- Case of Corporate Banking

2008:077

M A S T E R ' S T H E S I S

Relationship Portfolio Management

- Case of Corporate Banking

Hoda Talebi

Luleå University of Technology Master Thesis, Continuation Courses

Marketing and e-commerce

Department of Business Administration and Social Sciences Division of Industrial marketing and e-commerce

2008:077 - ISSN: 1653-0187 - ISRN: LTU-PB-EX--08/077--SE

MASTER'S THESIS

Relationship Portfolio Management Case of Corporate Banking

Supervisors:

Dr. Peter Naude Dr. Amir Albadvi

Referees:

Dr. Mohammad Aghdasi Dr. Mohammad Reza Amin Naseri

Prepared by:

Hoda Talebi

Tarbiat Modares University Faculty of Engineering Department of Industrial Engineering

Lulea University of Technology

Department of Business Administration and Social Sciences Division of Industrial Marketing and E-Commerce

Joint MSc PROGRAM IN MARKETING AND ELECTRONIC COMMERCE

2007

1

Abstract

After about three decades, it is a few years that Iranian banking industry is experiencing a more competitive market after the entrance of private banks to the market. Different objectives of a bank can be summarized to the optimized resource allocation between different customers. For achieving this objective, a very first step is defining a framework that can distinguish between different categories of customers. For developing such a framework, three main areas were identified within literature. These three areas include Relationship Portfolio Models, Network Theory and Business Banking Relationships. After reviewing these three areas, an initial model was extracted that was examined by several interviews with banking experts in Parsian Bank, Iran. After finalizing the model, a small sample group of corporate customers were chosen and relevant data were gathered. By using the MCDM (Multi-Criteria Decision Making) technique of PROMETHEE this customers were ranked based on the variables and their relevant indicators.

Keywords

Banking, Portfolio theory, Relationship management, Network theory, Business-

to-business marketing, Corporate

^banking

2

Table of Contents

List of Tables ………. 3

List of Figures ………. 4

Chapter 1: Introduction ……….. 5

1.1 Research Objective ……….. 6

1.2 Research Design ……….……. 6

1.3 Contributions ………... 7

1.4 Structure of the Thesis ………. 7

Chapter 2: Literature Review ………. 9

2.1 Relationship Management……….……… 10

2.1.1 Relationship Management Tools

……….. 12

2.1.2 Portfolio Theory

……….. 12

2.1.3 Marketing Portfolio Models

………... 16

2.1.4 Customer Portfolio Models

……….……….. 17

2.2 Network Theory ……….……….. 28

2.2.1 Network Portfolio Models

……….…….. 32

2.3 Business Banking Relationships………. 35

2.4 Initial Model ………. 49

Chapter 3: Research Design ……… 65

3.1 Research Approach ……… 67

3.2 Research Purpose ……… 68

3.3 Research Strategy ……… 70

3.4 Data Collection ……….……… 72

3.4.1 Collecting Primary Data Using Semi-Structured Interviews

………… 73

3.4.2 Sampling

……….……… 75

3.4.3 Validity and Genralizability

……….…… 78

3.4.4 Reliability

……….……… 79

3.4.5 Collecting Secondary Data

……….…… 79

Chapter 4: Data Analysis and Results ……… 80

4.1 Exploratory Study ……… 80

4.1.1 Interviewees’ Characteristics

……….……… 81

4.1.2 Qualitative Analytical Procedure

……… 83

Chapter 5: Conclusions and Implications ……….……… 99

5.1 Conclusions ……… 100

5.2 Descriptive Analysis ……… 104

5.3 Contributions ………. 109

5.3.1 Theoretical Contribution ……… 110

5.3.2 Methodological Contribution ……… 110

5.4. Managerial Implications ……… 111

5.5 Limitations ……… 111

5.6 Further Research ……… 112

References ……… 113

3

List of Tables

Table 2.1: A Summary of Important Customer Portfolio Models

..………

19 Table 2.2: Marketing Financial Services to Businesses

…….…….…….…….…

40 Table 2.3: Indicators of Customer's Business Attractiveness

…….…….………

50 Table 2.4: Indicators of Strategic Importance of the Account

…….…….……..

54 Table 2.5: Indicators of Complexity and Difficulty of Managing the Account

…..

57 Table 2.6: Indicators of Relative Stage of the Present Buyer-Seller Relationship

…

60 Table 3.1: Uses of Different Types of Interview

…….…….…….…….…….…..

74 Table 4.1: Distribution of Different Interviewees

…….…….…….…….………

83 Table 4.2: Interviewees Opinions about Indicators of Customer's Business

Attractiveness

…….…….…….…….…….…….…….…….………

84 Table 4.3: Some Points about Indicators of Customer’s Business Attractiveness

…..

86 Table 4.4: Additional Points about Customer’s Business Attractiveness

…….…….

89 Table 4.5: Interviewees Opinions about Indicators of Strategic Importance of the Customer

…….…….…….…….…….…….…….…….…….…….

90

Table 4.6: Some Points about Indicators of Strategic Importance of the Customer

…

⁹² Table 4.7: Additional Points about Strategic Importance of the Customer

…….…

92 Table 4.8: Interviewees Opinions about Indicators of Complexity and Difficulty of Managing the Account

…….…….…….…….…….…….………

93

Table 4.9: Some Points about Indicators of Difficulty of Managing each Customer

…

⁹⁴ Table 4.10: Additional Points about Difficulty of Managing each Customer ……… 95 Table 4.11: Interviewees Opinions about Indicators of Relative Stage of the

Present Buyer-Seller Relationship

…….…….…….…….…….…...

95 Table 4.12: Some Points about Indicators of Relative Stage of the Present

Buyer-Seller Relationship

…….…….…….…….…….…….…….

97 Table 4.13: Additional Points about Relative Stage of the Present Buyer-Seller

Relationship

…….…….…….…….…….…….…….…….…….

98 Table 5.1: Finalized Indicators of Customer’s Business Attractiveness

…….…….

100 Table 5.2: Finalized Indicators of Strategic Importance of the Customer

…….…..

102 Table 5.3: Finalized Indicators of Complexity and Difficulty of Managing each

Customer

…….…….…….…….…….…….…….…….…….…….

102 Table 5.4: Finalized Indicators of Relative Stage of the Present Buyer-Seller

Relationship

…….…….…….…….…….…….…….…….…….…

103

4

List of Figures

Figure 1.1: Structure of the Thesis ……….

8

Figure 2.1: Evolution of Portfolio Models ………..

