communi-
cations, uncertainty in the competitive environment, perceived risk, the
services portfolio, etc. Third, all concepts were measured at one point in
time. More insight is needed as to how the concepts develop over time.
Finally, the effects were estimated with regards to intentions. Actual beha-
vior such as repeat “buying” behavior and word-of-mouth were not included
in this study. However, they should be studied in order to complete the
picture.

Managerial implications
Focal point There are several factors management of service providers can use to
improve their specific service delivery process and to establish long-lasting
relationships with their customers. Most important is the creation of
affective commitment of customers in their relationship with the supplier.
Affective commitment has the strongest effect on intention to stay in the
relationship, and thus on future revenues. This affective commitment is
positively influenced by higher levels of satisfaction of customers, and by
the improvement of the technical quality of the service and trust in the
supplier. The creation of trust in the relationship should be a focal point of
attention for managers. It was shown that trust not only has a positive effect
on affective commitment, but also on satisfaction and the intention to stay in
the relationship. Trust arises if the manufacturer is honest and benevolent
with the customer. Fair treatment, open bi-lateral communication and
information exchange, a cooperative orientation as well as sound economic
results are factors that have been shown to be building blocks of trust (Berry,
1995; Geyskens et al., 1996). Opening communication lines can be achieved
by installing a customer contact program in which the company takes the

JOURNAL OF BUSINESS & INDUSTRIAL MARKETING, VOL. 13 NO. 4/5 1998 419
 initiative to keep in touch with its customers. Also, inviting customers to
product and service information meetings and seminars in regional and local
markets might be another communication initiative that managers may use
to establish trust as well as affective commitment. Delivering service to
customers through a team of cross-functional customer contact employees
ensures that customers can get a higher level of personalized service for a
wide range of services, as they keep meeting the same employees for every
service call. Service guarantees are also efficient instruments in establishing
trustworthy relationships with business-to-business customers. To these
customers service guarantees are signs of a company’s customer
commitment and form an attribute through which companies can distinguish
themselves from competitors. Moreover, companies could experiment with
other reward systems that encourage personnel to maintain affective
relationships with customers instead of focusing almost solely on sales
figures. Customer contact personnel view themselves as relationship
managers.
Demanding higher Finally, trust and affective commitment can be promoted by implementing
standards and demanding higher standards of conduct from their personnel. As Berry
and Parasuraman (1991, p. 145) state: “they must be willing to ask not only
‘Is it legal?’ but also ‘Is it right?’”
As mentioned, trust has a positive effect not only on affective commit-
ment, but also on customer satisfaction. Satisfaction in turn has, through
affective commitment, an indirect positive effect on intention to stay in the
relationship. Management should try to avoid business relationships in
which the partner only feels calculative commitment. Given the choice
between developing closeness to the service provider through attraction or
obligation, it is clear that management should cultivate the former over the
latter.
Finally, acknowledgment of the importance of customer commitment may
provide input for developing alternative segmentation strategies for firms
operating in business-to-business markets (Van der Walt et al., 1994).
Instead of traditional customer criteria, such as client size and type of
industry, clients could also be segmented on the basis of commitment
characteristics. These client segments would require different relationship
management tactics and strategies.
Dependence symmetry For instance, our study indicates that calculative committed customers have
a much weaker intention to stay than affective committed customers. As
soon as alternative suppliers emerge they probably will switch service
providers. The way to avoid high calculative commitment is to strive for
dependence symmetry and honesty toward the customer. A balanced and
honest relationship will result in less calculatively and more affectively
committed customers, more satisfied customers, and therefore in a long-
lasting profitable relationship for both parties.
In markets for business services in which a high emphasis is placed on
integrated networks between suppliers and customers, commitment and
commitment-related issues are likely to play a crucial role. It is hoped that
our study provides a small contribution to a better understanding of these
issues.

Note
1. Satisfaction was not included in our analysis as it was operationalized as a single-item
measure.

