Keimpe Pols

Low-Value Customers in the Social Services Sector:

The Conflict between Customer Value Management and

Corporate Social Responsibility

Keimpe Pols

Low-Value Customers in the Social Services Sector:

The Conflict between Customer Value Management and

Corporate Social Responsibility

Keimpe Pols

Master Thesis

Marketing Management & Marketing Research

Feb. 2010

Keimpe Pols Supervisors

Franciscushof 57 dr. J.E. Wieringa 8801 MV Franeker dr. S. Gensler Tel: 0644236717

Management Summary

A conflict between Customer Value Management (CVM) and acting socially responsible is a topic that applies to many organizations nowadays. In this research we study this subject in an organization that is operating in the social services sector. We identify groups of low-value customers and give recommendations on how to handle these segments. We provide tools for improving profitability and make sure that the strategy to follow is in line with the Corporate Social Responsibility (CSR).

A segmentation process leads to a better understanding of customers‟ needs and wants, allowing greater responsiveness in terms of the product or service offer. We use profitability as segmentation bases. The segments are presented in a customer pyramid. The pyramid is divided in tiers, each with its own level of profitability. Next to the high-profitable platinum tier and profitable gold tier, the iron tier contains X% of the total customers contributing X% to total profit and the lead tier represents X% of total customers, contributing -X% to the total profit. We find differences between certain segments regarding the profiling bases. The tiers differ on the type of sponsor, the months of unemployment of a client, the total services used and the usage of the services empowerment, psychosocial and administrative.

We also come up with possibilities to handle low-value segments and discuss the CSR of these options. At the company we study, CSR comes from an intrinsic motivation. Firing customers is not considered as ethical. Furthermore, a strategy is socially responsible if certain conditions are met: abusing sensitive groups should be prohibited, the future prospect of a client should be good and the intensity or urgency of the problem should not be too high. Avoiding acquisition in the first place is considered as an appropriate first step to take for improving the profitability of the segments. We came up with a model that can be used for predicting profitability of a (new) client. During an intake, employees can use this tool to discover in which segment a client belongs and what strategy to follow. The variables that are used in our model are found by a multiple regression analysis with dependent variable

profitability. If avoiding acquisition is not possible, increasing price or reducing costs are other steps to take. This can be done by providing certain services to low-value segments and by working more efficiently through lower time per visit.

Table of Contents

1 Introduction and Problem Statement

§ 1.1 Introduction 6

§ 1.2 Problem Statement 7

§ 1.3 Research Questions 7

§ 1.4 Academic and Practical Relevance 8

§ 1.5 Structure of the Thesis 8

2 Theoretical Framework

§ 2.1 Corporate Social Responsibility (CSR) 9

2.1.1 Motives for CSR 9

2.1.2 Stakeholder Approach and Profitability of CSR 10

2.1.3 Communicating CSR 11

§ 2.2 Customer Value Management (CVM) 12

2.2.1 Managing Relationships 12

2.2.2 ABCs of Unprofitable Customer Management 13 2.2.3 Reduce Costs or Increase Price 15

2.2.4 Firing Customers 15

2.2.5 Ethics when Firing Customers 16

§ 2.3 Segmentation 16

2.3.1 The Segmentation Process 17

2.3.2 Segmentation Criteria 18 2.3.3 Segmentation Bases 19

2.3.4 Customer Pyramid 21

2.3.5 Ethics with Segmentation 22

§ 2.4 Theoretical Model 23

2.4.1 Factors Connected to Profitability of Segments 24

2.4.2 Factors Connected to CSR 25

§ 2.5 Drivers of Customer Value 25

2.5.1 Measuring Customer Value 25

3 Research Methodology

§ 3.1 Data Collection 31

§ 3.2 Analysis 31

3.2.1 General 32

3.2.2 Multiple Regression Analysis 32

3.2.3 Segmentation Analysis 33

4 Discussion

§ 4.1 Validity and Reliability 35

§ 4.2 Results 35

4.2.1 Results Multiple Regression Analysis 36 4.2.2 Results Segmentation Analysis 38

5 Conclusions and Recommendations

§ 5.1 Predicting Profitability 45

§ 5.2 Profiling the Segments 45

§ 5.3 CSR at Company X. 46

§ 5.4 Strategy for Approaching the Low-Value Segments 46

§ 5.5 CSR Communication at Company X 51

§ 5.6 Limitations of this Study and Directions for Further Research 51

§ 5.7 Academic Relevance 52

References

Appendices

Appendix A: Should companies communicate corporate citizenship? 60

Appendix B: Four types of customers (Reinartz & Kumar, 2002) 60

Appendix C: Initial Model with User Status Variables 61

Appendix D: Adjusted R-square (final model) 62

Appendix E: Adjusted R-square (initial model) 62

Appendix F: Graphs to check for Normality 63

Appendix G: Test of Homogeneity of Variances 64

Appendix H: F-test 64

1 Introduction and Problem Statement

§ 1.1 Introduction

“Yesterday my colleague told me that he gave his client 50 euros so he could buy a train ticket. This client has thousands of debt and almost no money to spend because most of his budget directly goes to his creditors. His family lives at the other side of the country. He hardly ever visits them because he does not have the resources to get there. By providing him the 50 euros for a train ticket, he can visit his family for the first time in three years in order to celebrate the end of the Ramadan (which is the most important social and religious event for Muslims). My colleague knows that this 50 euros goes in expense of the profit of the

organization. However, he felt that his organization had the social responsibility to help this client without getting a direct return from it.”

Above example of a firm acting socially responsible occurred at the company named “Company X”. This company will be studied in this research. Company X is a profit

organization for care, employability and reintegration. Clients of the organization are persons who have psycho-social problems, are in debt, have problems at work or have trouble finding a job. Sometimes these clients pay for the services by themselves, but most of the time they have a sponsor. This in the way of a PGB (a personal budget, provided through government subsidy), local authorities that pay after an offer is made and sometimes by companies who pay for their employees. These customers have a negative motivation when making use of the services of Company X. They often have no choice but make use of the services the company provides. This is different to regular customers, who are driven by their “positive” needs (Arnold & Reynolds, 2003).

Company X provides an integrative solution to these customers. Meaning that the

organisation looks at the whole picture when someone starts a track. For example when a client is looking for a job (reintegration), they also check if the client has psychological issues, addiction problems or is in debt. Company X is capable to deal with the whole problem and provides a solution that fits the personal needs of the client.

mentioned in the beginning, other signs of social responsibility at Company X can be found in the mission statement of the firm, the organisational culture, the fact that they sometimes help people without getting paid or when financial returns are uncertain and in the fact that they do not turn down customers when they are not (that) profitable. However, the company has to make a profit in the end. This might give some friction now and then.

