Antecedents of continuance commitment in the B2B financial service sector

(1)

Antecedents of continuance

commitment in the B2B financial

service sector

Author:

N. van den Berg

(2)

UNIVERSITY OF GRONINGEN

FACULTY OF ECONOMICS AND BUSINESS

MSc Marketing Intelligence Master Thesis

Antecedents of continuance commitment in the B2B financial service sector

Author:

N. van den Berg (s1824147) Pleiadenstraat 34 2024TP Haarlem E: n.van.den.berg.1@student.rug.nl 1st Supervisor: Dr. J. T. Bouma

2nd Supervisor: Prof. Dr. J. E. Wieringa

(3)

Management summary

Since companies exist there has been the question of how to make customers loyal and hence more profitable. Uncountable academic scholars have devoted their work to find the underlying patterns and factors determining the loyalty of customers. Nowadays we know that customer loyalty is a complex construct deploying various forces of affective and behavioral nature and incorporating the construct of commitment. The latter can also be divided by its antecedents; affective and calculative. In the service industry and especially in the financial service industry between businesses, the specific knowledge of how customer commitment and loyalty are related and what drives them is of great importance for marketers. This comes due to the fact that most businesses maintain

relationship with various banks, where banks desire to have a more exclusive relationship with their clients.

This study gives an overview about the different concepts and their relation towards each other before focusing on calculative or continuance commitment and its antecedents. A literature study reveals potential predictors of

continuance commitment, which are then used in the analysis part in order to examine their role in the business-‐to-‐business (B2B) financial service sector. Two models including the drivers found in literature are estimated and tested on extensive market research data in the Dutch market.

The results indicate that in the prediction of the client’s most important bank, trust and the willingness to recommend the bank play a minor or insignificant role. The strongest predictors are the actual services taken from the banks with differing significant impact. Also brand preference plays a crucial role in the choice for banks services and the resulting switching costs.

(4)

(5)

Introduction

Recent years have seen troubling times for banks and other financial institutions throughout the world and especially in Europe and the US. Banks were driven into bankruptcy or had to accept bailouts from governments, which came as a package deal with newly imposed requirements. And therefore today, the backwater of the financial crisis requires banks to bolster their competitive positions in order to regain lost strengths and to comply with new regulations forced upon them by national and international commissions. According to Ernst & Young’s Global-‐Consumer-‐Banking-‐Survey from 2014, customer confidence in the banking sector is returning from the crisis. Furthermore, the report, which surveyed 32,000 private banks customers world wide, pointed out that

customers are on the move “with unprecedented access to competing banks and to new types of financial service providers.” (Ernst & Young, 2014; p. 2). The report further urges banks to gain a maximum of trust in order to broaden market share and create genuine loyalty.

In order to bolster the competitive position and market share banks can deploy a variety of measurements of which the customer loyalty is one. According to Boulding, et al. (1993) the repurchase willingness and the willingness to

recommend are consequences of customer loyalty. Also LaBarbera and Mazursky (1983) find advantages of loyal customers in their willingness to paying a

premium and showing lower defection rates.

(7)

(1990) and Fullerton (2003) split commitment up into two sub constructs; affective and continuance commitment.

This study aims at shedding light into the antecedents of continuance

commitment in order to provide an answer to the research question: What are the Drivers of bank patronage sentiment? The insights will give marketing

management a guide in the quest for the right balance in relationship marketing. The course of action follows the common two-‐step approach where a literature study is used to build up a conceptual model which is then tested in an empirical study itself. Therefore a literature study will summarize the concepts of loyalty and commitment and their relationship towards each other. Then, the concepts of trust, recommendation and brand preference are discussed as drivers of loyalty and commitment, resulting in the formulation of various hypotheses and the conceptual model. Subsequently the data and method of analysis are

introduced and discussed before the results are reported. The following critical discussion of the results is concluded with implication of the findings and suggestions for further improvement of the study.

(8)

Theoretical framework

Customer loyalty and commitment are intertwined concepts with various factors driving them. In order to position this study in the world of loyalty studies, the following theoretical framework will summarize and discuss the findings in academic literature, which ultimately lead to the formulation of the hypotheses and the conceptual framework.

