Relationships Between Marketing & Sales: the role of ... - CiteSeerX

Wolverhampton Business School Management Research Centre __________________________________________________________________________________________________

Relationships Between Marketing & Sales: the role of power and influence by Philip L Dawes & Graham Massey Working Paper Series 2001

Number

WP 009/01

ISSN Number

ISSN 1363-6839

Philip L Dawes Chair in Marketing University of Wolverhampton, UK Tel: +44 (0) 1902 323700 Fax: +44 (0) 1902 323178 Email: [email protected]

Graham Massey Lecturer in Marketing University of Western Sydney Macarthur, PO Box 555, Campbelltown NSW 2560 AUSTRALIA Tel: +61 246 203 522 Fax: +61 246 203 799 Email: [email protected]

© University of Wolverhampton 2001 - All rights reserved

Relationships Between Marketing and Sales: the role of power and influence _________________________________________________________________________________________

Copyright ©

University of Wolverhampton 2001

All rights reserved. No part of this work may be reproduced, photocopied, recorded, stored in a retrieval system or transmitted, in any form or by any means, without the prior permission of the copyright holder.

The Management Research Centre is the co-ordinating centre for research activity within Wolverhampton Business School. This working paper series provides a forum for dissemination and discussion of research in progress within the School. For further information contact: Management Research Centre Wolverhampton Business School Telford, Shropshire TF2 9NT !01902 321772 Fax 01902 321777 The Working Paper Series is edited by Kate Gilbert

2

Management Research Centre 2001


Abstract Gaining a better understanding of marketing’s cross-functional relationships (CFRs) has considerable importance for both academics and practitioners. The purpose of this paper is to explore the nature of one of the least understood of marketing’s CFRs viz., that with sales. Focusing on the key outcome variable of perceived relationship effectiveness, we develop and test a model to provide insights into the factors which either increase or decrease the effectiveness of the marketing/sales working relationship. Our proposed model hypothesizes that the power of the marketing department, the stakeholding of the marketing manager, and the interdependence between the sales and marketing manager, affect relationship effectiveness through the mediating variables of influence strategies, manifest influence, and interpersonal trust.

3



The authors Professor Philip L Dawes Philip Dawes is Professor of Marketing at the Business School of Wolverhampton University. His research interests include industrial marketing, organisational buyer behaviour, high-technology marketing in business markets, marketing organisation, and services marketing. Graham Massey Graham Massey is a Lecturer in Marketing at the University of Western Sydney, Australia. His research interests include marketing organisation, relationship marketing, new product development, and product life cycle theory.

4



Relationships Between Marketing and Sales: the role of power and influence Introduction Barely a decade ago, issues of marketing organization were peripheral to scholarly visions of marketing’s future (Day, 1997). Today however, these issues are high on the research agenda, and evidence of their importance can be seen in the large and rapidly growing body of literature examining cross-functional relationships (CRFs) and integration between marketing and other functional units (e.g. Fisher, Maltz, & Jaworski, 1997; Workman, Homburg, & Gruner, 1998). To date, the bulk of this research has focused on the CFR between Marketing and R&D, and has been driven largely by an interest in new product development and innovation in general. This is understandable, given the importance of innovation within modern organizations and the central role of marketing in driving these processes. Likewise, other scholars have noted marketing’s key strategic position within organizations and the need for marketers to manage a wide range of important CFRs (Hutt, 1995). Surprisingly, one key CFR remains relatively unexplored in the academic literature, i.e., the working relationship between the sister functions of sales and marketing. Accordingly, the purpose of our research is to begin to fill this gap in the literature and, as suggested by Ruekert and Walker (1987), we focus on dyadic interactions between the marketing manager and the sales manager because individual employees are the most appropriate starting point to examine CFRs. Drawing on theoretical perspectives of organizational and social exchange theories, we develop and test a model of relationship effectiveness. In total, six explanatory constructs emerged from this literature search. These six constructs are: interpersonal trust, manifest influence, influence strategies, departmental power, stakeholding, and total interdependence. Our paper is structured as follows. First, we review the theoretical foundations of our research and then develop a structural integrated model. Next, we describe the methodology and present our preliminary findings.

Theoretical foundations Essentially, our research uses an interaction approach to explain the relational outcome of perceived relationship effectiveness. This approach, which has been used in much of the research on relational outcomes, draws on exchange theories and game theory and focuses on understanding how constructs such as power, influence, trust, and cooperation, predict satisfaction, performance, and relationship continuity in buyer-seller, channel, and supplier contexts (e.g., Anderson & Narus, 1990; Morgan & Hunt, 1994; Moorman, Deshpandé, & Zaltman, 1993). Though our research draws mainly from this interaction approach, we also build on the more recent work of Smith and Barclay (1999) who combined the interaction approach with the ‘structural-contingency approach’. In short, this latter approach suggests that relational outcomes can be explained by the organizational, structural, and environmental characteristics facing the parties (e.g. Pfeffer & Salancik, 1978). As noted by Smith and Barclay (1999), an important variable identified by this literature is interdependence. Accordingly, we draw from both approaches to examine power, influence, interpersonal trust, and interdependence in the same working relationship model.