This conflict between making a profit and acting socially responsible is a problem. Especially when an organization is looking at Customer Value Management (CVM). The management of the company we study wants to improve CVM because they presume that there are some customers who are not contributing (much) to the profitability of the organisation. In other words, these customers are low-value. These low-value customers give a loss or generate a far lower profit than other customers. Company X does not have a clear view about the amount of resources spent at a customer compared to the budgeted resources. They want to know if there are certain low-value customers and once they are identified, they want to know how to manage these customers. Is there a solution that does not cause the organization financial problems and is social responsible as well?

§ 1.2 Problem Statement

Above mentioned problem gives the following problem statement:

§ 1.3 Research Questions

In order to answer the problem statement above, the following research questions are formulated:

1. What is CSR?

2. Why does an organization needs to take care of CSR? 3. What is customer value management?

4. How to identify low-value customers? 5. What to do about low-value customers? 6. How to balance customer value and CSR?

§ 1.4 Academic and Practical Relevance

Corporate social responsibility is a hot topic nowadays. Many people blame the large

companies and managers for not acting socially responsible and therefore accountable for the credit crisis. It seems more important than ever before for organizations to be perceived as respectable and socially responsible, i.e. to build a corporate reputation of social commitment (Hooghiemstra 2000; Fombrun & Van Riel 2003; Maignan & Ferrell 2004; Morsing & Beckmann 2006). Also in literature CSR gets frequent attention (Du et al., 2007; Sen & Bhattacharya, 2001; Klein & Dawar, 2004; Waller & Lanis, 2009). However, research on the combination of customer value and CSR is lacking. So when companies want to act socially responsible and want to maximize their profits through Customer Value Management (CVM), literature does not provide any guidelines on how to handle this. Regular CVM-theories do not apply for a socially responsible firm. Therefore it is time to start discussing this issue and we hope that this paper will contribute to that. Also the type of company that we consider in this study is not often discussed in previous research, due to recent privatization in the industry. So this research in this specific sector is both interesting for practitioners as well as for researchers. Especially for managers in this sector, but also for other managers that, for instance, want to look at CVM. Also practitioners and researchers that want to know more about the conflict between CSR and customer value can use this paper. So we may conclude that this research is of both theoretical as well as practical relevance.

§ 1.5 Structure of the Thesis

2 Theoretical Framework

In this chapter we discuss relevant literature regarding CSR, CVM and segmentation. We also come up with a conceptual model, which will give a schematic overview of the issues

discussed in this paper.

§ 2.1 Corporate Social Responsibility (CSR)

The term corporate social responsibility is mentioned a few times already in this thesis. But what is CSR exactly? Literature does not consistently gives one clear definition. According to Smith (2003) it refers to the obligations of the firm towards society. Podnar (2008) also underlines that it mostly is seen as a slightly fuzzy concept with numerous definitions. But all have in common that they acknowledge the fact that companies have responsibilities towards the society and environment that go beyond their own interests and legal obligations (De Bakker et al., 2005). According to Morsing et al. (2008), the attributes of social responsibility include responsibility towards society (community responsibility), concern for the environ-ment (environenviron-mental responsibility) and philanthropic behaviour or corporate social initiatives (support good causes) and that these attributes entail a broad variety of CSR activities. In this section we are going to discuss motives for CSR, we look at the stakeholder approach and profitability of CSR, and we discuss communication of CSR.

2.1.1 Motives for CSR

intrinsic and extrinsic motives. When consumers believe that CSR activities are derived from an intrinsic motivation, they have a more positive attitude towards the CSR and find it more believable. Perceived extrinsic motives will diminish the favourability of the CSR beliefs. Such extrinsic attributions are likely to be associated with the believe that the firm is not truly socially responsible (Du et al., 1997). Intrinsic CSR motives can be measured as the brand‟s genuine concern in being socially responsible and extrinsic CSR motives in terms of the competitive pressures the brand faces to engage in CSR activities (Du et al., 1997; Ellen et al., 2006).

2.1.2 Stakeholder Approach and Profitability of CSR

The extrinsic motive can be explained through the stakeholder approach. The theory behind this approach is that companies are not accountable only to their shareholders, but should also balance the interests of stakeholders that can affect or are affected by their operations

(Freeman 1984). Stakeholders are for example employees, customers, financers, communities or local governments. Clarkson (1995) makes a further distinction between primary and secondary stakeholders. A primary or participant stakeholder is one without whose continuing participation the corporation cannot survive as a going concern (Metcalfe 1998). This can be for example investors, employees, major customers or suppliers. Secondary or non-participant stakeholders are defined as those who influence or affect, or are influenced or affected by the corporation, but are not engaged in transactions with the corporation and are not essential for its survival (Metcalfe 1998). These can for example be minor customers, minor suppliers or neighbours.

That this stakeholder approach makes sense can be found in the argument that CSR leads to profitability. According to a meta-analysis of (Orlitzky et al., 2004) certain positive effects of CSR, especially the company‟s increased reputation as mediator, have a positive influence on the profitability of the firm. This increase in profits can come from a better reputation the company gets when acting socially responsible (Fombrun & Shanley, 1990; Miles & Covin, 2000). A better reputation can lead to more sales (Brown & Dacin 1997; Alexander 2002) and enhances product evaluations, while a bad reputation has a negative effect on product

et al., 1990). CSR does not only has a positive effect on current employees, but also helps to attract new talent (Grow et al., 2005).

2.1.3 Communicating CSR

Social and environmental activities need to be visible to stakeholders before they can, in various ways, reward the organisation for them (Orlitzky, 2005). So in order to get to the point that CSR pays off, communication of CSR view can play an important role.

In recent years annual reports has started to contain information on how the organisation is handling its social responsibilities (Sweeney & Coughlan, 2008) and the annual report represents the main communication method used by firms to disclose CSR information (O‟Dwyer 2003). This report is increasingly available online for other stakeholders to view (De Bussy et al., 2003). Companies are increasingly issuing special CSR reports to provide customers and other stakeholders with CSR information (Podnar, 2008).

However, we have to keep in mind that small firms are less inclined to use formal instruments to promote CSR of the organisation than large firms. Large firms are more visible to the public and the media. Investment in instruments useful in external communication such as a code of conduct, ISO certification and social reporting is therefore more important for large firms (Graafland & Van den Ven, 2006).

The study of Morsing et al. (2008) shows that recipients prefer that minimal CSR communication is released through annual reports or websites above the use of intensive communication channels (e.g. corporate advertising, public releases) and above no publication at all (see appendix A).

But how should this communication look like? According to Orlitzky (2005) publicity efforts need to be relentless, but believable. If the public regards the organisation‟s

On the one hand people are in favour of CSR efforts of companies, but on the other hand they dislike if companies actively communicate about it.