¹⁴

Figure 3.1: Forms of Qualitative Interview ………

75

Figure 3.2: Sampling Techniques ………

76

Figure 4.1: Organizational Chart of Parsian Bank ………

82

Figure 5.1: Data Entry to the Model for a Sample of Customers ………….

107

Figure 5.2: Partial Ranking of the Sample Data …….…….…….………

107

Figure 5.3: Complete Ranking of the Sample Data …….…….…….………

108

Figure 5.4: GAIA Plane of the Sample Data ………..

109

Figure 5.5: Stability Intervals …….…….…….…….…….…….…….……

109

5

Chapter 1 Introduction

1. Introduction

After the Islamic Revolution in 1979, all Iranian banks were nationalized. In 2001, Iranian

banking market was re-opened to private sector. Entry of private banks made the market more

competitive and growing. Reports of the Central Bank of I.R. Iran reveal enormous growth in

terms of total assets of banks and their loan receivables from private sector, banks' capital and

developing legal and contingency reserves to improve capital adequacy ratio and developing

operations

^,

deposits of the whole banking industry and loan receivable balances of the whole

industry. At the same time, entry of foreign banks to the banking market of Iran is not that far

while nationalization and lack of competition, limited interactions with global banking market

and different sanctions being imposed on Iran during last three decades have weakened Iranian

6

banking industry. These weaknesses will cause serious problems for Iranian banks encountering their strong foreign competitors and the growing banking market in the country; Iranian banks should become ready for the coming competitive environment. Developing competitive tools which consider the special characteristics of Iranian banking market seems essential for Iranian banks.

One of these competitive tools is accurate frameworks and models that can help banks to discriminate between different groups of their corporate customers and allocate their resources to an optimal combination of these customers.

1.1 Research Objective

The focus of this study is on classifying corporate customers in Iranian corporate banking environment and aims at presenting a model that enables bank managers to classify their corporate customers into an optimal combination. So the defined objective for this research project can be formulated as:

Developing a model for classifying relationships between corporate banks and their customers in an optimized way with the consideration of Network Approach

For achieving this aim, three main areas of literature have been identified as Relationship Portfolio Models, Network Theory and Business Banking Relationships which will be explored through chapter two and lead to an initial model. This model will be examined and validated by a series of semi-structured interviews with banking experts.

1.2 Research Design

This research project will follow the inductive approach for doing an exploratory study. Data collection technique that will be used is qualitative semi-structured interviews for performing a case study in a selected bank (Parsian Bank, Iran).

Sample of interviewees is chosen on a non-probability basis and in a purposive

7

manner within banking experts whose activities are related to corporate banking or have the relevant knowledge.

1.3 Contributions

This study contributes to both theory and practice. Findings will help to the development of theoretical constructs of a framework for categorizing corporate customers of a bank while at the same time provide guidelines for Iranian bank managers to allocate their limited resources in an optimal way.

1.4 Structure of the Thesis

This thesis consists of five chapters, as shown in Figure 1.1. In this chapter an

introduction to the research was given and research objective was clarified. In the

second chapter, relevant theoretical areas and literature is presented. In the third

chapter, research design appropriate for achieving the defined objective is being

explored. In the fourth chapter, data is being analyzed and finally, in chapter five,

conclusions as well as contributions and implications in addition to

recommendations for further research are being brought up.

8 Figure 1.1: Structure of the Thesis

Chapter One Introduction

Chapter Five

Conclusions and Implications Chapter Three Research Design

Chapter Two Literature Review

Chapter Four

Data Analysis and Results

9

Chapter 2

Literature Review

2. Literature Review

Research objective of this thesis is developing a Portfolio Model in the context of Iranian corporate banking with the aim of categorizing corporate customers of a bank in an optimal way to make the maximum use of limited resources of the bank.

Three main areas of literature related to the topic of this research have been identified.

These areas include Relationship Portfolio Models, Network Theory and Business Banking

Relationships. These three fields are being investigated in coming sections of this chapter.

10

2.1 Relationship Management

The main field of this research is Industrial Marketing and focus is being on the concept of Relationship Management in business markets, in general and corporate banking, in particular.

The concept of Relationship Management and the fact that all interactions between companies are not purely transaction-based was developed by the European IMP Research Group (Hakansson 1982); since then, customer and supplier relationships have attracted a great deal of attention from a wide range of perspectives, including business-to-business marketing (Dwyer et al. 1987; Ford 1990, 1997, 1998; Frazier et al. 1988; Gummesson 1987, 1999; Hakansson 1982;

K. Moller and Wilson 1995; Morgan and Hunt 1994; Turnbull and Valla 1986). It has been argued that Relationship Management is as important to marketing management as manipulating the marketing mix (Hakansson 1982; Pick 1999;

Turnbull and Cunningham 1981); it is important to all parties involved - buyers, sellers and any intermediaries (i.e. the network) (Zolkiewski and Turnbull 2002), and both individually and collectively is the most critical marketing challenge, particularly in a business-to-business situation because:

• Customers and suppliers are often an organization's greatest asset (Campbell and Cunningham 1983b; Zolkiewski and Turnbull 2002).

• In business markets, firms are often reliant on a small number of customers and suppliers (M. T. Cunningham and Homse 1982; M. T.

Cunningham 1982) and industrial concentration is high in many industrial markets (Campbell and Cunningham 1983b; Zolkiewski and Turnbull 2002).

• In business-to-business situations, markets are relatively static and maintaining existing relationships is often essential to on-going business success (Ford et al. 1998).

• In mature markets it is often difficult to gain new customers (consequently lost business cannot be easily replaced) (Campbell and Cunningham 1983b; Zolkiewski and Turnbull 2002).

• Understanding customers' needs and/or suppliers' capabilities aids new product development and innovation (Campbell and Cunningham 1983a).

• Choosing the right supplier relationships will give next-generational

competitive advantage (Sheth and Sharma 1997).

11

An implicit assumption, however, of much of the works about Relationship Management is that having "strong" customer or supplier relationships is necessarily

"good". When this assumption is stated explicitly it is immediately and obviously not so - as any production, sales or customer account manager knows (Zolkiewski and Turnbull 2002). On the other hand, according to Leek, Turnbull and Naude, the findings indicate that the great majority of companies have some form of relationship management system but of varying degrees of formality. Formal documented systems, personal judgment and meetings are the most likely methods to be used for managing relationships. A formal documented system provides a relatively objective view of the relationship which takes into account variables such as sales volume, profitability and etc. Suppliers use it to assess the sales and profitability of their relationships whereas buyers use it to assess technological competence, capability and innovativeness of their suppliers.