5

Power of

H5a



Figure 1. Conceptual Model

Model development As shown in Figure 1, three exogenous variables (power of the marketing department, stakeholding, and total interdependence) and three endogenous variables (influence strategies, interpersonal trust, and manifest influence) are linked to explain perceived relationship effectiveness. As part of this research, the six hypotheses denoted in Figure 1 will be developed and tested using data collected from firms in both the UK and Australia. In order to clarify the focus of our research, we start by examining the focal outcome variable of perceived relationship effectiveness and then develop our research hypotheses. The dependent variable of perceived relationship effectiveness was chosen because: (a) previous studies of effective working relationships have also focused on subjective outcomes (Anderson & Narus, 1990); and (b) objective measures of effectiveness, such as sales volume, may not accurately reflect the quality of a relationship due to confounding factors such as long sales cycles. Effect of the marketing manager’s manifest influence Consistent with recent research by Dawes, Lee, and Dowling (1998), we define manifest influence in terms of the actual effect that a marketing manager had on a specific project in terms of changing the opinions and behaviours of other members of the decision-making unit. The influence construct is included in our research because Ruekert and Walker (1987) argue that influence is likely to be an important co-ordinating dimension in CFRs, and because Homburg, Workman, and Krohmer (1999) suggest that much more empirical research is required in order to understand marketing’s influence on decision-making within the firm. A feature of our research is that it focuses on a specific project carried out by a cross-functional team which includes at least one person each from sales and marketing, along with a varying number of people from other departments. Needless to say, in such a 6



decision-making group, where often there is internal competition (Conrad 1990), people from the different functional areas are likely to be jockeying for positions of power in order to further the interests of themselves and their department. Perceived relationship effectiveness. Though there may well be some conflicts of interests between the ‘sister’ functions of sales and marketing, we argue that, on balance, in teams where the marketing manager is more effective in being able to change the opinions and behaviours of other members of the team, the sales manager will perceive their dyadic relationship to be more effective. In such instances, due to their increased manifest influence, marketing managers are likely to be able to obtain a greater part of the available resources, which they may share with the sales manager. If this sharing of resources happens, which is highly likely due to the need to build coalitions in cross-functional teams (Conrad 1990), it seems reasonable to expect the sales manager to think that his/her relationship with the marketing manager is effective. But why should a marketing manager share more of these resources with the sales manager as opposed to other members of the cross-functional team? Part of the answer to this question is due to the likelihood that the persons in the marketing/sales dyad will have higher domain similarity1 as compared to the persons in the other marketing dyads (e.g., marketing/finance) in the cross-functional team. Ruekert and Walker (1987) argue that the amount of resource flows between marketing people and those in other functional areas is positively related to the degree of domain similarity between them. Formally, we express the relationship between the marketing manager’s manifest influence and perceived relationship effectiveness in our first hypothesis. H1: As the manifest influence of the marketing manager increases, the perceived effectiveness of the sales/marketing manager relationship will increase. Effects of interpersonal trust We include this variable because a variety of studies in marketing (e.g., Moorman, Zaltman, & Deshpandé, 1992) and in organization behaviour suggest that this social psychological factor is likely to be important in determining relationship effectiveness. In contrast to previous studies in marketing, we follow McAllister (1995) who drew upon literature in both sociology and social pyschology and showed empirically that interpersonal trust has both cognitive and affective dimensions. Cognitive-based trust (CBT) derives from a person’s rational bases for trusting another person, e.g., previous occasions in which the other person has been competent, reliable, and dependable in performing tasks which affected the other person. In contrast, affect-based trust (ABT) is typified by emotional bonds between individuals, in which one party exhibits genuine concern and care for the welfare of the other person (Lewis & Wiegert, 1985). Perceived relationship effectiveness. The direct effects of CBT and ABT on the perceived effectiveness of the working relationship between two managers are not well understood. McAllister (1995 p. 32) for example noted that “existing research contains little on how trust affects performance outcomes.” However, he argues that trusting peers are likely to assess each other’s performance more favourably due to the beneficial effects of the behavioural consequences of trust, e.g., control-based monitoring, individual citizenship behaviour, and defensive behaviour. A study by Smith and Barclay (1997) examined the effects of trusting behaviours (e.g., investment of resources and effort into the working relationship) on perceived task performance and their findings generally support the proposition that greater trust is associated with greater perceived ‘task performance’, a construct which is conceptually similar to our dependent variable of ‘perceived relationship effectiveness’. Accordingly, we hypothesise that: H2a: As the sales manager’s cognition-based trust of the marketing manager increases, the perceived effectiveness of their working relationship will increase. H2b: As the sales manager’s affect-based trust of the marketing manager increases, the perceived effectiveness of their working relationship will increase.