For communicating CSR activities, an inside-out approach should be used (Morsing et al., 2008). It is best to start with involving and committing employees on all issues of corporate CSR policies; beginning with CSR efforts concerning employees themselves. CSR activities may be promoted and supported by top management, but if employees do not develop a sense of ownership, the activities will not find organizational support and not create a basis for the continuation of the activities themselves nor for a trustworthy communication about them (Morsing et al., 2008). Once an acceptable level of internal commitment is reached concerning internal CSR issues, employees and managers can communicate directly to external stakeholders in a believable way. The level of commitment to social responsibility is dictated by several factors, including the nature of the product, target market, corporate culture and mission statement (Machiette & Abhijit, 1994).

§ 2.2 Customer Value Management (CVM)

Customer value management represents the extent to which the firm can define and

dynamically measure individual customer value and use it as its guiding metric for marketing resource allocation decisions (Ramani & Kumar, 2008). Historically, customer data has primarily been used to identify the high value customers and to define ways for serving them in an optimal manner. Nowadays, companies have become more aware of low-value

customers, and the fact that these relationships can account for a substantial share of their total customer base (Haenlein & Kaplan, 2009). In recent years, several studies have shown that the share of customers with a negative contribution margin (revenue less direct cost and cost-to-serve) can reach up to 30% (Haenlein & Kaplan, 2009). In this section we provide options that are mentioned in literature about how CVM can be performed. We specifically look at the management of low-value customers.

2.2.1 Managing Relationships

relationship between loyalty and profitability in a way that makes it possible to better identify which customers to focus on. Relationships needs to be based on customers‟ profitability, not the revenue they generate. The cost of servicing customers who buy only small quantities of low margin products, may exceed the revenue they bring in. So next to the expected length of the relationship, an estimate of the average profit earned on each customer in any typical purchase period should be made.

After analyzing customers‟ profitability and projected duration of the relationship, a distinction between certain types of customers can be made. Reinartz & Kumar (2002) identify four types of customers who are different in their profitability and loyalty profiles (see appendix B):

Strangers: They are not loyal and are not profitable; identify them early and do not invest anything in this customers.

Butterflies: It is wonderful when they are around, yet unfortunately they leave easily; so they are profitable but disloyal. Milk them for the short time they are your

customer.

True Friends: They are loyal and profitable; use softly-softly approach (be cautious and patient and avoid direct action or force).

Barnacles: They are strongly attached to the firm but may cost the firm more in the long run; so they are loyal but unprofitable. Find out if they have the potential to spend more and become profitable.

2.2.2 ABCs of Unprofitable Customer Management

Haenlein & Kaplan (2009) describe a six step approach, which they call the ABCs of

Unprofitable Customer Management, for dealing with unprofitable customers. If a certain step does not provide a satisfied solution, proceed to the next step. These are the six steps as identified by Haenlein & Kaplan (2009):

1. Avoid their acquisition in the first place

2. Bear in mind potential rescue operations

Some types of unprofitable accounts can potentially be transformed into profitable ones, while others might simply appear to be unprofitable because the company has an incorrect

understanding of their needs. If the firm misunderstands the needs, it might offer a product that includes features the customer does not value and is not willing to pay for. Given that including these features creates cost, this can lead to unprofitable clients. If this is the case, firms may want to consider changing their business model. Most of the times it is the company that operates its business in an inefficient manner, resulting in unprofitable customers.

3. Catch the possibility of abandonment

Sometimes it is an option to say goodbye to your customers (see also the section ”Firing Customers” below). Three possible ways of handling this are mentioned:

“Cost escalation strategy”, involves increasing relational costs in the hope that unprofitable customers decide to leave the company.

“Withdrawal”, entails reducing the intimacy of the customer relationship.

“State-of-the-relationship talk”, consists of explaining to the unprofitable customer the reasons for its low profitability, and trying to convince him to change its behavior.

4. Draw up a cost-benefit analysis

Whether the strategy of abandonment is beneficial, depends on the benefits and costs associated with it. Also the cost of negative word-of-mouth should be taken into consideration.

5. Ensure familiarity with your environment

6. Facilitate biting the bullet

When previous five steps are reviewed and a customer abandonment strategy is chosen, the paper (for more details, see Haenlein & Kaplan, 2009) provides six recommendations for doing this the least painful way: be explicative, efficient, direct, open, personal and clever.

2.2.3 Reduce Costs or Increase Price

As mentioned before, most of the times it is the organization that runs it operations inefficiently (Haenlein & Kaplan, 2009). Therefore Zeithaml et al., (2001) advice that the company can make unprofitable customers profitable by cutting costs and serve the segment more efficient. An other option they mention is to increase price. An effective approach is to increase price for services customers have been receiving, but are not paying for. The paper of Mittal et al (2008) is also in line with these solutions. They recommend to implement

differential pricing and service strategies. Though it should be kept in mind that customers can become outraged when they find out about dynamic pricing and when they have the feeling that they pay more than other customers (Mittal et al., 2008).

2.2.4 Firing Customers

Another frequently mentioned option for dealing with unprofitable customers, is “firing customers” by getting rid of unprofitable customers (Haenlein & Kaplan, 2009; Mittal et al., 2008; Selden & Colvin, 2003; Zeithaml et al., 2001). This strategy is also called customer divestment or (selective) demarketing (Kotler & Levy, 1971, Selden & Colvin, 2003; Mittal et al., 2008). This method, when a company stops providing goods and services to existing customers, was once considered as not done. But nowadays it is seen as a viable strategic option to many organizations (Mittal et al., 2008). According to Mittal et al. (2008) these firms are taking advantage of new segmentation approaches and technologies that have made it easier to focus on retaining the right customers; those who will bring in the most revenue over time; and, by extension, to show problem customers the door. Mittal‟s research identifies four reasons why businesses terminate relationships with end users: the declining profitability of specific customers, the lower productivity of employees as they deal with unprofitable customers, changes in the capacity to serve large volumes of customers and shifts in a

with the greatest value growth potential and create strategies to migrate them from a lower value to a higher value group.

2.2.5 Ethics when Firing Customers

Regarding the topic of customer divestment, some literature questions the ethical and legal component of this option. The downside of firing customers is that these customers are likely to feel dissatisfied, angry, or betrayed, and speak negatively about the company (Mittal et al., 2008; Bougie et al., 2003; Richins, 1983). Mittal et al. (2008) mention that negative effects can be retaliation of customers, getting a bad reputation, sending business to your competitors and disturbing the relations with high-value customers. Next to these downsides, it may even violate ethical or legal obligations towards your customers. Furthermore, such a strategy may directly contradict the principles of corporate social responsibility deeply embedded in the organization. Mittal et al. (2008) also states that it should be considered that citizens of Western nations generally expect certain services (electricity, water, sanitation, and heating) to be universal, regardless of one‟s ability to pay. It should be noticed that because

differentiation and segmentation are the cornerstones of most customer divestment programs, certain initiatives (revenue management) may be perceived as discriminatory. Mittal‟s paper concludes that in certain situations customer divestment can be an effective strategy, although it should clearly be an option of last resort.