420 JOURNAL OF BUSINESS & INDUSTRIAL MARKETING, VOL. 13 NO. 4/5 1998
 References
Anderson, E. and Fornell, C. (1994), “A customer satisfaction research prospectus”, in Rust,
R.T. and Oliver, R.L. (Eds), Service Quality: New Directions in Theory and Practice,
Sage Publications, Thousand Oaks, pp.1-19.
Anderson, E. and Weitz, B. (1989), “Determinants of continuity in conventional industrial
channel dyads”, Marketing Science, Vol. 4 No. 4, pp. 310-23.
Anderson, J.C. and Narus, J.A. (1990), “A model of distributor firm and manufacturer firm
working partnerships”, Journal of Marketing Research, Vol. 54, pp. 42-58.
Armstrong, J.S. and Overton, T.S. (1977), “Estimating nonresponse bias in mail surveys”,
Journal of Marketing Research, pp. 396-402.
Assael, H. (1987), Consumer Behavior and Marketing Action, PWS Kent, Boston.
Bagozzi, R.P. and Phillips, L.W. (1991), “Assessing construct validity in organizational
research”, Administrative Science Quarterly, Vol. 36, pp. 421-58.
Bagozzi, R.P. and Yi, Y. (1988), “On the evaluation of structural equation models”, Journal of
the Academy of Marketing Science, Vol. 16, Spring, pp. 74-94.
Bearden, W.O., Sharma, S. and Teel, J.E. (1982), “Sample size effects on chi square and other
statistics used in evaluating causal models”, Journal of Marketing Research, Vol. 19,
November, pp. 425-30.
Becker, H.S. (1960), “Notes on the concept of commitment”, American Journal of Sociology,
No. 66, pp. 32-40.
Bentler, P.M. and Bonnett, D.G. (1980), “Significance tests and goodness of fit in the analysis
of covariance structures”, Psychometrika, Vol. 51, pp. 375-7.
Berry, L.L. (1983), “Relationship marketing”, in Berry, L.L., Shostack, G.L. and Upah, G.
(Eds), Emerging Perspectives on Services Marketing, American Marketing Association,
Chicago, IL, pp. 25-8.
Berry, L.L. (1995), “Relationship marketing of services – growing interest, emerging
perspectives”, Journal of the Academy of Marketing Science, Vol. 23 No. 4, pp. 236-45.
Berry, L.L. and Parasuraman, A. (1991), Marketing Services – Competing through Quality,
The Free Press, New York, NY.
Bollen, K.K.A. (1989), Structural Equations with Latent Variables, John Wiley & Sons, New
York, NY.
Brislin, R.W. (1980), “Translation and content analysis of oral and written materials”, in
Triandis, H.C. and Berry, J.W. (Eds), Handbook of Cross-Cultural Psychology:
Methodology, Vol. 2, Allyn and Bacon, Boston, MA.
Carman, J.M. (1990), “Consumer perceptions of service quality: an assessment of the
SERVQUAL dimensions”, Journal of Retailing, Vol. 66 No. 1, pp. 33-55.
Cronin J.J. Jr and Taylor, S.A. (1992), “Measuring service quality: a reexamination and
extension”, Journal of Marketing, Vol. 56, pp. 55-68.
Cudeck, R. and Brown, M.W. (1983), “Cross-validation of covariance structures”, Multivariate
Behavioral Research, Vol. 18, pp. 147-67.
Dwyer, F.R., Schurr, P.H. and Oh, S. (1987), “Developing buyer-seller relationships”, Journal
of Marketing, Vol. 51, pp. 11-27.
Fehr, B. (1988), “Prototype analysis of the concepts of trust and commitment”, Journal of
Personality and Social Psychology, Vol. 55 No. 4, pp. 557-79.
Fornell, C. and Larcker, D.F. (1981), “Evaluating structural equation models with
unobservable variables and measurement error”, Journal of Marketing Research,
Vol. XVIII, pp. 19-50.
Ganasan, S. (1994), “Determinants of long-term orientation in buyer-seller relationships”,
Journal of Marketing, Vol. 58, pp. 1-19.
Gerbing, D.W. and Anderson, J.C. (1988), “An updated paradigm for scale development
incorporating unidimensionality and its assessment”, Journal of Marketing Research,
Vol. XXV, May, pp. 186-92.
Geyskens, I. and Steenkamp, J.B. (1995), “An investigation into the joint effects of trust and
interdependence on relationship commitment”, EMAC Proceedings, Paris, pp. 351-71.
Geyskens, I., Steenkamp, J.B.E.M., Scheer, L.K. and Kumar, N. (1996), “The effects of trust
and interdependence on relationship commitment: a transatlantic study”, International
Journal of Research in Marketing, Vol. 13, pp. 303-17.