7



Manifest influence. While we can find no direct evidence to support our contention that greater trust in the marketing manager will lead a sales manager to attribute greater influence to this co-worker, there is some research which provides indirect support. For example, Yukl (1981), in his summary of the extant literature, noted that lateral influence is often exercised through rational persuasion and appeals based on friendships. As just noted, CBT derives from a person’s rational bases for trusting another person, while ABT is typified by emotional bonds between two individuals. Later empirical research by Keys, Case, Miller, Curran, and Jones (1987), lends further support to Yukl’s (1981) viewpoint regarding how lateral influence is exercised and, accordingly, we argue that CBT and ABT are likely to be positively associated with manifest influence. H2c: As the sales manager’s cognition-based trust of the marketing manager increases, the perceived level of manifest influence of the marketing manager will increase. H2d: As the sales manager’s affect-based trust of the marketing manager increases, the perceived level of manifest influence of the marketing manager will increase. Effects of the influence strategies used by the marketing manager In order to more fully understand how manifest influence acts as an informal co-ordinating mechanism in CFRs, two influence strategies of threats and legalistic pleas are included as predictor variables in our model. The threats type of influence strategy relates to instances when the source communicates to the target that he/she will apply negative sanctions if the target does not perform the desired action (Frazier & Summers, 1984). Research in organizational communication (Yukl, Falbe, & Youn, 1993) indicates that the threats (or pressure) strategy is used most often when an initial influence attempt has failed and is generally viewed as being an inappropriate form of influence behaviour since target resentment about the source’s use of coercion is likely to result. In addition, Fisher, Maltz, and Jaworski (1997), in their research on marketing’s CFRs with engineering, argue that coercive influence attempts are by their very essence confrontational, and as a result, can lead to negative psychological and psychosocial outcomes. Together, these two studies provide us with some support to speculate that when a marketing manager uses this coercive influence strategy it will reduce ABT, CBT, and perceived relationship effectiveness. In contrast, the use of the legalistic pleas influence strategy involves efforts to verify the legitimacy of a request and the source’s authority or right to make it. According to Yukl (1990), this tactic is most appropriate for a request that is unusual and of doubtful legitimacy to the target person. Moreover, this influence strategy is needed most in lateral (as opposed to upward or downward) communications because ambiguity about authority relationships and task responsibilities is greatest in this direction (Yukl & Tracey, 1992). Recall that lateral (horizontal) communication is the most important type of communication in CFRs (Ruekert & Walker, 1987). As above, little direct evidence exists for us to make strong predictions about the effects of using this strategy on the two dimensions of interpersonal trust. Nevertheless, because this strategy, like the threats strategy, involves coercion and manipulation, it is less socially acceptable than a soft coercive strategy, such as recommendations (Yukl & Tracey, 1992, p. 526). As a result, the use of this hard coercive strategy may lead to negative psychological and psychosocial outcomes. Using the same logic which was used with respect to the use of the threats strategy, we expect the use of legalistic pleas to reduce affect-based trust, reduce cognition-based trust, and reduce perceived relationship effectiveness.

Manifest influence. In this section, since research on the effects of the threats strategy and the legalistic pleas strategy on manifest influence is very sparse, we are forced to rely mainly on the buying center research of Venkatesh, Kohli, and Zaltman (1995) to justify our hypotheses. But since one their key findings was unexpected and somewhat contradictory, we have no strong justification. Specifically, Venkatesh, Kohli, and Zaltman (1995) hypothesized that the use of both of these hard coercive strategies would lead to increased manifest influence. Though they found that the use of the 8



threats strategy did result in increased manifest influence, they found that the use of legalistic pleas had the opposite effect i.e., its use led to reduced manifest influence. This finding was surprising because, on the basis of prior research (e.g., Yukl & Tracey, 1992), it was expected that the use of legalistic pleas would not only have a positive effect on manifest influence, but its effect would be stronger than that of the threats strategy. Unfortunately, no explanation for this unexpected finding was provided. However, since the study by Venkatesh, Kohli, and Zaltman (1995) is the only one which has specifically examined the use of these two influence strategies in a CFR setting, we use their empirical evidence as the basis for hypothesizing that the threats strategy will lead to increased manifest influence, while the use of legalistic pleas will lead to reduced manifest influence. Accordingly, we hypothesize: H3a: Greater use of threats by the marketing manager will lead to: (1) reduced cognition-based trust, and (2) reduced affect-based trust. H3b: Greater use of legalistic pleas by the marketing manager will lead to: (1) reduced cognitionbased trust and (2) reduced affect-based trust. H3c: Greater use of threats by the marketing manager will lead to greater manifest influence. H3d: Greater use of legalistic pleas by the marketing manager will lead to greater manifest influence. H3e: Greater use of threats by the marketing manager will lead to reduced perceived effectiveness of the sales/marketing manager relationship. H3f: Greater use of legalistic pleas by the marketing manager will lead to reduced perceived effectiveness of the sales/marketing manager relationship. Effects of the marketing manager’s stakeholding Stakeholding refers to the stake which an individual has in the final outcome of a decision made by an organizational work unit (Patchen, 1974). Accordingly, high stakeholders are those with most at risk in the final decision because they are most likely to be held accountable for the outcome of the decision. This construct is included because research by Dawes, Lee, and Dowling (1998) shows that it is a key determinant of an individual’s amount of manifest influence in buying centres, which are DMUs that are very similar in nature to the informal cross-functional groups studied here. Accordingly, we posit that marketing managers who have greater stakeholding in the focal project are likely to be more influential than marketing managers who have less stake in the final outcome of the project. Though no prior research has examined how stakeholding is likely to affect the use of influence strategies, we posit that because of self-interest, marketing managers with greater stakeholding are more likely to use threats and legalistic pleas during the specified project. Therefore, we hypothesize: H4a: Marketing managers with greater stakeholding are more likely to have greater manifest influence. H4b: Marketing managers with greater stakeholding are more likely to use the influence strategies of (1) threats and (2) legalistic pleas. Effects of departmental power Departmental power refers to the relative importance of a department to an organization in general (Kohli, 1989). Evidence provided by Perrow (1970) suggests that an individual’s influence in a DMU is related to the power of his or her department in the organization. Accordingly, we argue that marketing managers who belong to powerful marketing departments are more likely to have greater manifest influence as compared to marketing managers who belong to less powerful marketing departments. Additionally, we posit that marketing managers in powerful marketing departments are more likely to use threats and legalistic pleas during the project.