§ 2.3 Segmentation

particular customer groups with relatively homogeneous requirements (Choffray & Lilien, 1980). This leads to a more efficient application of resources and makes it possible to target your offerings specifically (Wind, 1978; Beane & Ennis, 1987; Blattberg & Sen, 1976). By applying a segmentation approach, firms are encouraged to perform a detailed customer analysis. This leads to a better understanding of customers‟ needs and wants, allowing greater responsiveness in terms of the product or service offer (Dibb & Simkin, 2001). In this section we describe the segmentation process, discuss segmentation criteria for effective

segmentation and look at segmentation bases that can be used. We also discuss the customer pyramid, which can be used when using profitability as segmentation bases. We conclude by discussing ethical implications with segmentation.

2.3.1 The Segmentation Process

The process of segmentation is in literature often presented as a chain of different phases. You start with the Segmentation phase, then you proceed to the Targeting phase and you conclude with the Positioning phase. This STP process (figure 1), developed by Kotler (1997) is presented and discussed below:

Figure 1: The Market Segmentation Process (Kotler, 1997)

Market Segmentation

The first step of the market segmentation process is to identify and profile distinct groups of buyers who might require separate products or marketing mixes. These groups have similar needs and buying behaviour. First, the bases for segmenting the market should be identified. An important step is the selection of variables used for the analysis. The results of the analysis are very sensitive to the selection of the variables. After the market has been segmented, the segments have to be described in order to get profiles of the segments.

Market Targeting

1995; Piercy, 1992). In this stage, decisions are made about where resources should be prioritized (Dibb & Simkin, 2001). To achieve this, marketers are required to make choices about the relative attractiveness to the business of a number of target markets (Simkin & Dibb, 1998). The study of Simkin & Dibb (1998) shows that profitability is regarded as the most important criteria for selecting a target market, followed by market growth and market size. According to Kotler (2000), the firm first must ask whether a potential segment has the characteristics that make it generally attractive. Second, it must be considered whether investing in the segment makes sense given the firm‟s objectives and resources. Some attractive segments could be dismissed because they do not match with the company‟s long-term objectives or if the company lacks one or more necessary competences to offer superior value.

Market Positioning

The final step is to develop a positioning and marketing mix. Positioning is the act of designing the company‟s offering and image to occupy a distinctive place in the target market‟s mind (Kotler, 2000). The products‟ key distinctive benefits in the market should be established and communicated. This stage concerns the design of marketing programmes which will reflect the proposition on offer and which will shape customers‟ perceptions about the nature of that offer (Dibb et al. 1997).

2.3.2 Segmentation Criteria

Marketers are urged to consider the segmentation variables before making final decisions about which segmentation scheme to adopt. Literature mentions different requirements for effective segmentation. The most used criteria for effective segmentation is Kotler‟s (1991) segmentation criteria. To be useful, market segments must be:

Measurable: Size, purchasing power and profiles of segments can be measured. Substantial: Segments must be large or profitable enough to serve. A segment should

be the largest possible homogenous group worth going after with a tailored marketing program

Accessible: Segments can be effectively reached and served.

Actionable: Effective programs can be formulated for attracting and serving the segments.

These segmentation criteria are used for evaluating the segments and to see if the segments are useful for the application of marketing activities.

2.3.3 Segmentation Bases

The selection of appropriate segmentation bases and methods is crucial with respect to the number and type of segments that are identified in segmentation research, as well as to their usefulness to the firm. The choice of different bases may lead to different segments being revealed. A segmentation basis is defined as a set of variables or characteristics used to assign potential customers to homogeneous groups (Wedel & Kamakura, 1998). Kotler (2000) makes a distinction between consumer characteristics and consumer responses as

segmentation bases. Frank et al. (1972) classify segmentation bases into general (independent of products, services or circumstances) and product-specific (related to both the customer and the product, service and/or particular circumstances) bases. Furthermore, they classify

segmentation bases into whether they are observable or unobservable. Those distinctions lead to the classification of segmentation bases as shown in figure 2 below:

Figure 2: Segmentation Bases (Wedel & Kamakura, 1998)

I Observable General Bases

These segmentation bases are relatively easy to collect, reliable and generally stable.

Segments derived from the bases are easy to communicate and resulting strategies are easy to implement. The segments are often readily accessible. Some differences in purchase

behaviour and elasticity‟s of marketing variables have been found among these types of segments, supporting the differentiable criterion, but in many studies the differences were in general too small to be relevant for practical purposes (e.g. Frank et al. 1972; Frank, 1968; Frank 1972; McCann, 1974). Although the lack of significant findings in those studies supports the conclusion that the general observable bases are not particularly effective, they are often applied in segmentation studies.

II Observable Product-Specific Bases

The bases in this class comprise variables related to buying and consumption behaviour. Purchase and usage data on consumers can be used as a data source for discovering differences between purchasing power of segments. Meaningful segmentations depend on finding patterns in customers‟ actual buying behaviour (Yankelovich & Meer, 2006). Accessibility of the segments identified from these bases appears to be somewhat limited in view of the weak associations with general consumer descriptors (Frank et al. 1972; Frank 1972). Segments are also less differentiable using these bases. Weak to moderate differences in elasticity‟s of marketing mix variables were found among segments identified by product-specific bases, indicating that those bases are measurable and substantial.

III Unobservable General Bases

The predictive validity of lifestyle with respect to purchase behaviour can be substantially better than that of general observable segmentation bases (Frank et al. 1972). Personality, values and lifestyle provide a richer perspective of the market based on a more lifelike portrait of the consumer and therefore provide actionable bases.

IV Unobservable Product-Specific Bases

Certain segmentation bases may be preferred depending on the specific requirements of the study at hand. It depends on the segmentation goal in order to determine which segmentation bases to use (Yankelovich & Meer, 2006; Wedel & Kamakura, 1998). In general, the most effective bases are product-specific unobservable bases. A variety of bases may be combined, each according to its own strengths.

2.3.4 Customer Pyramid

When the goal of the segmentation is to distinguish between valuable and non-valuable customers, customer profitability is often used as segmentation bases (Verhoef & Donkers, 2001). In the article of Zeithaml et al. (2001) they state that customer profitability can be managed and increased by segmenting customers into profitability tiers using a customer pyramid (see figure 3).

Figure 3: Customer Pyramid

The Platinum Tier: These are the company‟s most profitable customers. They are heavy users of the product or service, not overly price sensitive, they want to try new offerings and are committed.