Personal judgment takes into account the subjective relationship variables and the contextual variables such as relationship atmosphere, the importance of the relationship, company level variables, business environment variables and a number of additional intangible factors. Meetings allow the individuals to discuss problematic relationships and exchange information they have obtained from the use of the formal documented system and their own judgment. These are often used concurrently to enable information to be gathered from a variety of sources.

The emphasis on each method differs between suppliers and buyers. Buyers generally found formal, documented systems and personal judgment more useful than suppliers who preferred meetings. Buyers may find it easier to implement a formal, documented system, as they have a number of objective and even quantifiable purchasing criteria that their suppliers have to meet. The suppliers’

situation is more complex; they have to satisfy both their customers’ needs and their own needs (Leek et al. 2002).

In the current study, Relationship Management is perceived as the process of

efficiently and effectively allocating resources to different kinds of relationships

(Leek et al. 2004) and the main purpose of this thesis is creating a model including

both of objective and subjective variables for segmenting, categorizing and

12

managing relationships with current and potential corporate customers of Iranian banks in an efficient and effective way. In the first coming part, different relationship management tools are being explored and the most suitable tool for achieving the defined purpose of this research project is being selected and looked at in the second and third parts.

2.1.1 Relationship Management Tools

There are different relationship management tools such as Customer Relationship Management (CRM), Key Account Management and Portfolio Analysis (Leek et al. 2002).

CRM has been recently became popular in consumer retail markets while many practitioners are now operationalizing the concept of Key Account Management in different business contexts including banking. Yet the selection of key accounts, preferred suppliers and key relationships can be problematic and this is where Portfolio Analysis and Portfolio Management make a major contribution to management (Zolkiewski and Turnbull 2002).

Among different aspects of relationship management, two themes of research are critically reviewed in this research: the first theme is Relationship Management in the context of Portfolio Analysis and the second one is the concept of Relationship Networks.

2.1.2 Portfolio Theory

The central tenet of Portfolio Approach is to enable managers to refocus from a product orientation towards a customer orientation and thus to invest their resources in the most efficient and effective way (Zolkiewski and Turnbull 2002).

Turnbull states that Portfolio Concept is a useful management tool for enforcing a

discipline in the allocation of the company’s limited resources to an optimal

combination of business operations which will maximize long-term returns at a

13

given levels of risks (Turnbull 1990); clearly, new and developing relationships will require different resource inputs compared to mature or declining relationships (Ford 1980).

In addition to resource allocation question, Portfolio Analysis can also act as an important source of knowledge about customers and suppliers. Thus, Portfolio Models can provide a framework for relationship management at both the macro/strategic level - planning the acquisition and development of new relationships and micro/tactical (operational) level -efficient management of the existing portfolios (Turnbull and Valla 1986; Zolkiewski and Turnbull 2002). In other words, all Portfolio Models claim that they will help managers to manage their current and potential customers optimally. In short, it could be said that Portfolio Analysis gives answers to these questions (Zolkiewski and Turnbull 2002):

• Do new relationships need to be created?

• Which relationships should be developed?

• Which maintained?

• Are there any that should be broken/discarded?

The development of customer and supplier portfolio models has, to date, largely been related to business-to-business markets. This is probably because of the small number of players in such markets (Zolkiewski and Turnbull 2002); it is common for a firm serving business markets to be highly dependent on a small number of customers (Turnbull and Cunningham 1981). Therefore, the addition or loss of a major customer or supplier can have dramatic effects on the company's turnover, profitability and its viability. In such circumstances, Portfolio Analysis can act as a very useful tool by identifying key strategic relationships (Zolkiewski and Turnbull 2002).

Portfolio Theory was first developed to be used in financial investment decision-making during the 1950s (Markowitz 1952) as a mechanism for reducing risk (see Figure 2.1 for the evolution of Portfolio Theory). The main inputs for portfolio evaluation in financial investment decisions were postulated as being

“expected return” and “degree of risk”. Markowitz's point of departure was that

14

rational investors will (or should) select portfolios which maximize the individual investor's utilities by maximizing the expected return for a given level of risk or minimizing the risk for a given level of expected return. It was further refined in the 1960s by Sharpe (Sharpe 1963), who suggested that the risk of an individual investment should not be seen in isolation, rather it should be viewed in terms of how it contributes to the overall balance and performance of an investment portfolio. It follows from this that risk can be decreased (and ultimately eliminated) by holding a widely diversified portfolio of securities (see (Yorke 1984), for a detailed discussion).

Figure 2.1: Evolution of Portfolio Models

Although the Portfolio Concept put forth by Markowitz was an instrument for the management of equity investments, the concept has been recognized to have viable applicability in other fields (Turnbull 1990). Portfolio Theory has since been applied in areas other than finance.

The next area of application of Portfolio Theory was in auditing product programs, where individual products or groups of products were analyzed in terms of their current and future market share, sales, volume, costs and investment requirements (Henderson 1970; Marvin 1972).

Subsequently, Portfolio Approach received increasing attention from corporate strategists (Abell and Hammond 1979; H. I. Ansoff and Leontiades 1976; BCG 1968; Hedley 1977; Hofer and Schendell 1978; Wind and Douglas 1981) all of whom have been primarily concerned with the classification of

1950 1970 1975 1980 1995 2000

Financial Investments

Auditing Product Programs

Corporate Strategies

Customer Portfolio Models Supplier Portfolio Models

Indirect

Portfolio Models

Marketing Strategies

15

products and/or businesses on certain key dimensions in order to assist in the achievement of corporate strategic objectives used by a wide range of multi- product, diversified organizations (see (Turnbull 1990) for a detailed discussion).

Key dimensions have included market share, market growth, market attractiveness and competitive position depending on which model has been offered. Regardless of the dimensions used, the basic idea is that the positions of the units on the grid, determine the formulation of the most appropriate strategy.

The next application of Portfolio Theory was in marketing planning and communication as differentiated selling/purchasing behavior (Dubois and Pederson 2002). During last two decades, there has been a steady flow of research concerning academic Portfolio Models in this field (Bensaou 1999; Campbell and Cunningham 1983b; Canning Jr. 1982; M. T. Cunningham and Homse 1982; M.

T. Cunningham 1982; Dickson 1983; Dubinsky and Ingram 1984; Dubinsky 1986;

Fiocca 1982; C.J. Gelderman and Van Weele 2000; C.J. Gelderman and Van Weele 2001; C.J. Gelderman and Van Weele 2003; Kraljic 1983; Krapfel et al.

1991; LaForge and Cravens 1982; Olsen and Ellram 1997; Rangan et al. 1992;

Shapiro et al. 1987; Turnbull and Zolkiewski 1997; Van Stekelenborg and Kornelius 1994; Wynstra and Pierik 2000; Yorke 1984; Yorke and Droussiotis 1994; Zolkiewski and Turnbull 2002). Application of Portfolio Theory in marketing is specifically addressing the allocation of resources to both channels of supply and distribution (i.e. suppliers and customers) (Zolkiewski and Turnbull 2002).