Loyalty

Customer loyalty has evolved into a major subject resulting in an essential

concern for managers and a strategic obsession for many executives. The reason for this is the increasing competition, particularly in the service industry, and the focus on the relationship between the company and its customers, which comes forth in the relational marketing approach (Bodet 2008).

The rather lucrative effects of customer loyalty on business performance have already been mentioned. In order to become a beneficiary of those effects companies have to master the antecedents and deploy factors such as product and service quality, which have a positive impact on customer loyalty. This was proven by Yonggui, Hing-‐P. & Yer (2003) who investigated these factors in the Chinese retail banking sector and concluded that it is of utmost importance for Chinese banks to improve both service and product quality in order to gain customer loyalty and hence a competitive advantage in the long run.

(9)

Dick and Basu (1994) and Bandyopadhyay and Martell (2007) succeeded to expose the positive but rather latent influence of attitudinal loyalty on behavioral loyalty.

As one of the earliest scholars in customer loyalty studies Tucker (1964) argued that the entire construct of loyalty is to be found in behavior namely in the form of past purchases of brands or products. Jacobi and Chestnut (1978) confirm in their study that behavioral loyalty studies emphasize the interpretations of purchase patterns from panel data as the nature of loyalty. From this

perspective, loyalty is believed to be stochastic rather than deterministic (Uncles and Laurent 1997). However, an attitudinal component of loyalty cannot not be neglected as it includes for example the concept of word-‐of-‐mouth discussed by Zeithamel, Berry and Parasuraman (1996) or encouragement to use the product or service as found by Bettencourt and Brown (1997).

After all, a holistic approach as discussed by Day (1976) accounts for all expected and unexpected effects by including behavioral as well as attitudinal loyalty effects on firm performance. This holistic approach also includes non-‐users of a product or service as a customer groups because of their potential to become customers and the attitude they hold even when not personally using the good or service. The holistic perspective also allows the distinction between true loyalty and spurious loyalty as suggested by Day (1976).

In order to gain a better understanding of the loyalty concept researchers succeeded in integrating the construct of customer commitment into the view of loyal customer behavior (Fullerton, 2005; Zins, 2001). The studies by Fullerton (2005) and Zins (2001) deploy commitment as a pivotal mediator in the

relationship between customer’s evaluation of a firm’s performance and the customer’s intention concerning the future relationship with the firm (Fullerton 2005). Fullerton (2005) and Bloemer and Kasper (1995) integrate these

concepts in an attempt to discriminate between true and spurious loyalty. According to Li and Petrick (2010), the attitudinal component of loyalty can be equated to the concept of commitment. Although they investigate customer loyalty in the cruising tourism sector, their conclusion remains of interest since it is yet another view on the loyalty -‐ commitment relationship.

(10)

Commitment

According to Bügel, Buunk and Verhoef (2010), there are significant differences in the level of commitment among different industry sectors. The authors compared the banking industry with health insurance, supermarkets, mobile telecom and automotive sectors and found that the banking industry shows the highest commitment among its clients. The authors suggest that these findings are the result of the continuous service banks offer compared to the switch-‐over moments in the other sectors.

The concept of commitment is comprised of an unknown share of affective and calculative or continuance commitment. Zins (2001) argues that there are different antecedents, contents and consequences, which allow the distinction between liking and identification (affective) and dependence and switching costs (continuance). Allen and Meyer (1990) even distinguish between three forms of commitment in their study of employee-‐ employer relationships in the field of organizational behavior. The two already mentioned are completed by normative commitment, which describes a moral pledge of the employee to stay with the company.

The concept of affective commitment is characterized as the consumer’s sense of belonging and involvement with a business partner similar to emotional bonding (Rhoades, Eisenberger and Armeli, 2001; Fullerton, 2003).

Whereas De Ruyter, Wetzels and Bloemer (1998) describe calculative

(continuance) commitment as the manner the customer is forced to remain loyal against an ulterior desire. Fullerton (2005) underlines that continuance

commitment stems from the feeling that ending the relationship entails a social or economic sacrifice also called switching costs.

Whereas affective commitment has been studied in brand image studies and relationship marketing, continuance commitment has not yet received broad attention in marketing scholarship (Gruen, Summers & Acito 2000).