The Gold Tier: These customers have a lower profitability than the platinum tier. They might not be as loyal to the firm, even though they are heavy users in the product category; they might minimize risk by working with multiple vendors

The Iron Tier: This tier contains customers that provide the volume needed to utilize the firm's capacity but whose spending levels, loyalty and profitability are not substantial enough for special treatment.

The Lead Tier: This tier consists of customers that are costing the company money. They demand more attention than they are due given their spending and profitability and they are sometimes problem customers; complaining about the firm to others and tying up the firm's resources.

2.3.5 Ethics with Segmentation

Market targeting sometimes generates public controversy (Machiette & Abhijit, 1994). Especially when it concerns targeting of sensitive groups. A sensitive group is a segment of the population generally perceived as being disadvantaged, vulnerable, discriminated against, or involved in social issues which consequently influence their consumer behaviour

(Machiette & Abhijit, 1994). The following typology identifies the primary sources from which groups derive their sensitivity:

Culturally dictated sensitive groups: These groups elicit a high impact media profile and, despite fluctuations in the intensity of public awareness, have captured the nation‟s attention and invite public scrutiny for an enduring period of time. They are usually associated with specific social issues.

Situationally influenced sensitive groups: These groups result from temporary environmental or personal circumstances which place the consumer in a sensitive group for a period of time. Divorce, temporary unemployment, family death or other circumstances can create a situation of consumer vulnerability.

Issue-driven sensitive groups: In contrary to previous categories, these consumers need only to be sympathetic to cause-related or social issues, rather than being an actual member of a vulnerable group. These are defined as groups of consumers responding to social issues through observable changes in their purchasing behavior. Examples are members of cause-related marketing groups, like rain forest advocates.

Above classification can be used to identify certain sensitive groups that are affected by the companies marketing activities. In examining these groups, not only the perceptions of sensitive group members should be considered, but also attitudes of the public at large concerning related social issues. This often gives the real pressure on marketers for social responsibility (Machiette & Abhijit, 1994).

Machiette & Abhijit (1994) state that concern with social responsibility will continue to create problems and opportunities for marketers targeting sensitive groups. Anticipating and

incorporating the issues into a systematic social responsibility program is essential for competing in the environment. Kotler (2000) also underlines that the public is concerned when marketers take unfair advantage of vulnerable groups or disadvantaged groups (Kotler, 2000). Kotler states that the issue is not who is targeted but rather how and for what. Socially responsible marketing calls for targeting that serves not only the company‟s interests, but also the interests of those targeted.

§ 2.4 Theoretical Model

This section gives an overview of the theoretical discussion so far. We provide a graphical overview of the content of this paper. Next to the information extracted from literature, we use information gathered from employees of the company studied in this paper.

Figure 4: Conceptual model of the decision on low-value segments with the conflict of profitability and CSR

2.4.1 Factors Connected to Profitability of Segments

The factors that are related to the profitability of segments are derived from our previous discussion of the literature. We distinguish variables that are drivers of profitable segments and variables that are used for the identification of segments.

One of the profit drivers of segments is the extent to how efficient an organisation operates. By not being able to deliver the right services on the right way to the right customers, inefficiencies can occur resulting in higher costs. Another factor that has a relationship with the profitability of segments, is the marketing actions of the company. This can contain the offering of different products to segments or the implication of other marketing mix

instruments. A third factor that has a relationship with segment profitability is dynamic

pricing. When different prices to separate segment are charged, profitability can be increased.

factor is the segmentation criteria. This criteria tests the usefulness of the segments that are identified.

2.4.2 Factors Connected to CSR

The factors that are connected to CSR are derived from previous discussed literature, but also from interviews with employees of the company. From previous studies we found that there are two distinct motives for CSR. An intrinsic motive is embedded in the organisational culture and is more an ethical motive. An extrinsic motive results from the belief that CSR is good for the company‟s reputation and the employee satisfaction. Other factors that can make a company act socially responsible, are the future prospect of a client; the urgency and

intensity of the problem a client has; and if the client is a member of a sensitive group. When a firm perceives that the future prospect of the client is desperate without the help of the organization, they may act socially responsible. The same thing occurs when the problem of a client is urgent or the intensity is high. Than the company feels obliged to offer services to this person. When a client is a member of a sensitive group, CSR has to be taken into account. It should be prevented to take advantage of these sensitive groups. Finally CSR communica-tion is an important factor regarding CSR. For optimal CSR communicacommunica-tion, an inside-out approach is recommended. Also make sure that CSR communication is not to intensive.

The model discussed gives an overview of the problem we are dealing with: the conflict between the identification and decision on low-value segments on the one side and CSR on the other side. The rest of this report will discuss this issue and at the end we will come to a number of recommendations on how to decide regarding the low-value segments.

§ 2.5 Drivers of Customer Value

In this section we explain how customer value is measured and discuss what variables drive customer value. In this section we also formulate hypotheses.

2.5.1 Measuring Customer Value

costs consist of the labour hours that can be accounted to a client times the hourly cost-price (direct costs and an add-on for indirect costs) of Company X. In the classification of

segmentation bases as suggested by Wedel & Kamakura (1998), profitability would best fit in the “Observable Product-Specific Bases”. Profitability can be observed (is objective) and it relates to buying behaviour or user status.

2.5.2 Hypotheses

We discovered three distinct drivers of customer value. These drivers are also used for profiling the segments. They are general observable variables in the way of demographics of the clients; the observable product-specific basis in the way of the user status of a client; and the observable product-specific basis in the way of the service usage. We now discuss each and come up with hypotheses.

Demographics

The general observable variables are the demographics of the clients. Problems clients have and the course of the track depend on the demographic characteristics of a client.

Demographics make each problem unique.

Common used demographics are age and gender (McDonald & Dunbar, 1995). We include these variables in our research. Because of the lack of research in this specific sector and due to inconclusive evidence from the interviews, we cannot predict the direction of effect. Therefore we expect that:

H1: There is a relationship between age and profitability; and H2: There is a relationship between gender and profitability.

We include family size, because we expect that a large family size will have a negative influence on the profitability. When children are involved, more complex problems may occur. It also happens that not just one client is taken care off, but the whole family makes use of the services of the company. This has a negative influence on the profitability, because for only one person is getting paid. Therefore we expect that:

H3: There is a negative relationship between family size and profitability.

We also include marital status, because for example ex-partners may cause extra

sector and due to inconclusive evidence from the interviews, we cannot predict the direction of effect. Therefore we expect that:

H4: There is a relationship between marital status and profitability.

There can be even more implications if a client has children with their ex-partner. We expect that this will have a negative influence due to extra complications. Therefore we expect that:

H5: There is a negative relationship between having children with an ex and profitability.