JOURNAL OF BUSINESS & INDUSTRIAL MARKETING, VOL. 13 NO. 4/5 1998 421
 Grönroos, C. (1990), Service Management and Marketing: Managing Moments of Truth in
Service Competition, Lexington Books, Lexington, MA and Toronto.
Grönroos, C. (1995), “Relationship marketing: the strategy continuum”, Journal of the
Academy of Marketing Sciences, Vol. 23 No. 4, pp. 252-4.
Gundlach, G.T., Achrol, R.S. and Mentzer, J.T. (1995), “The structure of commitment in
exchange”, Journal of Marketing, Vol. 59, January, pp. 78-92.
Hoelter, J.W. (1983), “The analysis of covariance structures: goodness-of-fit indices”,
Sociological Methods and Research, Vol. 11, February, pp. 325-44.
Iacobucci, D., Ostrom, A.L., Braig, B.M. and Bezjian-Avery, A. (1996), “A canonical model
of consumer evaluations and theoretical bases of expectations”, Advances in Services
Marketing and Management, Vol. 5, pp. 1-44.
Jöreskog, K.G. and Sörbom, D. (1989), LISREL® 7: A Guide to the Program and Application,
Jöreskog & Sörbom/SPSS Inc., Chicago, IL.
Kumar, N., Hibbard, J.D. and Stern, L.W. (1994), The Nature and Consequences of Marketing
Channel Intermediary Commitment, MSI Working Paper, Report No. 94-115.
Kumar, N., Scheer, L.K. and Steenkamp, J.B.E.M. (1995), “The effects of perceived
interdependence on dealer attitudes”, Journal of Marketing Research, Vol. XXXII,
pp. 348-56.
MacCallum, R. (1986), “Specification seraches in covariance structure modeling”,
Psychological Bulletin, Vol. 100, pp. 107-20.
Marsh, H.W., Balla, J.R. and McDonald, R.P. (1988), “Goodness-of-fit indexes in
confirmatory factor analysis: the effect of sample size”, Psychological Bulletin, Vol. 103
No. 3, pp. 391-410.
Morgan, R.M. and Hunt, S.D. (1994), “The commitment-trust theory of relationship
marketing”, Journal of Marketing, Vol. 58, pp. 20-38.
Morrow, P. (1983), “Concept redundancy in organizational research: the case of work
commitment”, Academy of Management Review, Vol. 8, pp. 486-500.
Oliver, R.A. (1996), Satisfaction. A Behavioral Perspective on the Consumer, McGraw-Hill,
New York, NY.
Reichheld, F.F. and Sasser, W.E. Jr (1990), “Zero-defections quality comes to services”,
Harvard Business Review, September-October, pp. 105-11.
Scanzoni, J. (1979), “Social exchange and behavior interdependence”, in Burgess, R.L. and
Huston, T.L. (Eds), Social Exchange in Developing Relationships, Academic Press, Inc.,
New York, NY.
Steenkamp, J.B.E.M. and van Trijp, H.C.M. (1991), “The use of LISREL® in validating
marketing constructs”, International Journal of Research in Marketing, Vol. 8 No. 4,
November, pp. 283-99.
van der Walt, S.D., Scott, D. and Woodside, A.G. (1994), “CPA service providers: a profile of
client types and their assessment of performance”, Journal of Business Research,
pp. 225-33.
Williamson, O.E. (1975), Markets and Hierarchies, Analysis and Antitrust Implications, The
Free Press, New York, NY.
Williamson, O.E. (1979), “Transaction-cost economics: the governance of contractual
relations”, Journal of Law and Economics, Vol. 22, pp. 233-61.
Williamson, O.E. (1983), “Credible commitments: using hostages to support exchange”,
American Economic Review, Vol. 73, pp. 519-40.
Williamson, O.E. (1984), “Credible commitments”, The Antitrust Bulletin, Spring, pp. 33-76.
Williamson, O.E. (1985), The Economic Institutions of Capitalism, The Free Press, New York,
NY.

Appendix: Measures
Construct Sample item
Technical quality The mean time between failure of company X equipment (plus
disconfirmation scale).
Functional quality The understanding of your needs by company X staff (plus
disconfirmation scale).
Trust benevolence Though circumstances may change, we believe that the supplier
will be ready and willing to offer us assistance and support.

422 JOURNAL OF BUSINESS & INDUSTRIAL MARKETING, VOL. 13 NO. 4/5 1998
 Trust honesty We can count on this firm to be sincere.
Dependence There is really no alternative for this firm.
Satisfaction Please indicate how satisfied you are with the overall service of
company X.
Affective commitment We want to remain a customer of this firm because we genuinely
enjoy our relationship with them.
Calculative commitment There is just too much time, energy and expense involved in
terminating our relationship with this firm.
Intention to stay We expect our relationship with the supplier to continue for a long
time.
■

JOURNAL OF BUSINESS & INDUSTRIAL MARKETING, VOL. 13 NO. 4/5 1998 423