(11)

influence on staying intentions than continuance commitment. More consistency is found in the field of relationship marketing where Fournier, Dobscha and Mick (1998) and Grayson and Ambler (1999) find that CRM programs, which are supposed to tie the customer tighter to a company, may not have the anticipated positive effects on customer retention. The complex interplay between

continuance and affective commitment is further investigated by Fullerton (2003) and the main findings are shown in figures 1 2 and 3. The author shows the effect of high versus low continuance commitment and high versus low affective commitment on switching intentions, advocacy intentions and the willingness to pay more. The essence is that customers are not content with an increased feeling of continuance commitment. This can be seen in the rising switching intentions and falling advocacy intentions in the low affective condition. The high affective condition however shows a reversal in effects. If customers feel affective commitment to the supplier, they do not mind the increased continuance commitment feeling. The odd one out is the case of the willingness to pay more. Here, the continuance commitment feeling drives up the willingness to pay more in the low affective condition whereas the high affective condition sees a decline in the willingness to pay more while increasing

continuance commitment.

Figure 1 – Switching Intentions Figure 2 – Advocacy Intentions Figure 3 – Willingness to Pay More Fullerton, 2003

(12)

Whereas the effect of affective commitment seem to exceed the effects of continuance commitment in the business to customer relationship, Verhoef (2003) finds that in the business to business environment affective commitment exerts a significant, but rather small positive influence on customer retention and share development. In fact, the consequences of affective commitment are very similar to the consequences of economical incentives like loyalty programs on customer retention and share development. Therefore in the business to business (B2B) context it is of utmost importance to investigate both fragments of commitment and especially the concept of continuance commitment.

Continuance Commitment as the Dependent Variable

The Dutch and the German parlances deploy the term “huisbank” and

“Hausbank” respectively, in order to label a bank as the most important one for a business. In academic literature however, there is not much consensus about the terminology and definition of whatever constitutes a ‘house bank’ (literally translated from ‘huisbank’), or a customer’s ‘preferred bank’. The German scholar Fischer (1990) defined the term ‘house bank’ (pp. 3; 4) by introducing four features, which help to characterize a bank as a ‘house bank’. The first feature is that the bank accounts for the largest share of external finance and the largest share of financial services consumed by a customer. Secondly, the

relationship must involve a long-‐term orientation from both sides including a high level of trust between the parties. The two mentioned features result in the third, which is the influential role of the financial institution in the customer’s decision making. The last feature is the role of the financial institution in times of distress or need of restructuring. Whereas these are intuitively logic

(13)

The bank patronage sentiment as indicated by the appointment of “the most important bank” will be used as the dependent variable in this study.

Commitment sentiment is a very common dependent variable in business relationship studies (Anderson, Lodish and Weitz 1987; Anderson and Weitz 1990; Jackson 1985; Dwyer, Schurr and Oh 1987) and very well suited to discriminate between defectors and loyal customers (Mummalaneni 1987). According to Wilson (1995) “commitment implies importance of the relationship to the partners and a desire to continue the relationship into the future” and hence resembles bank patronage as deployed in the dependent variable.

Continuance commitment, bank patronage sentiment, house bank and the “most important bank” are hereafter used as synonyms describing patronage and dependence sentiment towards a financial institution as the provider of the most vital financial services.

Trust

(14)

Customer Recommendation

Since the publishing of Fred Reicheld’s “The one number you need to grow” in HBR in 2003, customer recommendation and the Net Promoter Score (NPS) as predictors for individual customer loyalty and even as predictors for firm growth have been a hotly debated topic. Despite various immediate critics from the academic field (Kristensen and Westlund, 2004; Morgan and Rego, 2004) the NPS has gained broad popularity and is now applied by numerous companies serving as an indicator of customer loyalty and ultimately even as indicator for company growth potential. The diligent application of the NPS by practitioners throughout any sectors makes customer recommendation an interesting factor for the analysis of individual bank patronage sentiment.