Ethnicity is also regarded as an important variable. Communication between a client and other parties may be troublesome due to cultural and language problems. Therefore we expect that:

H6: There is a positive relationship between being a “native” and profitability and a negative

relationship between being a “foreigner” and profitability.

Education is also seen as a factor that has a big influence on the profitability of a client. Clients with a higher education can be helped with less effort. Also because they self can provide a positive contribution to the solution of the problem. Lower educated people require more attention. Therefore we expect that:

H7: There is a negative relationship between educational level and profitability

Clients with physical disabilities or the intensity of the mental disability of a client are also factors that can influence profitability. Clients with physical handicaps or severe mental problems may get a higher budget from sponsors more easily. Therefore we expect that:

H8: There is a positive relationship between having a physical disability and profitability; and H9: There is a positive relationship between intensity of mental disabilities and profitability.

Difficult or extraordinary cases have a different influence on profitability than “standard” issues. An alcohol or drugs-addiction or a criminal record can therefore also give implications during a track. Therefore we expect that:

H10: There is a negative relationship between having an alcohol or drugs-addiction and

profitability; and

Finally, the amount of time someone is unemployed may also play a role. If someone is employed, debts can be paid off more easily. When someone is out of a job for a long time, it often is more difficult to provide a successful reintegration track. Therefore we expect that:

H12: There is a negative relationship between months of unemployment and profitability.

User status

Next to demographics, the user status of a client can be drivers of customer value. User status is an observable product-specific basis. User status tells us something about the way a client makes use of the services of Company X.

It often occurs that a client is getting help from other people or organisations. For example help from social workers, youth-care, psychologists, a general practitioner, psychological treatment by the GGZ (mental healthcare institution), help regarding their finances by the GKB (institution for people that are in debt) or help from other organizations. It is not sure if help from these organisations lead to a higher profitability for Company X. It often does save time because a diagnose made by the GGZ can be used for acquiring a budget for a client. This results in a higher efficiency. However, it is possible that clients are not helped in a good way by other organisations. This leads to extra problems and inefficiency due to additional time-consuming actions. Due to these contradictions we can not predict the direction of the relationship. Therefore we expect that:

H13: There is a relationship between getting help from other organizations and profitability.

The sponsor, so the organization or person that pays the bill for the client, is also regarded as an important factor affecting profitability. Employees in the organisation mentioned that some sponsors result in a higher profitability than others. They expect that clients from the UWV (institution for unemployment) lead to the highest profits. The structure of a track and the nature of the problems from these sponsor or rather stable and therefore more easy to handle. For example the problems from clients that are redirected by the local authorities are more complex. However, there is a risk with clients from UWV. In order to receive payment for the services Company X delivers, 70% of the clients needs to be successfully treated (e.g. a job is found). Because of this uncertainty we can not predict the direction of effect. Therefore we expect that:

H14: UWV clients have the highest relative impact on profitability, followed by PGB2, third

The average duration of a contract is also a factor that may influence profitability. This relationship can go either way. One possibility is that in a shorter track, the problem is more clear and therefore can be handled efficiently. Another option is that in a shorter track, time and resources are lacking in order to provide help in a good way. Due to this contradiction we can not predict the direction of effect. Therefore we expect that:

H15: There is a relationship between the duration of a contract and profitability.

Frequency of visits can be another factor that influences profitability. The (planned)

frequency of visits is the amount of visits a month that is getting paid for. We expect that this has a positive influence on profitability, because when this frequency is higher, employees do not need to provide much extra help outside these visits. This same reasoning goes up for time per visit. We expect that when the budgeted or planned time for a client is higher, that this will have a positive influence on profitability. When an employee has long planned

contact hours, he can help this client better and he does not need to spent additional uncovered time. Therefore we expect that:

H16: There is a positive relationship between the frequency of visits and profitability. H17: There is a positive relationship between the time per visit and profitability.

Service Usage

Finally, service usage is a driver for customer value and used for profiling. This is also an observable product-specific basis. Service usage is defined as the actions and help that Company X provides for the benefit of the client. This includes twelve different services (see figure 5). The more services a client receives, the better (meaning: more efficient) the client can be helped. Most of the time the problems clients face are not resolved by using a single service. Often there are more problems to be taken care of for helping a client in a good way. Therefore a whole range of services are offered to clients. Because Company X beliefs in this integrative solution, we expect that all the services offered have a positive relationship with profitability. Therefore we expect that:

H18: There is a positive relationship between service usage and profitability.

Figure 5: Drivers of Customer Value Demographics

Age Gender Family size Marital status

Number of children with ex Ethnicity Education Physical disabilities Intensity of mental disabilities Alcohol/Drugs addiction Criminal record Months of unemployment User Status Help other organizations Sponsor PGB1

Sponsor PGB2 or higher Sponsor Local Authorities Sponsor UWV

Duration Frequency Time per visit

Service Usage Budgeting Training/coaching Empowerment Psychosocial Administrative

3 Research Methodology

In this chapter we discuss the research methodology of this paper. We first look at the data collection. After that, we explain how we analyzed the data.

§ 3.1 Data Collection

In this section we discuss how the data that is used in our analysis is collected. The

information we need for determining the profitability of a client, is found in internal reports. We look at the latest annual report (2008), hour-registration forms that employees fill in, ledgers and bills.

In order to describe the segments in a useful way, we need to know which variables are important to consider in the profiling stage. We are looking for the variables that have a relationship with profitability. We find out which variables we need by having interviews with all employees that have to deal directly with clients. With these six interviews, we get an understanding of the care process and about the clients of the company. Next to that, we randomly read reports about clients that are available in the database of the company. We read reports about 35 different clients, which is approximately 15% of the company‟s database. In the reports we search for information about the nature of the care clients get, the difficulties that employees and clients run into during the process and the availability of customer specific information (e.g. demographics). Based on the information generated from the interviews and reports, we select the variables used for profiling.

The data used for segmentation bases and the value-drivers are found in the same internal reports from the database of the company. There are 223 client-files available in that database. The database only consists of clients that have completed their track. 107 of these 223 files contain sufficient information that can be used for analysis, which is 48%. So 116 cases contain no information about clients or hour-registration and therefore can not be used. Therefore the database we are going to work with contains data about 107 clients.

§ 3.2 Analysis

3.2.1 General

We start our analysis by checking if there are any outliers in the database. We do this by looking at descriptive statistics. Next, we make segments by dividing the clients in segments based on their profitability. Because our database consists of just over a hundred cases, four segments are chosen. If we would divide the database into more groups, the number of cases in these groups would be rather low. The segments are presented in a customer pyramid.