9



H5a: As the power of the marketing department increases, the marketing manager is likely to have greater manifest influence. H5b: As the power of the marketing department increases, the marketing manager is more likely to use the influence strategies of (1) threats and (2) legalistic pleas. Effects of total interdependence Total interdependence. In line with the majority of the CFR literature, we define the sales manager’s dependence on the marketing manager in terms of how much he/she needs to maintain a relationship with the marketing manager in order for him/her to achieve his/her goals and responsibilities (Fisher, Maltz, & Jaworski, 1997). According to Ruekert and Walker (1987), interdependence is the key internal variable affecting marketing’s interaction with other functional areas. Manifest influence. Ruekert and Walker’s (1987) cross-functional interaction theory posits that the amount of influence exercised by a member of one functional area over a member of another, depends partly on their relative resource dependence. For example, if a marketing manager tightly controls market research information about the re-launch of an existing product, we would expect the sales manager to be dependent on the marketing manager and that the marketing manager would have significant influence over sales decisions. In an empirical test of this part of their theory, Ruekert and Walker (1987) found good support for the proposition that as the dependence of another functional area on marketing increased, the greater was the level of marketing’s influence on the decisions and operations of that other functional area. Accordingly, we hypothesize that: H6a: Greater total interdependence between the sales manager and the marketing manager will lead to the marketing manager having greater manifest influence. Cognition-based trust and affect-based trust. Our basic premise is that we expect that greater total interdependence between a sales manager and a marketing manager will lead to the sales manager having a greater amount of both cognition-based trust and affect-based trust in the marketing manager. Though we propose separate hypotheses for these two dimensions of trust, the following discussion does not distinguish how total interdependence is likely to affect them individually. This is because very little theory or data exists at the interpersonal level to distinguish cognition-based trust from affect-based trust either in terms of their antecedents or consequences (McAllister, 1995 p. 54). Moreover, we draw from the interfirm level of research, as opposed to the more appropriate interpersonal level, because this is the only research that we could locate that has addressed the relationship between total interdependence and trust. Specifically, in their research on dealer attitudes, Kumar, Scheer, and Steenkamp (1995) posited that increased total interdependence in the channel relationship would lead to increased trust. Their underlying logic being that as total interdependence intensifies, it becomes increasingly dangerous for either channel partner to engage in opportunistic behaviour, negative tactics, or coercion, because they have too much to lose. In such situations, due to self-interest, each partner has a strong motivation to build, maintain, and reinforce their relationship. Though Kumar, Scheer, and Steenkamp (1995) argue that high interdependence by itself will not directly create trust, they suggest that it is likely to lead to an intrachannel environment in which trust can be cultivated and flourish because of the convergence of the partners’ interests. More recently, another interorganizational study (Smith & Barclay, 1999) provides further and more direct empirical evidence to show that greater interdependence between two selling partners leads to greater mutual trust in their working relationships. So, using these two studies as the basis of our justification, it is hypothesized that: H6b: Greater total interdependence between the sales manager and the marketing manager will lead to greater cognition-based trust. H6c: Greater total interdependence between the sales manager and the marketing manager will lead to greater cognition-based trust. 10



Perceived relationship effectiveness. Finally, we expect total interdependence to positively affect perceived relationship effectiveness. The justification for this assertion is based on the work of Smith and Barclay’s (1999) research on selling partners where they found strong support for the hypothesis that greater interdependence led to greater perceived relationship effectiveness. Part of the reasoning behind this viewpoint was that interdependence is thought to increase group cohesion, commitment, and other positive group outcomes. Accordingly, we hypothesize: H6d: Greater total interdependence between the sales manager and the marketing manager will lead to greater perceived relationship effectiveness.