Brand Preference

Marketing practitioners have long deployed brand preference in its role as antecedent of customer loyalty. The degree of preference for a specific brand compared to competing alternatives is a central component in the construct of customer loyalty (Rundle-‐ Thiele and Mackay, 2001). A shared antecedent of brand preference and purchase intention is brand equity as proven by Cobb-‐ Walgren, Ruble and Donthu (1995). Furthermore, brand preference was found a significant moderator on the customer satisfaction – customer behavioral loyalty (as measured by share of wallet) relationship (Keiningham, et al., 2005). Other scholarly works (Bennett and Rundle-‐Thiele, 2002) have described attitudinal loyalty as preference for a brand. Therefore it is very interesting to find the role brand preference plays in the B2B financial service setting.

Hypotheses

Predictors and drivers of customer loyalty and commitment are to be found numerously in literature. That is the reason why this theoretical framework will focus on the service sector in the B2B area however, without ignoring influences by predictors discussed in other fields. Drivers of customer loyalty and/or commitment found in literature are discussed hereafter.

(15)

Services

Fullerton’s (2003) findings, that consumers face themselves in a (continuance) committed relationship if they perceive significant switching costs or if the product or service is not easily replaceable by another provider, are confirmed by Bendapudi & Berry (1997) and Dwyer, Schurr & Oh (1987). This form of commitment can also occur when one of the partners feels dependent on the relationship (Heide and John 1992). Another source of continuance commitment lies in the action of one party to bind the other party by pledges and investments (Becker, 1960) or contracts and service agreements (Anderson and Weitz 1992). Perceived switching costs, the lack of alternatives and dependence compile the antecedents in the concept of continuance commitment.

(16)

sentiment increases with every single product taken. This means also that every product makes a contribution on its own.

Hypothesis 1a: The possession of a credit at the focal bank is positively related to the likelihood that the client sees this bank as the most important bank.

Hypothesis 1b: The possession of a mortgage at the focal bank is positively related to the likelihood that the client sees this bank as the most important bank.

Hypothesis 1c: The possession of a savings account at the focal bank is positively related to the likelihood that the client sees this bank as the most important bank.

Hypothesis 1d: The possession of an investment portfolio at the focal bank is positively related to the likelihood that the client sees this bank as the most important bank.

Hypothesis 1e: The possession of a real estate leasing contract (Leasing) at the focal bank is positively related to the likelihood that that the client sees this bank as the most important bank.

Hypothesis 1f: The possession of a leasing contract for company goods (Leasing 1) at the focal bank is positively related to the likelihood that the client sees this bank as the most important bank.

Hypothesis 1g: The possession of a leasing contract for the vehicle park (Leasing 2) at the focal bank is positively related to the likelihood that the client sees this bank as the most important bank.

(17)

Number of Insurances

As described earlier, the bandwidth of products or services taken by a customer influences the satisfaction – repurchase chain. Even in continuous relationships with very few interactions as in the case of insurances the commitment felt by clients increases with every insurance procured. Therefore, bank patronage is expected to increase with the number of insurances procured.

Hypothesis 2: The number of insurances procured at the focal bank is positively related to the likelihood that the client sees this bank as the most important bank.

Mobile Banking

Mobile banking implies the usage of mobile devices to execute bank transaction and other financial related activities with an app, provided by the bank. Since Biedenbach and Marell (2010) point out the importance of positive customer experience in the B2B setting through direct interaction, this driver is included in the analysis. Murray and Häubl (2007) found that frequent customer

experience of a product increases the habit-‐forming usage of the product, which in turn exerts a cognitive lock in effect. This cognitive lock in effect positively influences consumer preferences for the particular product (Muray and Häubl, 2007). Johnson, Bellman and Lohse (2003) investigated the cognitive lock in concept in the field of website usage and found that frequent, habit forming usage of websites increases the browsing speed. The websites with the fastest learning curves also show the highest rates of purchasing (Johnson, Bellman and Lohse, 2003). Therefore, due to mobile interaction with the bank through the application the bank client relationship is believed to influence the patronage probability positively.

Hypothesis 3: The usage of mobile banking at the focal bank is positively related to the likelihood that the client sees that bank as the most important bank.