Models in this area consist of three main categories:

1. Customer portfolio models

2. Supplier (purchasing) portfolio models 3. Indirect portfolio models (network models)

In theory, marketers can check the basic soundness of each customer or

supplier against its position on the portfolio grid. In addition, they can assess the

mix and balance of these customers/suppliers and whether they are likely to meet

marketing objectives. Portfolio Analysis can therefore enhance and promote

marketing planning and communication (Yorke and Droussiotis 1994).

16

In overall, Portfolio Theory is essentially concerned with facilitating interdependent decisions (Wind and Douglas 1981) in the allocation of finite resources (human, technical and financial (Zolkiewski and Turnbull 2002)) among different assets, be it financial investments, products or strategic business units or marketing issues (Yorke and Droussiotis 1994).

In coming section, Marketing Portfolio Models are being explored in depth.

2.1.3 Marketing Portfolio Models

Since 1982, a variety of 2D and 3D Marketing Portfolio Models have been specifically developed to address customer and supplier relationship management utilizing single, two or three-step analysis phases with differing numbers of relationship categories being created. The models not only differ in the variables they use but also in their construction which consequently leads to differing numbers of relationship categories being created; for example Turnbull and Zolkiewski’s model identifies eight categories of relationships (Turnbull and Zolkiewski 1997) while both Shapiro et al.’s model and Krapfel et al.’s model produce four types of relationship (Krapfel et al. 1991; Shapiro et al. 1987).

As it was mentioned earlier, these models can be divided into Customer Portfolio Models, Supplier Portfolio Models and Indirect Portfolio Models (Network Models).

Most of these models have taken the dyadic relationship as the unit of analysis and

are based, at least implicitly, on an understanding that long-term, interactive

relationships are often the norm in business-to-business markets (Zolkiewski and

Turnbull 2002). The principle objective of these models is to categorize the

relationships in order to enable the managers to determine a strategy for

improving their business, but only a few of the models (Krapfel et al. 1991; Olsen

and Ellram 1997; Shapiro et al. 1987) provide any guidance on what to do with the

information that results from carrying out the Portfolio Analysis. The action plans

suggested by the models are inevitably vague as without the details of a specific

17

company, its aims and its relationships it is impossible for academics to devise models that provide prescriptive advice (Leek et al. 2002).

An important question that managers need to address is which model is appropriate and at which level of their business should they be applied: SBU, market segment or individual customers and/or suppliers (Zolkiewski and Turnbull 2002)? For answering these questions within the defined boundaries of this research, Customer Portfolio Models are being investigated in the coming section and Indirect Portfolio Models will be investigated after discussing about Network Theory.

2.1.4 Customer Portfolio Models

As mentioned before, the main focus of this research project is the relationship between banks and their corporate clients; emphasis of this part is being on Customer Portfolio Models.

Since the customer becomes the core of the analysis in industrial marketing strategy, it can be convenient for the selling company to divide its business among accounts, rather than among products or product lines (Fiocca 1982). Several models have been presented by academics with the aim of categorizing the relationships with customers in order to enable the managers to determine a strategy for improving their business (A summery of these models is presented in Table 2.1).

The pioneering Customer Portfolio Model belongs to Cunningham and Homse

who conceptualized a one-step model with two dimensions of technical

interaction and sales volume (M. T. Cunningham and Homse 1982). Their model

was improved by the three-step model of Campbell and Cunningham (Campbell

and Cunningham 1983b). The first step of this model focuses on the nature and

attractiveness of all customers and categorizing them into different life cycle stages

18

by using various data. The second step of analysis focuses on the customer's own performance as an important aspect of customer portfolio planning of the key customers. And at last, the third and final step involves the selection of the key customers for analysis. Step three is, in fact, broken down into two sub-steps where first, all the key customers are analyzed together and then the most important are analyzed individually. Main criticisms to this model include:

1. The conceptual validity and practicality of using a life cycle approach to customer analysis can be challenged (Ford 1982; Gronroos 1983;

Zolkiewski and Turnbull 2002).

2. Although the implicit assumption of any customer portfolio model is identifying the key customers, the input of the third step of Campbell and Cunningham's model are the customers whose selection criteria has remained vague by the composers of the model.

Another significant Customer Portfolio Model is the Fiocca model (Fiocca 1982). This two-step model firstly analyzes customers at a general level according to the strategic importance of, and the difficulties in managing the relationship with each customer (account). The second step of analysis requires another two- dimensional matrix to be constructed for the key accounts identified in step one, with the customers' business attractiveness on one axis and the strength of the supplier-customer relationship on the other. Main criticisms to this model include:

1. Not using customer profitability. In fact Fiocca used his second matrix simply to infer that different cells of the portfolio matrix could be associated with different levels of profitability (Yorke and Droussiotis 1994).

2. Problems with data calculation (Turnbull and Topcu 1994); subjective axes scales and use of different measures for each of the two dimensions.

3. Different interpretations from variables (Turnbull and Topcu 1994).

19 Table 2.1: A Summary of Important Customer Portfolio Models (continued)

Author Year No.

of steps

No. of Dimensions in Each Step

Action

Plans Unit of analysis Suggested dimensions

Cunningham

and Homse 1982 1 2 NA NA Technical interaction and sales volume

Fiocca 1982 2 2 Yes

Step 1: Accounts (customers)

Step 2: Relationships (e.g.

one product or one homogenous product group)

Step 1 (general): difficulty in managing the account and strategic importance of the account

Step 2 (key accounts): customer's business attractiveness and relative buyer-seller relationship

Dickson 1983 2

Step 1: 5 Step 2: 2

Yes

Step 1: brand specific level or a product group level sold to each customer Step 2: Manufacturer and distributors

Step 1 (distributor portfolio analysis): the distributor's rate of sales growth after adjusting for inflation, the

manufacturer's share of distributor's sales of a particular product or group of products, the manufacture's sale of the product or product group to each distributor, direct

manufacturing costs, and gross profit

Step 2 (Channel dependence matrix): Manufacturer's market share, distributor's/retailer's market share

20 Table 2.1: A Summary of Important Customer Portfolio Models (continued)

Author Year No.

of steps

No. of Dimensions in Each Step

Action

Plans Unit of analysis Suggested dimensions

Campbell and Cunningham

1983 3

Step 1: 5 Step 2: 5 Step 3: 3

No

Step 1: Customers of each market segment (one product or one

homogeneous product group)