Method In order to provide a common general context for the respondents, we asked them to focus on a specific, major cross-functional project in which both the respondent and the marketing manager had had significant involvement in during the previous 18 months. Also, respondents were told to nominate a project which involved staff from at least two other functional areas. The bulk of these self-nominated projects (54.2%) related to new product development, whilst the remaining 45.8% covered a wide range of activities including: promotion and public relations, assessing key processes within the firm, and producing a business plan. On average, 4.36 functional units were involved in the nominated projects. Data Collection Data was collected from Australia and the UK using an identical pre-tested, self-administered mailed questionnaire. The sampling frame in each country was generated from a proprietary mailing list of firms purchased by each of the two authors in their respective countries. The criteria for inclusion in the sampling frame were firstly that the company should have an identified (named) sales manager or senior sales executive. Secondly, there must also be a named marketing manager, or senior marketing executive. Executives who had dual responsibilities, e.g., ‘sales and marketing managers’ were excluded from the sample. In the UK, the final sample of consisted of 716 firms, while the Australian sample consisted of 325 firms. After one follow-up reminder letter, 201 questionnaires were returned. However, one of these was deemed unusable. Moreover, because it was anticipated that the proprietary mailing lists may not have been as accurate as the providers claimed, a stamped, self-addressed reply paid card was attached to each follow-up questionnaire so as to facilitate a reply. Knowing the reason why some respondents in our sampling frame did not respond, allowed us to take this into account when determining the net response rate. The card simply asked the respondent to choose one of five categories to represent the best reason for not completing the questionnaire. After allowing for the 92 organizations which had returned the postcards, we calculated our overall response rate to be 20.3%. Sample Characteristics The sampling unit for this dyadic research is the sales manager. We did this for two reasons. First, since three constructs (threats, legalistic pleas, and manifest influence) regarding the influence of the marketing manager form a key part of our model, using sales managers to provide the data is more acceptable on methodological grounds. This is because it is now generally accepted that influence is best assessed by peer reports rather than by self-reports since the latter measures are prone to selfinflation bias (McQuiston & Dickson, 1991). Second, sales managers were chosen because we thought that this might lead to an increased response rate. Our reasoning being that in the two countries where data was collected (i.e., the UK and Australia), sales managers are much less likely to be the target of academic research, as compared to marketing managers. Since the focus of this research is business marketing, the sixty firms that operated exclusively in consumer markets were omitted from the analysis. As a result, the final sample of firms for this research study consisted of 131 firms. Eighty of these firms (61.1%) operated exclusively in business

11



markets while the remaining fifty one firms (38.9%) operated in both business and consumer markets. The average size of the firm was 2002 employees. Evaluating the Quality of the Data Collected Test of nonresponse bias. Tests of nonresponse bias were conducted and indicated that there were no significant differences between the early and late respondents in terms of six variables. Three of these variables (positional level; level of education; and, amount of marketing training) relate to the individual, while the other three variables (goods vs service firms; the number of other divisions the sales force sells products for; and, whether the firm is a single entity or part of a corporation) are organizational characteristics. Test of key informant competence. On average, the sales managers had worked for 11.6 years in their firm, which suggests that our respondents were experienced and knowledgeable about the issues covered in this research. The average duration of the working relationships between the two managers was 3.8 years. Operational Measures Whenever possible, we used existing measures and adapted them for our study context. The survey instrument included three types of measures: formative multi-item; reflective multi-item; and, singleitem. When the construct was a summary index of observed variables, a formative measurement model was used. In such situations, the observed variables cover different facets of the construct and cannot be expected to have significant inter-correlations. However, if the observed variables were manifestations of underlying constructs, we used a reflective measurement model. Importantly, when this is the case, the scales’ psychometric properties can be assessed by means of criteria based on confirmatory factor analysis (Anderson & Gerbing, 1988). Formative multi-item measure. One such measure was namely, total interdependence, and it was assessed by using 6 items (Fisher, Maltz, & Jaworski, 1997). Specifically, each sales manager, in terms of how he/she could accomplish his/her own goals and responsibilities during the project, was asked to rate how dependent he/she was on the marketing manager for: (a) resources e.g., information, (b) support e.g., advice, and (c) outputs e.g., reports. Next, each sales manager was asked to rate how dependent the marketing manager was on him/her for (a) resources, (b) support, and (c) outputs. Following previous researchers in the marketing channels area (e.g., Kumar, Scheer, & Steenkamp, 1995), we conceptualised total interdependence as a multidimensional composite index and summated the scores for the six items. Accordingly, we assumed that each item represented a dimension of interdependence. Multi-item reflective measures. Eight such measures were used: power of the marketing department; stakeholding; threats; legalistic pleas; cognition-based trust; affect-based trust; manifest influence; and perceived relationship effectiveness. The measures employed were adapted from existing scales and the details for each scale can be found in the Appendix. Metric Equivalence Early inspection of the means, standard deviations, and alpha coefficients of the key constructs suggested a very close similarity between the UK and the Australian data. However, in order to confirm this similarity, and to justify the pooling of the two datasets for measure refinement and model testing, we compared the UK and Australian data using multivariate analysis of variance (MANOVA: GLM procedure of SPSS, 1999). This analysis produced two key outputs. First, no departure from multivariate variance homogeneity was detected (Box's test: F45, 28278 = .492, p = .998). Second, the MANOVA results strongly suggest that there were no significant differences between the UK and Australian datasets (F9, 94 = .352, p = .954). Accordingly, we felt confident in pooling these two data sets. Measure Refinement After data collection, the reflective multi-item measures were first subjected to exploratory factor analysis. Then, confirmatory factor analysis (CFA), within the structural equation modelling program of AMOS