(18)

Enabled Negative Balance Facility

Most banks offer the possibility to enable the client to overdraw his or her

account in order to buffer oscilliating in-‐ and outflow of money. The arrangement for this facility often requires an application by the client. An existing facility of overdrawing the account is hence an indication of intended regular to frequent usage by the client. Since Murray and Häubl (2007) found that frequent

customer experience subsequently leads to a cognitive lock in, frequent usage of the bank account is expected to increase bank patronage. The mentioned lock in effect and the dependence on the frequently used account are also synonyms of switching costs, which would occur if the account would be canceled.

Hypothesis 4: The possibility to overdraw the bank account at the focal bank is positively related to the likelihood that the client sees that bank as the most important bank.

Number of services from another financial institution

The failure to not spreading the finanincial services among different institutions can result in a perceived or actual dependency of the client. According to

Dowling and Uncles (1997), clients who perceive a scarcity of other option, except continuing the ongoing relationship, exert a passive behavior and

sometimes even develop emotional bonds. Therefore, the number of products or services taken from a competitive financial institution are negatively related to the felt dependency of a client towards the first bank. The financial sector in the Netherlands knows three large players who share the vast majoriyty of market share and a rather small rest market share which is divided by more than 12 smaller institutions. Therefore, the services taken by the two largest competitor banks to the focal bank will be included in the analysis.

Hypothesis 5a: The number of products taken from another bank

(19)

Hypothesis 5b: The number of products taken from another bank

(competitor B) than the focal bank is negatively related to the likelihood that the client sees the focal as the most important bank.

Number of Services taken from the Focal Bank

According to Bloemer and de Ruyter (1999) the degree of involvement makes a difference in moderating the satisfaction-‐ loyalty relationship. The authors test the relationships of customer satisfaction and customer loyalty and assess the role of product involvement in a moderating function to this relationship. They find that services in a high involvement context affect the consumer’s emotional states, which in turn moderates the satisfaction-‐ loyalty relationship. Low involvement services do not play a significant role in the satisfaction-‐ loyalty relationship. Since Biedenbach and Marell (2010) prove that positive product experience is important as a driver of loyalty it is assumed that the number of products drives the feeling that the bank is the main bank of the customer.

Hypothesis 6: The number of services taken from the focal bank is positively related to the likelihood that the focal bank is seen as the most important bank by the client.

Trust

Next to the behavioral, product-‐related drivers there are also other attitudinal-‐ based antecedents of customer loyalty to be found in academic literature. Rauyruen, Miller and Barret (2007) investigate the relationship quality as a predictor of B2B customer loyalty. For this purpose they survey more than 300 SMEs in the courier and freight delivery service industry in Australia. The authors use trust, commitment, satisfaction and service quality as elements

comprising relationship quality. They find that all four aspects increase customer loyalty of which trust should actively be promoted by the supplier and

(20)

than rational characteristics. In order to maintain purchase intentions, satisfaction is a crucial element, whereas service quality increases both purchases intentions and attitudinal loyalty.

Service quality is also to be found a strong determinant of behavioral intentions by Zeithamel, Berry and Parasuraman (1996). In fact, the perceived absence of service quality leads to unfavorable customer behavior like complaining, which is assumed to predict or accompany defection (Richins, 1983; Scaglione, 1988). Theron, Terblanche and Boshoff (2010) also examine the antecedents of

relationship commitment in the B2B financial service sector. They however, focus more on affective characteristics of the commitment in the bank – client relationship. The investigated drivers are trust, communication, shared values and attractiveness of alternatives. All those drivers play a significant role in the management of commitment in a relationship.

Since trust is found to be a positive driver of customer loyalty it is assumed that with increasing trust the number of products are taken increase and also the patronage behavior increases.

Hypothesis 7a: Trust (Trust1) is positively related to the likelihood that the focal bank is seen as the most important bank by the client.

Hypothesis 7b: Trust (Trust2) is positively related to the number of products taken from the focal bank

Willingness to Recommend

The Willingness to Recommend (WTR) has been a debated topic since Reichheld’s publication in 2003. Also the role of WTR in the loyalty construct has been

discussed by Fullerton (2003). Advocacy intentions have according to Fullerton (2003) a significantly influenced by affective commitment in the two conditions of high versus low continuance commitment. Since a high WTR is believed to be an indication of attitudinal commitment (Zeithaml, Berry & Parasuraman, 1996) WTR should moderate the products taken – customer patronage feeling

(21)

feelings are being strengthened with an increase in the willingness to recommend the service provider. Also the relationship between brand

preference and patronage feelings is possibly moderated by the willingness to recommend. Finally, since a high WTR is an indicator of attitudinal commitment, it also has a positive direct effect on patronage feelings.