Step 2: Customers of each market segment

Step 3: Key customers and sub-customers

Step 1 (general-life cycle classification of customer

relationships): sales volume, use of strategic resources, age of relationship, supplier's share of customer's purchases and profitability of customer to supplier

Step 2 (customer-competitor analysis): Growth rate of customer's demand for product or service, customer's share of its market, the volume of the supplier's product purchased by each customer, the share of each

competitor and power balance

Step 3 (portfolio analysis of key customers): Growth rate of customer's market, competitive position (the share the supplier holds of the customer's purchases relative to the share held by the largest competitor) and the sales volume of each customer

Dubinsky

and Ingram 1984 1 2 Yes Existing customers over recent planning periods

Present profit contribution (net sales to a particular customer - cost of goods sold - direct selling expenses of salesperson) and potential profit contribution

21 Table 2.1: A Summary of Important Customer Portfolio Models

Author Year No. of steps

No. of Dimensions in Each Step

Action

Plans Unit of analysis Suggested dimensions

Shapiro et al. 1987 1 3 Yes

Net profit (net price minus cost to serve) by customer and by order

Net price, cost to serve (presale costs, production costs, distribution costs and post-sale service costs) and customer buying behavior

Rangan et al. 1992 2

Step 1: 2 Step 2: 2

Just for the mentioned case

Customers of a product or one homogeneous product group)

Step 1: Price-service tradeoffs and buyer power Step 2: Price and cost to serve

Yorke and

Droussiotis 1994 3

Step 1: 2 Step 2: 1 Step 3: 2

Just for the mentioned cases

Step 1: All customers Step 2: All customers Step 3: Profitable customers

Step 1: Difficulty in managing the account and strategic importance of the account

Step 2: Profitability

Step 3: Perceived strength of the relationship and current percentage share of the customer's business held by the agency

Turnbull and

Zolkiewski 1997 1 3 No Customers of each

market segment Cost to serve, net price and relationship value

22

Yorke and Droussiotis undertook an empirical study to test and develop the Fiocca matrix using a Cypriot textile agency (Yorke and Droussiotis 1994).

Their model has three steps. In the first step, the same as Fiocca's model, two variables of strategic importance of the account and difficulty in managing the account are considered and in the second step, profitability of different customers is calculated and finally, in the third step, perceived strength of the relationship and current percentage share of the profitable customers' business held by the agency is considered. Mentioned weaknesses of this model by academics are:

1. This analysis was conducted over a very short timescale (two months) and the authors recognize that it may not be representative of the usual situation in the industry and the company (Zolkiewski and Turnbull 2000).

2. The way indirect and direct costs are allocated raises important questions; very often it is not easy to simply apportion management time and costs or even sales time and costs to a particular customer or contract (Zolkiewski and Turnbull 2000).

3. Variables for calculating profitability are too closely associated with one specific organization and are not easily transferable to comparable situations (Turnbull 1990).

Dickson developed a two-step model in 1983 focused on the relationship between a manufacturer and its distributors (Dickson 1983). In the first step, which is named Distribution Portfolio Analysis (DPA), the attractiveness of customer accounts as cost-profit centers is monitored. Five dimensions are considered for calculating customer attractiveness: the distributor's rate of sales growth after adjusting for inflation, the manufacturer's share of distributor's sales of a particular product or group of products, the manufacture's sale of the product or product group to each distributor, direct manufacturing costs, and gross profit.

In the second step, which is named Channel Dependence Matrix, channel members' shares of each others' sales is investigated to assess potential power.

Dubinsky and Ingram developed their one-step Customer Portfolio Model in

1984 (Dubinsky and Ingram 1984). This model has two dimensions: customers'

present profit contribution and their potential profit contribution and the main

23

goal is persuading marketing personnel to adopt a profit orientation rather than a sales volume orientation and assisting them to assess the current status and affect the desired balance in the composition of their customer bases. Profit contribution here refers to contribution margin, where Contribution margin of customer = (Net sales – Cost of goods sold – Direct selling expenses of salesperson). Present profit contribution is the current contribution margin a customer provides the vendor over a given planning period (such as a month, quarter, or year). Potential profit contribution margin is the contribution margin a customer is expected to provide in the future. To ascertain the potential profit contribution of each account, any expected changes in sales strategies or service levels, and concomitant costs must be considered. Addressing this issue is analogous to the budgeting process practiced in many firms. Future sales could be forecasted by traditional means, such as by surveying the customer or projecting past sales into the future. Cost of goods sold estimates may be derived from accounting or production records. Based on these two dimensions, customers are categorized into four main groups: undesirable accounts, undeveloped accounts, desirable accounts and developed accounts (Dubinsky and Ingram 1984).

The next model was developed by Shapiro et al. in 1987 (Shapiro et al.

1987). This model focuses on customers just as profit centers and they are categorized just based on their profitability which is calculated by subtracting costs (presale, production, distribution and post-sale service costs) from net price.

Shapiro et al. argue that profitability of customers is affected by their behavior, but the concept of buying behavior and its effects on profit remained vague in this model. Main weakness of this model is:

1. Leaving the interpretation of low and high values to the discretion of

the analyst, which could cause difficulty when comparable data sets are

required, especially if management make subjective judgments as to

these values (Zolkiewski and Turnbull 2002).

24

2. The manner in which pre and post-sale costs are recorded can prove to be extremely difficult to implement in a technically complex product context. The amount of time spent by R&D staff, sales engineers, managers etc., can be difficult to determine exactly or even approximately, especially if the relationships are long term. Also, the manner in which costs such as R&D and the preparation of detailed bids are apportioned is complex, as these costs are often directed towards the needs of both existing and potential customers (Zolkiewski and Turnbull 1999).

Rangan et al. improved Shapiro et al.'s model and developed a buying- behavior-based framework suitable for micro-segmenting customers in mature industrial markets before using the two-dimensional grid of price and cost to serve (Rangan et al. 1992). They argued that though the concept of buying-behavior- based segmentation (Webster and Wind 1972) had been recognized from two decades before the presentation of their model, few applications of the approach had been reported in the industrial marketing literature. The basic supporting idea for Rangan et al.'s model was that in contrast to the works before their model that attempted to uncover existing segments as a way to position products strategically, in their approach the assumption is that at a micro-segment level, firms can influence and shape the buying behaviors of their potential customers by altering marketing mix variables (e.g. price). This model, categorizes customers based on their buying behavior variables and then map them on the two dimensional grid of price and cost to serve. By using the information gained from the buying- behavior-based categorization they claim that a firm can justify price and cost in a way that customers become profitable.