12



Version 4 (Arbuckle & Wothke, 1999), was used to evaluate the internal and external consistency of these measures. As part of this procedure, several measures of overall fit are used to evaluate how well the CFA model reproduces the observed variables’ covariance matrix. For example, the goodness-of-fit index (GFI) is a descriptive overall fit measure for which a minimum of .9 is usually considered acceptable. Another measure is the comparative fit index (CFI). In contrast to the GFI, this is an incremental fit index where again a minimum value of .9 is viewed as being acceptable. The root mean squared error of approximation (RMSEA) is another commonly used fit measure where values of up to .08 are considered to indicate a reasonable model fit. So that we could maintain an acceptable ratio of observations to variables, we conducted the CFA in two stages. In the first stage, we analysed a 13-item, four-factor model containing power of the marketing department (4 items), threats (3 items), legalistic pleas (3 items), and stakeholding (3 items). In order to detect possible model misspecification, AMOS yields two types of information— standardized residual covariances and modification indices (Byrne, 2001). On the basis of these two types of information, and in order to improve construct validity, we deleted one item from stakeholding. The revised 12-item model produced an acceptable fit with a chi-square of 80.839 (df = 47, p = .002), GFI = .91, CFI = .96, and a RMSEA = .074. In the second stage, we analysed an 18-item, four-factor model containing manifest influence (5 items), cognition-based trust (5 items), affect-based trust (3 items), and perceived relationship effectiveness (5 items). After inspecting the modification indices and the standardized residual covariances for crossloadings greater than 2.58, we decided to delete two items from the manifest influence measure. The revised 16-item model produced an adequate fit with a chi-square of 125.854 (df = 95, p = .019), GFI = .90, CFI = .98, and a RMSEA = .050. Reliability analysis reveals that the alpha coefficients for all the resultant scales are .79 or higher (see appendix), which suggests that for each construct, there is a reasonable degree of internal consistency between the corresponding indicators. The descriptive statistics and correlations for the key constructs are shown in Table 1. Overall, these measurement results are satisfactory and suggest that it is appropriate to proceed with the evaluation of the structural model and to be able to draw the results into a meaningful discussion.

Table 1. Descriptive statistics Scale Mean

S.D.

1

2

1. Power of the Marketing Department

15.32

5.83

2. Stakeholding of the MM

5.79

3.08

.09

3. Interdependencea

26.25

6.28

.12

.25**

4. Threats

5.51

3.28

.23**

.24**

13

3

4

5

6

7

8

.12



Scale Mean

S.D.

1

2

3

4

5

6

7

5. Legalistic Pleas

6.44

3.81

.27**

.27**

.10

.74**

6. Cognition-based Trust

27.00

6.52

-.07

-.09

.20*

-.28**

-.24**

7. Affect-basedTrust

15.68

5.24

-.08

-.00

.17

-.28**

-.27*

.79**

8. Manifest Influence

14.22

3.36

.23**

.24**

.35**

.31**

.18*

.14

.13

9. Perceived Relationship Effectiveness

26.80

6.67

-.11

-.10*

.23**

-.38**

-.36**

.69**

.69**

8

.19*

a

Denotes a formative indicator * Pearson correlation coefficients significant at the 0.05 level (two-tailed test). ** Pearson correlation coefficients significant at the 0.01 level (two-tailed test).

Model Estimation AMOS Version 4 was used to estimate the model using structural equation modelling (SEM), with observed variables. Recognition of the reliability of AMOS computations has been established by its increasing use in published studies in reputable journals over the last few years. Prior to model estimation, each of the multi-item constructs were represented by summated scores using equally weighted scales developed from the results of the CFA. This path analytic procedure was used due to the complexity and difficulty of using a full structural equation model. For a similar use of this technique, see Li and Calantone (1998, p. 88) and the references cited by these authors to justify this approach. Results The structural model was assessed using established measures and evaluative criteria for model fit. The results suggest that the data fit our conceptual model well, with a χ2 of 16.569 (df = 16, p = .419), GFI = .973, CFI = .999, and RMSEA = .017. Furthermore, the squared multiple correlation for perceived relationship effectiveness is equal to .59, which indicates that the variables included in our model explain 59% of the variance in this key outcome variable. The results of the hypotheses testing are summarized in Table 2. Table 3 also provides a summary of the direct, indirect, and total effects of the independent variables on our dependent variable perceived relationship effectiveness.

Table 2. Construct structural model

14



Linkages in the Model

Hypothesis Number

Hypothesis Sign

Result

H H H H H H H H H H H H H

+ + + + + + + + +

.146** .318*** .339*** .773*** -.303*** -.060 .343*** -.259*** -.178* .711*** .281*** .234*** .093*

H H H H H H H

+ + + + + + +

.125* .044 .253*** .246*** .155** .042 .243***

Between endogenous variables Manifest influence Í PRE CBT Í PRE ABT Í PRE CBT Í ABT Threats Í CBT Threats Í ABT Threats Í Manifest influence Threats Í PRE Legalistic pleas Í Manifest influence Legalistic pleas Í Threats Total interdependence Í Manifest influence Total interdependence Í CBT Total interdependence Í PRE Exogenous Í Endogenous variables Stakeholding Í Manifest influence Stakeholding Í Threats Stakeholding Í Legalistic pleas Stakeholding Í Total interdependence Departmental power Í Manifest influence Departmental power Í Threats Departmental power Í Legalistic pleas Model diagnostics χ2 = 16.569, p = 0.414, df = 16 GFI = 0.973, AGFI = 0.923, CFI = 0.999 RMSEA = 0.017 *** Significant at ≤ 0.01 level (one-tailed test) ** Significant at ≤ 0.05 level * Significant at ≤ 0.10 level