3.2.2 Multiple Regression Analysis

We want to know what to do with underperforming segments, so it would be useful to develop a model that predicts the profitability of a client on beforehand. Because we have data on individual client level, we can use multiple regression analysis to discover which variables have a relationship with profitability.

The variables mentioned in the hypotheses (see also figure 5) are used in the multiple regression analysis in SPSS as independent variables. We use profitability as the dependent variable. As estimation technique for the multiple regression analysis, we use the sequential search method. We apply the procedure of “forward addition and backward elimination” (Hair et al., 2006). This is a trial-and-error process for finding the best regression estimates. We first enter user status variables because these have the highest correlations with profitability. This model is shown in Appendix C. Next, we repeatedly re-compute the regression equation by adding en deleting variables at each stage. We delete independent variables that are highly insignificant, but try adding them later on when other variables are entered. In this way we prevent that we exclude variables that are not significant by itself, but are significant given the presence of another variable. We also look at the contribution of each variable to the adjusted R-Square at each stage. This results in a model in which independent variables contribute significantly. The hypothesized relations and directions are tested to see if there is enough evidence on a 90% significance level to support the previously mentioned hypotheses.

We check if there are multicollinearity problems by looking at the VIF values. To see if there is a normal distribution, we look at graphs of the standardized residuals.

3.2.3 Segmentation Analysis

We segment our database by using profitability as a segmentation bases. The customer pyramid consists of different tiers (segments). The first segment, the platinum tier, represents the most profitable clients. These are the top 7 clients (7%) in our database. There is a big jump in the data between the 7 highest and 8 highest client in profit (a 60% increase) and therefore we separate the customers at this point. Therefore this platinum tier contains clients with ≥€4590,- profit.

The second segment, the gold tier, represents client who also have a high profit; though lower than the platinum tier. These 33 clients (32%) have above average profit. Therefore the gold tier contains clients with a profit ≥€2061,- and <€4590,-.

The third segment, the iron tier, represents the clients that are low-value; yet they still make a profit. These 36 clients (35%), have below average profit. Therefore the iron tier contains clients with a profit ≥€0,00 and <€2061,-.

The fourth segment, the lead tier, represent the clients that generate a loss. These 27 clients (26%) have a negative profitability. Therefore the lead tier contains clients with a profit <€0,00.

The customer pyramid also shows the total profit of a segment, the average profit of a client in that segment, the percentage of the total customers that are in that segment and the

percentage of the total profit that is gained in that segment.

We make profiles of the segments by applying analysis of variance (ANOVA). ANOVA compares if the means of the profiling bases of a number of groups are different. There is a difference when there is enough evidence at a 95% confidence interval. Profitability is the segmentation basis and the drivers of customer value are used as profiling bases.

One of the assumptions in ANOVA is that the group variances are equal. With the Levene test we check whether the data meets this assumption. If there are equal variances we use a normal F-test. If this assumption is not met, we use the Brown-Forsythe and Welch test.

A chart of group means is presented in order to get a quick overview of whether there are large differences among the groups and which groups score high or low in a certain variable.

4 Discussion

In this section we discuss the results of our analysis. We first look if our database contains outliers. The Levene test is performed before we can do ANOVA. We go into the validity and reliability of the regression analysis by looking at the adjusted R-square, multicollinearity issues, and normal distribution graphs. Furthermore, the results of the regression analysis and the segmentation analysis are presented. In the end, we see if the segmentation criteria are met.

§ 4.1 Validity and Reliability

Let us start with discussing validity and reliability of our analysis. We found four outliers, so we continue our analysis with 103 cases. Our adjusted R-square is 0,435, which indicates that the variables we use in our regression analysis explain 43,5% of the variance in profitability (see Appendix D). This R-square is higher than the R-square of our initial model, which has an adjusted R-square of 0,314 (see Appendix E). So by adjusting our model, our R-square has improved.

We also checked for multicollinearity problems (figure 6). Large VIF values, usually 10 or more, suggest multicollinearity (Cooper & Schindler, 2006). There are no VIF values are above 10, so this is an indication that multicollinearity is not an issue.

Appendix F shows that we are dealing with a normal distribution. The distribution of the residuals matches the normal distribution rather well.

§ 4.2 Results

In this section we discuss the results of the analysis. We first present the results of the

4.2.1 Results Multiple Regression Analysis

Now let us examine if we can accept the stated hypotheses. We therefore look at the significant relationships found from the regression analysis and that are used in our final model. The variables that are entered in our final model are shown in figure 6:

Results Regression Analysis with dependent variable Profit

Model Unstandardized Coefficients Standardized Coefficients t Sig. Collinearity Statistics

B Std. Error Beta Tolerance VIF

(Constant) 3094,178 1149,302 2,692 0,009 Demographics Children ex -452,118 211,252 -0,223 -2,140 0,037 0,741 1,350 Physical Disabilities 1575,427 618,054 0,283 2,549 0,014 0,655 1,526 Alcohol Drugs -464,351 241,525 -0,191 -1,923 0,060 0,815 1,227 User Status GGZ 1916,582 610,685 0,333 3,138 0,003 0,718 1,394 Other -1258,890 625,266 -0,210 -2,013 0,049 0,742 1,348 PGB1 2049,797 803,039 0,342 2,553 0,013 0,450 2,223 PGB2 or higher 6739,395 1294,428 1,040 5,206 0,000 0,202 4,948 Local Authorities 1771,120 751,900 0,343 2,356 0,022 0,380 2,631

Time per Visit -2000,920 490,793 -0,842 -4,077 0,000 0,189 5,292

Service Usage Social Activation -1694,030 626,138 -0,269 -2,706 0,009 0,816 1,225 Application Assistance -1639,710 701,926 -0,260 -2,336 0,023 0,649 1,540 Housing 2525,023 606,961 0,421 4,160 0,000 0,788 1,270 Assistance at Home 1127,371 597,269 0,188 1,888 0,064 0,813 1,230

Figure 6: Significant Variables Used in Model

Regarding the demographics, we did not find evidence to support the hypotheses for most of the relationships. As expected, we did find a negative relationship between having children with an ex and profitability. We also find a positive relationship between having a physical disability and profitability. Furthermore we find a negative relationship between having an alcohol or drugs-addiction and profitability. Therefore we:

accept H5: There is a negative relationship between having children with an ex and

profitability.

accept H8: There is a positive relationship between having a physical disability and

accept H10: There is a negative relationship between having an alcohol or drugs-addiction

and profitability.

For the other demographics we did not find significant evidence to accept the hypotheses, therefore we:

reject H1: There is a relationship between age and profitability; and reject H2: There is a relationship between gender and profitability.

reject H3: There is a negative relationship between family size and profitability. reject H4: There is a relationship between marital status and profitability.

reject H6: There is a positive relationship between being a “native” and profitability and a

negative relationship between being a “foreigner” and profitability.

reject H7: There is a negative relationship between educational level and profitability reject H9: There is a positive relationship between intensity of mental disabilities and

profitability.

reject H11: There is a negative relationship between having a criminal record and

profitability.

reject H12: There is a negative relationship between months of unemployment and

profitability.