Based upon Shapiro et al. and Krapfel, Salmond and Spekman matrices

(Krapfel et al. 1991; Shapiro et al. 1987), Turnbull and Zolkiewski proposed a one-

step three-dimensional basis for Customer Portfolio Analysis in 1997 (Turnbull and

Zolkiewski 1997). They argue that three-dimensional analysis based upon cost to

serve, net price and relationship value is appropriate when segmenting the

customers of any firm. They use these variables to analyze the case study data and

suggest that this three-dimensional approach is much more beneficial as a strategic

25

planning aid because it provides a more refined set of analysis criteria. Using three-dimensional analysis allows a more refined customer classification in terms of dimensions of analysis and number of classifications. For example, in this case, eight classifications are possible - compared to the four groups in the two-by-two matrices. These eight segments provide a mechanism for combining hard data (profitability of customers) with more judgmental data (relationship value).

Although Customer Portfolio Models are inherently appealing as a means of analysis, there are difficulties in their application and definition and using this tool causes some challenges. The main criticisms against portfolio models can be divided into technical issues and fundamental/meta-theoretical criticisms.

Technical issues are as follows:

1. The difficulty of choosing appropriate dimensions of analysis (Zolkiewski and Turnbull 2002). The variables used in each of the models vary and the inclusion or omission of certain variables may lead to criticism (Yorke and Droussiotis 1994; Zolkiewski and Turnbull 2000, 2002).

2. The difficulty of definition and measurement of the variables used (Zolkiewski and Turnbull 2002). These difficulties exist because of the subjectivity of variables, the imprecision of the scales proposed for axes and problems in calculating financial data.

Subjectivity of variables. Many of the variables used in the models are a mixture of actual, subjective and judgmental which creates problems with their definition and measurement (For further discussion see (Leek and Turnbull 2001)). It is these more judgmental variables which become the most difficult to use, partly because of the subjective nature of the calculation and partly because of the difficulty in replicating these calculations and also using in future comparisons. Other portfolio analyses, such as BCG, etc., also suffer from similar problems of subjectivity (Zolkiewski and Turnbull 2002).

Imprecision of the scales proposed for axes (Zolkiewski and Turnbull

2002). Generally the scales proposed for axes are imprecise; for

instance, what are low and high values? Again, such values implicitly

involve subjective judgments and therefore become more difficult to

assess. Such values can only be useful if a rough conceptual guide is

needed for sorting out the major customers from the mass of

customers.

26

Problems in calculating financial data (Zolkiewski and Turnbull 2002).

All the models include financial data and the calculation of this financial data can be problematic. For example, by investigating different models, it can be seen that profitability/value of relationships is one of the common themes of analysis (albeit in different guises). Profitability is determined from an accurate analysis of the real costs of serving a particular customer. The problem is that profitability is a very difficult variable to define and measure for most companies. Few companies have accounting systems that can provide relationship costs and profits.

Campbell and Cunningham recognized the importance of including customer profitability in any analysis and the difficulty associated with collecting such information (Campbell and Cunningham 1983b).

Shapiro et al. put forward a mechanism for calculating customer profitability (subtracting cost to serve from net price) (Shapiro et al.

1987), which is similar to that proposed by Dubinsky and Ingram (Dubinsky and Ingram 1984), while Yorke and Droussiotis use a very different set of variables for profitability (Yorke and Droussiotis 1994).

However, the Yorke and Droussiotis variables are too closely associated with one specific organization and are not easily transferable to comparable situations. Cost to serve is more operational - it simply requires the availability/collection of appropriate data, but while the allocation of direct and variable costs should be relatively simple, the apportionment of overhead costs can be extremely difficult. Although the cost to serve dimension suggested by Shapiro et al. includes pre- and post-sale costs, little guidance is given as to their calculation and, therefore, it can be extremely difficult to decide how R&D costs, etc., should be apportioned while many companies do not have adequate mechanisms for allocating indirect costs. In short, the real problem lies in the fact that the definitions simply do not involve easily collected

"hard" data and many organizations do not have mechanisms which allow them to calculate the real 'cost to serve' of individual customers or even market segments. Some models, like Krapfel et al.'s use the concept of relationship value instead of customer profitability. This variable includes a more strategic set of variables (e.g. switching costs, reliance on the buyer, etc.) some of which are subjective and difficult to quantify (Krapfel et al. 1991). Wilson and Jantraria analyzed the concept of value and concluded that the value can be seen narrowly as only economic value or in a broader way including also so called

“psychological” or “behavioral” and “strategic” value (Wilson and

Jantraria 1997). Despite the problems associated with subjectivity, such

calculations can be a very useful aid to management and can be

extremely useful in the process of prioritizing relationships. Thus, the

concepts of customer profitability and relationship value need careful

consideration, but, they should not be taken as an absolute; the context

of the relationship also needs consideration and although a particular

27

relationship or market segment may not be profitable in financial terms there may be good reasons to 'invest'. For example:

• The relationship is an integral part of competitive strategy, e.g.

entering a new market, an early stage relationship or to gain technological competencies.

• The relationship may be needed for R&D purposes.

Increasing scientific and technical specialism means that R&D is increasingly expensive and it is difficult for a single firm to acquire all the knowledge needed to be truly innovative.

• The unprofitability may be temporary as a result of environmental influences, e.g. strikes or fluctuations in exchange rates.

• The customer is attractive or strategically important.

3. Not considering macro environmental pressures (such as exchange rate fluctuations or government policy) (Zolkiewski and Turnbull 2002). Dubois and Pedersen argue that by simplifying the nature and context of exchange and not considering environmental pressures, Portfolio Models fail to capture the vital aspects of buyer- supplier relationships (Dubois and Pederson 2002).

4. Different units of analysis (Zolkiewski and Turnbull 2002). Suppliers may have a range of relationships within a single but multi-dimensional customer or a multidimensional supplier may have a range of relationships with a single customer. The question then is whether analysis should be at the individual relationship level or aggregated across the whole customer or supplier. Clarity about the unit of analysis is essential. Care must be taken to avoid simplistically classifying all dealings with a diversified customer or supplier as one relationship. The manager must make an informed decision as to whether to treat each relationship individually or to aggregate them at the level of the customer or supplier (Zolkiewski and Turnbull 2002). It is very probable that analysis of accounts place them in very different portfolio positions according to the level at which the analysis takes place. An illustration of this dilemma was demonstrated in: (Turnbull and Zolkiewski 1997; Zolkiewski 1994).

5. Dependency to historical data. In all existing Portfolio Models, part of needed variables in the model are related to the relationship history between buyer and seller and this issue is contradictory with the claim of helping managers to manage potential customers who do not have any relationship history.