Hypothesis 8a: Willingness to recommend (WTR1) the focal bank has a positive influence on patronage sentiment towards the focal bank

Hypothesis 8b: Willingness to recommend (WTR2)has a positive influence on the relatioship between the number of products taken from the focal bank and the patronage sentiment towards that bank

Hypothesis 8c: Willingness to recommend (WTR3) has a positive influence on the relatioship between brand preference for the focal bank and the patronage sentiment towards that bank

Brand Preference

Hellier, et al. (2003) define brand preference as the “extent to which the

(22)

recommend, assessing whether there are differences among customers who would like to recommend the bank and customers who would not.

Hypothesis 9a: The statement of brand preference (BP1) for the focal bank is positively related to patronage sentiment towards that bank

Hypothesis 9b: The statement of brand preference (BP2) for the focal bank is positively related to the number of products taken from the focal bank

Hypothesis 9c: The relationship between the brand preference

(BPmediated) for the focal bank and patronage sentiment towards the focal bank is mediated by the number of products taken from the focal bank.

Control Variables

In order to control for a potential effect of company size and age, two more variables are included. The two variables indicate whether the company recently started operations (Starter) and whether the business is only run by a self

employed individual (ZZP). However, since the emphasis of this study does not lie on these mechanisms, these variables will merely be included as control variables.

Starter

This control variable is dummy coded, showing a “1” for businesses younger than three years old and a “2” for businesses older than three years. Due to the

turbulences in the financial markets and the resulting blame on banks and their practices one might assume that younger businesses tend to favor small and young financial institution who have not had the same negative press coverage as the focal bank and the large competitors. This estimation however is based on intuition and has not been confirmed by academic research yet.

(23)

Self-‐employed/ ZZP

The abbreviation ZZP comes from the Dutch “Zelfstandig Zonder Personeel” and describes the business form in which the entrepreneur represents the entire workforce of the business. As the previous variables, also this variable is dummy coded showing a “1” in case the respondent is a ZZP-‐ business and a “2” in case the respondent is not answering as a ZZP business. Also for this mechanism there has not been a study, which can be cited here, and therefore the analysis will hopefully shed more light into the matter.

(24)

(25)

Data

In order to test the hypotheses, which have earlier been extracted from theory, quantitative analyses of two Dutch market research datasets are chosen. In the following section the datasets are introduced and discussed before the samples are examined and prepared for use.

The Dutch Market

This study will be conducted in the Dutch market. In comparison with the

European average, the Dutch market is relatively large in size as measured in the percentage of GDP. At the end of 2012 the size of the banking sector in the

Netherlands amounted to 4.5 times the national GDP whereas the European average (EU-‐ 15) lies at 3.6 times (Dutch Finance Ministry, 2013). At the same time, the Dutch bank sector is relatively concentrated since the market shares of the three largest Banks ING, ABN AMRO and Rabobank combined amount to 74% in 2011 (Dutch Finance Ministry, 2013). Of those 74%, 34% belong to the ING Bank, 26% to the Rabobank and 14% to the ABN AMRO (Dutch Finance Ministry). One of those three banks is chosen and hereafter referred to as “the focal bank”, whereas the other two large banks are referred to as the two competitors.

Data Source

The core of this study consists of two quantitative analyses of two cross sectional market research datasets, which are called “financial monitor” and “image-‐

building monitor”. In this section, both datasets are described and discussed one at a time. The focal bank and the competitors are the same across the two

datasets.

The Financial Monitor

(26)

10.600 respondents who represent a business with an entry in the Commercial Registry of the Netherlands. Those respondents are approached via a panel, which TNS NIPO maintains, and via the entries in the Commercial Registry of the Netherlands. The issuer claims that the proportions in the segments of small to medium sized enterprises (SME’s) mirror the proportions in the Dutch market relatively accurate. In order to achieve the representativeness of the sample the proportions are tested for number of employees, turnover, age of business and sectors.