Table 3. Determinants of perceived relationship effectiveness Construct

Direct Effect (1)

Indirect effect (2)

Total effect (1) + (2)

Manifest influence

.146

-

.146

Cognition-based trust

.318

.262

.580

Affect-based trust

.339

-

.339

Threats

-.259

-.146

-.405

-

-.314

-.314

.093

.177

.270

MMs stakeholding

-

-.012

-.012

Power of the marketing department

-

-.071

-.071

Legalistic pleas Total interdependence

Summary Overall, 17 out of the 20 hypotheses were found to be statistically significant. Consequently, very strong support is found for our model.

Notes

15



1.

Domain similarity refers to the degree to which two different individuals or departments share the same goals, skills, or tasks (Ruekert & Walker, 1987 p. 6).

Appendix Measures Construct

No. of Items

Reliability (alpha)

Perceived Relationship Effectiveness (adapted from Ruekert and Walker 1987)

5

.95

Manifest Influence (adapted from Kohli and Zaltman 1988)

3

.86

CognitionBased Trust (adapted from McAllister 1995)

5

.90

Affect-Based Trust (adapted from McAllister 1995) ‘Threats’ (adapted from Venkatesh, Kohli, and Zaltman 1995)

3

.95

3

.83

‘Legalistic Pleas’ (adapted from Venkatesh, Kohli, and Zaltman 1995)

3

.79

6

NA

Departmental Power (adapted from Kohli 1989)

4

.88

Stakeholding (McQuiston and Dickson 1991)

2

.90

Total Interdependence (Fisher, Maltz, and Jaworski 1997)

Operationalization Seven-point scale anchored by 1 ‘Completely Disagree’ and 7 ‘Completely Agree.’ Respondents were asked to rate: their satisfaction with the working relationship; their belief that the marketing manager (MM) carried out their responsibilities and commitments; the value of the time spent developing and maintaining the relationship; the MM’s response to feedback and advice; and, overall success of the working relationship. Seven-point scale anchored by 1 ‘Very Little’ and 7 ‘A Great Deal.’ How much importance other team members gave to the MM’s input*; how much influence the MM had on the selection of decision criteria; the rating of the various options; changes induced in preferences of other team members; and, impact upon the thinking of other team members*. Seven-point scale anchored by 1 ‘Completely Disagree’ and 7 ‘Completely Agree.’ Whether most people trust and respect the MM; whether the MM approaches his/her job with professionalism and dedication; whether the SM doubts the MMs competence and preparation; whether the SM can rely on the MM to not cause problems through careless work; and, whether other work associates consider the MM to be trustworthy. Seven-point scale anchored by 1 ‘Completely Disagree’ and 7 ‘Completely Agree.’ Whether the SM can freely share ideas, feelings and hopes with the MM; whether the SM can talk openly to the MM about work difficulties knowing that the MM will want to listen; and, whether the MM would respond constructively and with understanding to the SM’s problems. Seven-point scale anchored by 1 ‘Never’ and 7 ‘Very Often.’ Respondents were asked to rate how often the MM: used threats of retaliation for failure to comply; threatened to disrupt the decision-making process; and, used threats to discontinue specific benefits for non-compliance with his/her demand(s). Seven-point scale anchored by 1 ‘Never’ and 7 ‘Very Often.’ Indicated that s/he expected others to comply with him or her because of his or her job position; reminded us of our obligations as stipulated in our firm’s rules and procedures; and, made a point to refer to specific company policies when attempting to influence our actions. Seven-point scale anchored by 1 ‘Didn’t Need his/her Help at All’ and 7 ‘Completely Dependent on his/her Help.’ Respondents were asked to rate how dependent the SM was on the MM for resources; support, or outputs. Also, the SM was asked how dependent MM was on the SM for resources, support, and outputs. Seven-point scale anchored by 1 ‘Completely Disagree’ and 7 ‘Completely Agree.’ The functions performed by the marketing staff are generally considered by top management to be more critical than others; top management consider the marketing staff to be more important than others; the marketing staff tend to dominate staff from other areas in the affairs of the organization; and, the marketing staff are generally regarded as being more influential than others. Seven-point scale anchored by 1 ‘Completely Disagree’ and 7 ‘Completely Agree.’ Respondents were asked: If the results of this group project do not work out as well as they are supposed to, the MM will receive most of the blame; if the results of this group project do not work out as well as they are supposed to, the status of the MM in the firm will go down; and, if the group project works out well, the MM will receive most of the credit*.