6. Not depicting the interdependencies. This problem is about not

depicting the interdependencies between two or more items in analysis

i.e. they concentrate on categorizing products or relationships (Ritter

2000).

28

7. Visualization rather than analysis. There have been many critics

about this issue that a portfolio simply facilitates visualization rather than serving as an analytical and prescriptive tool in itself (Yorke and Droussiotis 1994).

8. General models. All existing Customer Portfolio Models are general models not focusing on any particular sector while all of them have production-based views and do not consider special characteristics of service sector.

Even with existing difficulties about Portfolio Analysis, academics believe that the construct is valid and potentially very valuable as it can provide significant guidance as to how resource optimization can be achieved. It also provides an excellent mechanism for the organization of information about customers and suppliers (Nellore and Soderquist 2000; Olsen and Ellram 1997), but it is said that all existing models result in simplistic analyses and multi-variable (or dimension) analysis may be more appropriate, despite the difficulty of visualization. The development of complex database and modeling softwares facilitates this mode of analysis (Zolkiewski and Turnbull 2002).

In short, despite of extensive theoretical and empirical work in the field of Relationship Portfolio Modeling, there is still much further work needed to clarify the most appropriate measurement/analysis of variables and the associated managerial process. Until this is done, the practicality of the concept must remain suspect (Leek et al. 2002).

2.2 Network Theory

Another evolving research theme along with Portfolio Theory is in the concept of Relationship Networks. This concept recognizes the impact that a variety of "actors"

in the network of relationships surrounding a focal firm can have on a firm's

strategy and tactics. In other words, the network perspective encourages strategic

vision by emphasizing the importance of all interactions within a network and

demonstrating the wider impact of many actions (Zolkiewski and Turnbull 2002)

29

by having inter-firm relationships as focal units of analysis rather than firms (Axelsson and Easton 1992; Hakansson and Snehota 1995; Iacobucci 1996).

In this research, the Interaction and Network Approach developed by the “IMP Group” in the context of industrial (manufacturing) systems (Axelsson and Easton 1992; Ebers 1997; Ford 1997; Ford et al. 2002a; Ford and Ritter 2004;

Hakansson 1982; Hakansson and Snehota 1995; Hakansson et al. 2004; Turnbull and Valla 1986) is being combined with the Portfolio Theory. The IMP Group’s way of looking business was different than before - the emphasis was not on atomistic markets or on discrete purchasing decisions. The IMP Group’s fundamental new insight was the idea that different actors were interdependent, their relationships evolve over time and different actors make adaptations creating long-term exchange relationships (Interaction Approach) (Turnbull et al. 1996). After a while, the focus of the IMP Group was moved into the network level even though it was stated that firms and relationships were the sine qua non of industrial networks and they must be studied if networks are to be understood (Easton 1992). In other words, the original IMP research was based essentially on dyadic relationships (Hakansson 1982) but later recognized the importance of analyzing and managing a diverse array of dyadic relationships in the network of relationships within which a supplier or purchasing company exists (Easton 1992;

Turnbull and Wilson 1989).

Traditional markets are being replaced by networks; firms build complex

relationships and the macro environment surrounding firms is getting more

complex thus the Network Approach is of increasing importance (K. Moller and

Wilson 1995; K. Moller and Halinen 1999). At the same time, several researchers

have linked competitiveness to a company’s capabilities to develop and manage its

array of network relationships (Turnbull et al. 1996). In such circumstances,

ignoring the network surrounding a firm is just not more than a simplistic

approach.

30

What makes the Network Management highly interesting is the interconnectedness of actors in the networks. Relationships do not exist independent from each other-they are interconnected because a given relationship does not only affect itself and the two actors involved; a relationship may also have an effect on other relationships. This has been called secondary, indirect or network function of relationships (Hakansson and Snehota 1995). Moreover, the actions of different actors may have effects in several levels: effects in the relationship, effects on the relationship, and effects on the portfolio of several relationships or effects on the whole network (Ford and McDowell 1999).

Hakansson and Snehota provide a description of the network perspective in which actors perform activities and/or control resources (Hakansson and Snehota 1995). They view networks as a combination of all the elements which they comprise (activity links, resource ties and actor bonds) and their functionality (dyad, single actor and network). They also acknowledge the effects on the network of changes which develop over time. Zolkiewski also suggested that the role of indirect relationships within a network (such as the influence of government on procurement decisions) need consideration (Zolkiewski 1999).

This suggests that managers must manage relationships in the context of the impact of these actions on other actors in the network (Turnbull et al. 1996).

However, all of these studies refer to the outcome of interconnectedness and not to interconnectedness itself (Blankenburg-Holm 1995; Blankenburg-Holm and Johanson 1997).

On the other hand, the network surrounding a company is difficult to define and delimit and it has no objective boundaries (Ford et al. 2002b). Views about manageability of networks can be separated crudely into two extremes (Ritter et al. 2002). The first extreme has its home in Strategic Network Approach.

This American-based view sees networks as a strategic entity managed by a hub

firm (Tikkanen 1998). However, much criticism has been presented against this

approach. One central theme of criticism is that any view of a network centered

on a single company, or defined by the company itself is inevitably restricted. Such

31

a view is often associated with the illusion that the company controls that network or more simply that it is company’s own network (Ford et al. 2002a). The second extreme sees that the firm level is not appropriate for analysis as networks occur and evolve “of their own” and thus are unmanageable (Ritter et al. 2002). From this point of view, networks are seen independent, organism-like, evolving and self-organizing systems. However, most writers conclude that management of networks is possible in some extent (the IMP Group approach). This is based on the fact that even if the management of networks can be seen impossible, a firm can still manage in networks (Ford et al. 2002b; Ford et al. 2002a; Hakansson and Ford 2002). The IMP tradition has little sympathy with an understanding of the networks as a governance structure in the sense of an imposed structure of a dominant organization. Rather, governance is achieved through relationships and the network is a way of understanding the generalized connectedness that prevails through relationships. In addition, networks are not viewed as a priori structures to be imposed on organizations. Instead, they are considered as structured by the enactment of selective ties and relationships between autonomous actors (McLoughlin and Horan 2002). This is also the accepted point of view throughput this research.

But the IMP Network Approach has some problems. The main problem of this approach is that it is mainly descriptive in its nature and concentrates to a great extent on understanding the structures and dynamics of networks. It has been criticized of being too academic as a concept and lacking practical implications or management perspective (Brennan and Turnbull 2002; Ford 1998;

Ojasalo 2002). For covering this weakness, the strengths of Portfolio Models are

being combined with the concepts proposed by the IMP Network Approach; nor do

Portfolio Models, neither Relationship Networks provide complete insight for managers

in managing relationships. Portfolio Models have somehow simplistic view of

relationships and Relationship Networks do not provide applicable tools, practical

implications or management perspective for managers. In short, none of these

32

two are sufficient alone. In this research, the goal is to take advantage of the strengths and recover the weaknesses of them.