The survey is being conducted throughout the entire year, which, according to the research agency, levels out fluctuation during a year. Respondents are approached once a year as a maximum to avoid a bias due to weariness of the population and also in order to decrease the likelihood of null-‐ responses. Another measurement, which is supposed to increase valid responses, is the deployment of the ‘mixed-‐mode technique’ that enables the respondent to participate in the survey via phone, online or face-‐to-‐face, whichever way suits the respondent best. The survey is conducted in Dutch, which is appropriate since the receivers are owners or leaders of businesses operating in the Netherlands. Abstruse language and other language-‐based problems are

minimized since the survey is conducted for more than eight consecutive years including continuous feedback and improvement. The yearly results are bases to multi-‐million Euro decision on the highest management levels of the three main banks ING, Rabo and ABN Amro. The, for this study relevant part of the

questionnaire, can be found in appendix exhibit J.

Sample Description of the Financial Monitor

(27)

0 5 10 15 20 25 30 35 Representation of Company Sizes % 0 5 10 15 20 25 30

Representation of Sectors

%

needs of the companies. After the exclusion of the large companies, the

companies with a maximum of 50 employees remain in the dataset amounting to 9344 cases. In order to examine the representativeness of the sample, various investigations have been executed where the most significant are described hereafter.

Figure 4 and 5 show the distribution among the sectors and the number of employees. From the figures we can see that the sample covers all the industries not equally but well. Also the different sizes of companies are distributed well. According to the Dutch bureau for statistics (CBS) 19% of all businesses in the Netherland in 2007 operated in the commercial service sector. With a slight increase in recent years, this sector amounts to 24% in 2013 in our dataset. This random test shows the adequate accuracy of the proportion in the sample.

Figure 4 – Representation of Sectors (left)

Figure 5 – Representation of Company Sizes (right)

After the examination of the proportions, the sample is checked for oddities and systematic missing values on sight using various descriptive techniques.

(28)

though not all of the earlier mentioned proportions are precisely met, the sectors and company sizes are sufficiently covered.

All in all it is concluded, that the dataset is an adequate base for the testing of the hypothesis.

It is chosen to include all the sectors in order to gain universal insights into the mechanisms at work in the Dutch market.

Preparation of the Dataset from the Financial Monitor

The variables indicating the possession of one of the products (hypothesis 1a – 1g) will be coded categorical with an “0” indicating no possession of the product at the focal bank and “1” indicating the possession of the service from the focal bank. The variables of the usage of mobile banking (h. 3) and for the negative balance facility (h. 4) are coded the same way. The variables indicating the number of insurances procured at the focal bank (h. 2) are coded “0” for none and “1” to “6” indicating the number of insurance produced at the focal bank. The variables indicating the number of products taken from the competitor banks (h. 5a & h. 5b) are coded with the values covering the range of “1” to “10”. With the exception of the dependent variable “most important bank” are none of the variables mutually exclusive between the banks. This means that respondents can have accounts and procure insurances at several banks.

Figure 6 visually portraits the characteristics of the market penetrations across the three Banks. The variables indicating the number of insurances and number of products are analyzed using the cases of which the values comply with >= 1.

Figure 6 – Market Penetration of the three largest Banks

(29)

The Image-‐Forming Monitor

The image-‐forming monitor (IFM) will be serving as the base for the second part of the analyses. The IFM consists of an online-‐based survey of the market

research agency Kien and is conducted three times a year in order to publish an annual report. The population consist of individuals of 16 years and older who work in positions with financial authority in SME companies with an entry in the Dutch chamber of commerce. The publication is representative with respect to gender, age, educational level, family situation and work participation. The research is conducted according to the norms of ISO 20252 (market research) and ISO 26362 (access panels). In the first part of the survey respondents are asked demographical questions followed by a set of questions that clarify at which bank the respondent takes which products. Subsequently only the clients of the three largest banks are further asked if they trust their bank and if they would recommend it further. The respondents are also asked which bank they prefer in the hypothetical event of a product need. The, for this study relevant part of the questionnaire, can be found in appendix exhibit K.