References

16



Anderson, J. C. & Gerbing, D. W. (1988) Structural Equation Modeling in Practice: a review and recommended two-step approach Psychological Bulletin 103(3) pp. 411-423. Anderson, J. C. & Narus, J. A. (1990) A Model of Distributor Firm and Manufacturer Firm Working Partnerships Journal of Marketing 54(January) pp. 42-58. Arbuckle, J. L. & Wothke, W. (1999) AMOS 4.0 User’s Guide (Chicago, IL: SmallWaters Corporation). Byrne, B. M. (2001) Structural Equation Modelling with AMOS: basic concepts, application and programming (Mahwah, NJ: Lawrence Erlbaum Associates Publishers). Conrad, C. (1990) Strategic Organizational Communication: an integrated perspective (Fort Worth, TX: Holt, Rinehart & Winston, Inc.). Dawes, P. L., Lee, D. Y. & Dowling, G. R. (1998) Information Control and Influence in Emergent Buying Centers Journal of Marketing 62(July) pp. 55-68. Day, G. S. (1997) Aligning the Organization to the Market, In: D. R. Lehmann & K. E. Jocz (Eds) Reflections on the Futures of Marketing (Cambridge, MA: Marketing Science Institute) pp. 67-93. Fisher, R. J., Maltz, E. & Jaworski, B. J. (1997) Enhancing Communication Between Marketing and Engineering: the moderating role of relative functional identification Journal of Marketing 61(July) pp. 54-70. Frazier, G. L. & Summers, J. O. (1984) Interfirm Influence Strategies and Their Application within Distribution Channels Journal of Marketing 48(Summer) pp. 43-55. Homburg, C., Workman, J. P. Jr. & Krohmer, H. (1999) Marketing’s Influence Within the Firm Journal of Marketing 63(April) pp. 1-17. Hutt, M. D. (1995) Cross-Functional Working Relationships in Marketing Journal of the Academy of Marketing Science 23(4) pp. 351-57. Keys, B., Case, T., Miller, T., Curran, K. E. & Jones, C. (1987) Lateral Influence Tactics in Organizations International Journal of Management 4(September) pp. 425-37. Kohli, A. K. (1989) Determinants of Influence in Organizational Buying: a contingency approach Journal of Marketing 53(July) pp. 50-65. Kumar, N., Scheer, L. K. & Steenkamp, J. E. M.(1995) The effects of perceived interdependence on dealer attitudes Journal of Marketing Research 32(August) pp. 348-356. Lewis, J. D. & Wiegert, A. (1985) trust as a social reality Social Forces 63 pp. 967-985. Li, T. & Calantone, R. J. (1998) The Impact of Market Knowledge Competence on New Product Advantage: conceptualisation and empirical examination Journal of Marketing 62(October) pp. 13-29. McAllister, D. J. (1995) Affect- and Cognition-Based Trust as Foundations for Interpersonal Cooperation in Organizations Academy of Management Journal 38(1) pp. 24-59. McQuiston, D. H. & Dickson, P. R. (1991) The effect of perceived personal consequences on participation in organizational buying Journal of Business Research 23(September) pp. 159-177.

17



Moorman, C., Deshpandé, R. & Zaltman, G. (1993) Factors Affecting Trust in Market Research Relationships Journal of Marketing 57(January) pp. 81-101. Morgan, R. M. & Hunt, S. D. (1994) The Commitment-Trust Theory of Relationship Marketing Journal of Marketing 58(July) pp. 20-38. Patchen, M. (1974) The Locus and Basis of Influence on Organizational Decisions Organizational Behavior and Human Performance 11(April) pp. 195-221. Perrow, C. (1970) Departmental Power and Perspectives in Industrial Firms, In: M. N. Zald (Ed) Power in Organizations (Nashville, TN: Vanderbilt University Press) pp. 59-89. Pfeffer, J. & Salancik, G. R. (1978) The External Control of Organizations: a resource dependence perspective (NY: Harper and Row). Ruekert, R. W. & Walker, O. C. (1987) Marketing’s Interaction with Other Functional Units Journal of Marketing 51(January) pp. 1-19. Smith, B. J. & Barclay, D. W. (1997) The Effects of Organizational Differences and Trust on the Effectiveness of Seller Partner Relationships Journal of Marketing 61(January) pp. 3-21. Smith, B. J. & Barclay, D. W. (1999) Selling Partner Relationships: the role of interdependence and relative influence Journal of Personal Selling & Sales Management 19(Fall) pp. 21-40. Venkatesh, R., Kohli, A. K. & Zaltman, G. (1995) Influence Strategies in Buying Centers Journal of Marketing 59(October) pp. 71-82. Workman, J. P. Jr., Homburg, C. & Gruner, K. (1998) Marketing Organization: an integrative framework of dimensions and determinants Journal of Marketing 62(July) pp. 21-41. Yukle, G. (1981) Leadership in Organizations (Englewood Cliffs, NJ: Prentice Hall Inc.). Yukl, G. (1990) Skills for Managers and Leaders (Englewood Cliffs, NJ: Prentice Hall Inc.). Yukle, G., Falbe, C. M. & Youn, J. Y. (1993) Patterns of influence behavior for managers Group & Organization Management 18(March) pp. 5-28. Yukl, G. & Tracey, J. B. (1992) Consequences of Influence Tactics Used with Subordinates, Peers, and the Boss Journal of Applied Psychology 77(4) pp. 525-35.

18