We did find a relationship between getting help from other organizations and profitability. Getting help from GGZ has a positive influence. Getting help from other instances or people has a negative relationship with profitability. So we can conclude that the relationship can be either positive or negative. Therefore we:

accept H13: There is a relationship between getting help from other organizations and

profitability.

Regarding the sponsors, we can not accept H14. The reason for this is that we had to few cases with UWV clients to find significant evidence. For the remaining variables we can conclude that PGB2 had the highest relative impact on profitability, followed by PGB1 and last by the local authorities. Due to the lack of sufficient cases, we:

reject H14: UWV clients have the highest relative impact on profitability, followed by PGB2,

We did not find a significant relationship between de duration of a contract and profitability and between the frequency of visits and profitability, therefore we:

reject H15: There is a relationship between the duration of a contract and profitability. reject H16: There is a positive relationship between the frequency of visits and profitability.

In contrary to our expectations, we find a negative relationship between time per visit and profitability. Therefore we:

reject H17: There is a positive relationship between the time per visit and profitability.

Regarding service usage, we did not find positive relationships with profitability for all variables. However, we did find a positive relationship between the service “housing” and profitability. The services “social activation” and “application assistance” have a negative relationship with profitability. Therefore we:

reject H18: There is a positive relationship between service usage and profitability.

4.2.2 Results Segmentation Analysis

The results of the segmentation analysis are displayed in the customer pyramid (figure 7). X% of the total customers are represented in the lead tier, with an average loss of € X,

contributing -X% to the total profit. X% of the total customers are represented in the iron tier, with an average profit of € X, contributing X% to the total profit. X% of the total customers are represented in the gold tier, with an average profit of € X, contributing X% to the total profit. X% of the total customers are represented in the platinum tier, with an average profit of € X, contributing X% to the total profit.

In order to profile the different tiers in the customer pyramid, we look if there are significant differences between the tiers. The results of the Levene test are shown in Appendix G. For the variable “administrative” we can use a normal F-test (Appendix H). For the remaining

variables we use the Brown-Forsythe and Welch test (Appendix I). The variables that show significant differences between tiers are the following:

PGB1

Welch and Brown-Forsythe test show that “PGB1” is significant on a 95% confidence interval (figure 6). The mean-plot (graph 1) shows that the platinum tier has the most people with PGB1. Lead Total profit: € -12.239,84 Average profit: € -453,33 Iron Total profit: € 36.739,20 Average profit: € 1.020,53 % of profit: -6,12% % of customers: 26,00% Gold Total profit: € 104.659,75 Average profit: € 3.171,51 Platinum % of profit: 18,37% % of customers: 35,00% % of profit: 52,33% % of customers: 32,00% % of profit: 35,42% % of customers: 7,00% Platinum tier Total profit: € 70.847,59 Average profit: € 10.121,08

Graph 1: Mean-plot PGB1

PGB2 or Higher

The Brown-Forsythe test show that “PGB2 or Higher” is significant on a 95% confidence interval (figure 6). The mean-plot (graph 2) shows that the platinum tier has the most people with PGB2 or higher.

Graph 2: Mean-plot PGB2 or Higher

Local Authorities

Graph 3: Mean-plot Local Authorities

Months Unemployed

Welch test shows that “Months Unemployed” is significant on a 95% confidence interval (figure 6). The mean-plot (graph 4) shows that the iron tier has many people with a high number of unemployment. The gold tier has a low mean, while the platinum tier has no people who were unemployed at start of the track.

Graph 4: Mean-plot Months Unemployed

Empowerment

Welch and Brown-Forsythe test show that “Empowerment” is significant on a 95%

Graph 5: Mean-plot Empowerment

Psychosocial

Welch and Brown-Forsythe test show that “Psychosocial” is significant on a 95% confidence interval (figure 6). The mean-plot (graph 6) shows that the platinum tier has many people who get psychological treatment. For the gold tier however, the mean is rather low.

Graph 6: Mean-plot Psychosocial

Administrative

tier, he probably does not receive administrative support.

Graph 7: Mean-plot Administrative

Total Services

Welch test shows that “Total Services” is significant on a 95% confidence interval (figure 6). The Post-Hoc analysis Games-Howell shows that the iron tier differs from the platinum tier. The mean-plot (graph 8) shows that starting from the iron tier, the higher the number of services are, the higher the profitability is.

Graph 8: Mean-plot Total Services

Segmentation criteria

In order to see if segmentation is effective for Company X, we briefly discuss each segmentation criteria:

Measurable: The customer pyramid shows that size, purchasing power and profiles can be measured.

Substantial: The segments that we focus on in this research, the lead tier (26% of the customers) and iron tier (35% of the customers), are large enough. They are not profitable enough, but the scope of this research is to improve this.

Accessible: The segments can be reached and served during an intake, this will be discussed further in chapter 5.

Differentiable: Analysis of the differences showed that there are significant differences between segments.

Actionable: Effective programs can be formulated, as will be discussed in chapter 5.

5 Conclusions and Recommendations

In this chapter we draw conclusions and come up with recommendations. We provide a model for predicting profitability, we profile the segments, look at the CSR at Company X, come up with a strategy for approaching the segments, give advice regarding CSR communication and provide limitations of this study and directions for further research.

§ 5.1 Predicting Profitability

We can use the results of the regression analysis for making a model for the prediction of the profit for a client. The significant variables that are presented in figure 6 are used for our model to predict the profitability of a client. Therefore the model is:

Profit Client = 3094,18 + 2049,797* PGB1 + 6739,395* PGB2_or_higher + 1771,120*

Local_Authorities - 2000,923* Time_per_Visit + 1127,371* Assistance_at_Home + 1916,582* GGZ - 1258,893* Other - 452,118* Children_ex + 1575,427*

Physical_Disabilities - 464,351* Alcohol_Drugs - 1694,030* Social_Activation – 1639,708* Application_Assistance + 2525,023* Housing

This model can be used during an intake. By asking specific questions, an employee can retain all the variables in our model. By filling in the model in an excel-sheet, profit can be estimated and the employee will know in which profitability tier the client belongs. In that way an employee can decide what strategy to follow regarding this client.

§ 5.2 Profiling the Segments

From the discussion of the segmentation criteria we can conclude that we meet all the criteria for effective segmentation. Therefore we can use the results of the segmentation analysis for profiling the segments. We also use some general demographics, but more importantly are the significant differences found between the tiers as identified early.

Lead Tier