2.2.1 Network Portfolio Models

Related Network Management theories are partly overlapping the issues of the portfolio theories and their applications (Ford et al. 2002b). Several authors who have researched management of the networks have separated different levels of Network Management. For example, Möller and Halinen and Möller et al. suggest four network management levels: network visioning, network management, portfolio management and relationship management (K. Moller and Halinen 1999; K. Moller et al. 2002a, 2002b) while Ojasalo also suggests four very similar elements in managing networks: scanning and analyzing an opportunity network, identifying a manageable key network, strategies for managing actors in key network and developing and applying operational method for managing actors (Ojasalo 2002). Ritter et al. propose in their model two external and two internal dimensions of management: relationship level, portfolio and net level, internal relationships and inter-functional coordination (Ritter et al. 2002). Turnbull et al.

punctuate the importance of the separation between the short-term management of the relationships and the longer-term development of a strategy for the company’s portfolio of supplier and customer relationships. They continue that this issue is linked tightly to the network positions and finally to the competitive advantage of the firm (Turnbull et al. 1996).

On the other hand, several authors have applied the value of the political- economy framework and its capacity for identifying socio-economic interactions between partners in terms of their internal structure and external environment to conceptualize structure and process in channels of distribution (Achrol et al. 1983;

Anand and Stern 1985; Arndt 1983; Dwyer and Summers 1986; Dwyer and Sejo

1987; Frazier 1983; Gaski 1986; Heidi and John 1988; Schul et al. 1983; Stern and

33

Reve 1980; Thorelli 1986) during the time that different Customer Portfolio Models were being presented by different researches. But in spite of these overlaps between Network Management Theories and the Portfolio Theory and vast researches about political-economy framework, all the portfolio models discussed thus far are dyadic and focus on one-to-one relationships not the network and socio- economic interactions, and although all the works mentioned at the beginning of this section suggest considering both Portfolio Theory and Network Theory, they do not try to combine and integrate these two concepts together. The only two researches that try to integrate the concepts of Portfolio Theory and Network Theory are the works of Zolkiewski and Turnbull, and Ritter (Ritter 2000; Zolkiewski and Turnbull 2002). Turnbull and Zolkiewski suggest that for investigating relationships within the context of all network relationships, the interactions within and between the separate portfolios of customers, suppliers and other influential organizations should be considered. According to their work, sets of portfolios provide an alternative conceptualization of networks and cover the weakness of descriptiveness of Network Theory. For doing so, they propose a three-step process:

1. Identifying the individual portfolio constituents (customer relationship portfolios, supplier relationship portfolios, and indirect relationship portfolios, e.g. government, competitors, potential suppliers or customers, suppliers' suppliers, regulatory bodies or even other divisions of the firm itself).

2. Identifying interactions between members of the individual portfolios, by ascertaining if there are any connections between the portfolio constituents, e.g.:

• Are they different departments in the same firm?

• Are they separate business units in a larger organization?

• Are they owned by the same multinational company?

• Do they work closely with each other, e.g. would they consider themselves to be "partners"?

• Are they involved in joint ventures or strategic alliances with each other, etc?

• Do they share information (both technical and commercial)?

• Do you treat closely allied firms differently?

3. Identifying interactions between the constituents of different

portfolios.

34

Rationale behind this perspective is that because of significant pragmatic and practical difficulties in the analysis of networks, it may be more appropriate to view a network in terms of these portfolios. Viewing a network in terms of portfolios can also partially solve the problem discussed by Ritter of portfolios not depicting the interdependencies between two or more items in analysis i.e. they concentrate on categorizing products or relationships (Ritter 2000). Thus, they tend to result in strategies that are independent of each other ignoring possible interrelatedness of subjects examined (Olsen and Ellram 1997). But Zolkiewski and Turnbull's model can not eradicate this weakness completely because this model can not propose a solution to help managers to know how they can consider this interconnectedness between different portfolios. The importance of not developing strategies in isolation was stressed before in Network Theory, the problem is the way of doing so.

Ritter presents a framework developed as a tool to analyze interconnections in business networks (Ritter 2000). According to his work, in order to address the issue of interconnectedness between relationships it is sufficient to analyze triads because every greater system (the network) can be deconstructed into triads for analytic purposes and network effects can be demonstrated using only a triad (Hummel and Sodeur 1987; Kappelhoff 1987;

Smith and Laage-Hellman 1992). By this definition of interconnectedness that relationships are connected when a given relationship affects or is affected by what is going on in certain other relationships (Hakansson and Snehota 1995) and assuming that interconnectedness between two relationships can be positive or negative (Anderson et al. 1994; Blankenburg-Holm and Johanson 1992; Cook and Emerson 1978), ten possible cases of interconnectedness between any two relationships and the impact of two relationships on a third one is identified.

Although the initial concern of these two mentioned researches and the

current research project is the same (integrating Portfolio Theory and Network Theory),

the direction of this research is somehow different. The focus of this research is

35

being on the integration of variables used in the Portfolio Models with the variables depicting interconnectedness between different actors of the network. In other words, the assumption of this research is that categorizing customers into different portfolios should not be done without considering some indirect relationships affecting the relationship between our organization and the customer. Consider an example in Iranian corporate banking: as Zolkiewski and Turnbull mention in their work, the interactions between members of the indirect portfolio and a firm's existing customers/suppliers are the most problematic in management terms (Zolkiewski and Turnbull 2002). Because of clear differences between private and state customers in Iranian corporate banking market, most Iranian corporate banks, consider this issue as an indicator of customer business attractiveness and the categorization of private and state customers into different portfolios is very probable. But it is not the end of story; there is an indirect relationship that has profound impacts on the categorization of state customers:

there is a law that does not let the banks forfeit the pledges of some state organizations instead of their unliquidated loans.

In short, the basic orientation of this research is that understanding the interactions between different network actors affects the way of categorizing customers/suppliers into different portfolios and variables/indicators showing these external interactions should be incorporated into Portfolio Models.

2.3 Business Banking Relationships

There are business markets in banking sector, as happens in the industrial field.

However, some specificities are found in banking that call for further research.

Such specificities affect nature of relationships. First, supplier/customer interactions in the banking field must be considered in the context of services (J.

Proenca and Castro 1998). During the past few decades, considerable attempts

have been made to define what constitutes a service, but there is not as yet any