Sample Description of the Image-‐Forming Monitor

(30)

exhibit A Figure 1 and 2. As can be seen, various company sizes and young as well as old are included.

Preparation of the Dataset from the Image-‐Forming Monitor

The sample is scanned visually for oddities, however there were none detected. Next to the products Account, Credit, Saving, Insurance and Investment also trust and likelihood of recommendation are asked. Trust is measured on a scale from “1”, indicating no trust, to “9”, indicating full trust. The likelihood of

recommendation is measured on a scale of “0” to “10” according to the NPS measurement system. As can be seen in Appendix exhibit A Figure 3 the differences in average trust and recommendation willingness across the three most important banks are relatively small also concerning their standard deviation. Whereas Appendix exhibit A Figure 4 shows the market share of the banks as most important banks in the customer perception.

The dependent variable “most important bank” is of dichotomous nature and thus identical to dependent variable from the analysis of the FM. The variable “services from focal bank” is coded “0”, indicating no product from one bank to “5” indicating all services (from this survey) are taken from one bank. The variables “recommendation” and “trust” are coded as described before. The variable “brand preference” is coded dichotomous with “1” indicating no preference for the focal bank and “2” indicating brand preference for the focal bank. Finally, the two control variables are coded the same way as in the previous analysis.

Figure 7 depicts a summary of all the variables and their different names used. The name of the focal bank and of the competitor banks are indicated by “[…]” in order to preserve the anonymity of the banks.

(31)

Variable

Name Short Name (if applicable) Name in Dataset (if different) Coding Dataset

Credit -‐ -‐ Binary FM

Mortgage -‐ -‐ Binary FM

Saving -‐ -‐ Binary FM

Investment -‐ -‐ Binary FM

Leasing -‐ V740_1: […] Lease Binary FM

Leasing 1 -‐ V740_2: […] Lease Binary FM

Leasing 2 -‐ V740_3: […] Lease Binary FM

Insurances -‐ Numberinsurances[…] Scale (0 -‐ 6) FM Mobile

Banking -‐ V40_1: Marktaan […] Binary FM

Negative Balance Facility

Negative Balance Negativesaldo[…] Binary FM

Products from Competitor 1 -‐ Numberproducts[…] Scale (0 – 10) FM Products from Competitor 2 -‐ Numberproducts[…] Scale (0 – 10) FM Services from Focal Bank Number of products

Products[…] Scale (0 – 5) IFM

Trust -‐ Trust[…]/

Vertrouwen[…] Scale (1 – 9) IFM Willingness

to

Recommend

WTR Recommend[…]/

Aanbevelen[…] Scale (0 – 10) IFM Brand

Preference -‐ Brandpreference[…] Binary IFM

Starter -‐ -‐ Binary FM/IFM

ZZP -‐ -‐ Binary FM/IFM

Figure 7 – Summary of variable characteristics

(32)

Methodology

After the careful preparation of the datasets, the method of analysis will be introduced and discussed. In order to test the hypotheses in the data, quantitative descriptive methods are needed. The methods chosen are a

multinomial logistic regression analysis for the Financial Monitor and a logistic regression analysis with moderated mediation effects for the Image-‐ Forming Monitor. Both methods and their proper application are examined and their use accounted for.

Multinomial Logistic Regression

The purpose of this study is to determine and characterize the antecedents of bank patronage sentiment. From the prior literature study it can be concluded that there is a multitude of drivers at work jointly influencing the outcomes. The aim of the analysis is to explore which of those drivers play a significant role in the B2B financial service sector. Therefore the statistical significance of the drivers as well as the magnitude of the influence on the outcomes is of interest. A statistical method that is able to shed light into that matter is a regression

analysis. In a regression analysis the joint influence of various driver variables on an outcomes variable can be measured and the significance of the effect of each driver variable determined. The goal is to find a model, which explains a significant amount of the forces that are driving an outcome. Although, in reality, the chance that a model can explain is very small, it is possible to construct models, which are able to supply enough certainty so that practical conclusions can be reached.

Since the dependent (outcome) variable of this model is of dichotomous nature, the deployment of a linear or linear probability model bares to many

Antecedents of continuance commitment in the B2B financial service sector