ESSENTIAL INSIGHTS FOR
ORGANIZATIONS THAT CONSIDER
TO BECOME CUSTOMER CENTRIC
revealed by means of a cross-classified multilevel model
at the country and industry level
Master thesis, MSc Marketing Management & Marketing Intelligence University of Groningen, Faculty of Economics and Business
January 20, 2016 Silke Mous Studentnummer: s2038269 Fagelstraat 16-3 1052 GB Amsterdam tel: +316-16162136 e-mail: mous.silke@gmail.com Supervisors University of Groningen
First supervisor: Prof. Dr. J. E. Wieringa Second supervisor: Dr. J. T. Bouma
Internship Company – Millward Brown Vermeer
ABSTRACT
The business world is changing and customers are demanding more and more. Customer centricity has gained considerable attention in recent years, arguing that this more recent philosophy better taps into this changing business environment. Many academics have argued that the previously dominating product centric philosophy has become obsolete. However, is customer centricity really a strategy that satisfies the needs of the current business world? Or would it be better to remain product centric in certain situations? This research takes a closer look at the differences between countries and industries in order to uncover the situations in which certain product or customer centric capabilities appear to be most effective. Three customer centric and three product centric capabilities have been tested to assess their influence on revenue growth. The customer centric capabilities that are included are taking the voice of
the customer into account in all major business decisions, a customer centric culture and NPD based on customer needs. The product centric capabilities consist of a strong focus on core products/services, transactional customer relationship and NPD based on technological innovation. This research uses a global dataset of 2557 respondents derived from 66 countries
and 17 industries. By means of a cross-classified random effects model; the study reveals that both, customer centric and product centric capabilities appear to have a positive influence on business growth. The capability of taking the voice of the customer into account in all major
business decisions, of NPD based on customer needs and of a strong focus on core products/services have a positive influence on revenue growth. These results contribute to the
customer and product centric literature, as this is the first research that confirms the positive effects on business growth. Furthermore, the results indicate that customer centric capabilities should not be seen as mutually exclusive with product centric capabilities. This is confirmed by the retail industry and the business & financial services industry specifically, which are capable of growing organizations by utilizing both product and customer centric capabilities simultaneously. Additionally, at the country level, it can be seen that Asian countries are leading in creating product and customer centric capabilities that contribute to business growth. This research shows that customer centricity can lead to business growth, however, product centricity has not become obsolete and can result in business growth as well.
TABLE OF CONTENTS
1. INTRODUCTION ... 7
1.1 Product centricity ... 8
1.2 Customer Centricity ... 9
1.3 Influence of product centric and customer centric capabilities on growth ... 10 1.4 Industry differences ... 11 1.5 Country differences ... 11 1.6 Approach ... 12 2. THEORETICAL FRAMEWORK... 134 2.1 Business growth ... 14 2.2 Product centricity ... 14
Strong focus on core products/services ... 15
Transactional customer relationship ... 15
New product development based on technological innovations ... 16
2.3 Customer Centricity ... 16
Voice of the customer ... 18
Customer-oriented organizational culture ... 19
New product development born out of customer needs ... 20
2.4 Industry differences ... 21
2.5 International differences ... 22
3. METHODOLOGY ... 24
3.1 Data collection and sample characteristics ... 25
3.2 Variable selection ... 25
Revenue growth ... 26
Individual level variables ... 26
Industry level variable ... 27
Country level variables ... 27
Industry identifier ... 28 Country identifier ... 28 3.3 Data analysis ... 28 Modelling methodology ... 28 Estimation technique ... 29 Model comparison ... 29
3.4 Model building process ... 30
4. RESULTS ... 32
4.1 Data exploration and preparation ... 32
Correlation ... 32
Missing data ... 32
Outliers ... 33
4.2 Null model ... 33
4.3 Country and industry interaction classification ... 35
4.4 Individual level explanatory variables ... 36
4.5 Industry level explanatory variable ... 37
4.6 Country level explanatory variables ... 38
4.7 Random slopes of individual level variables ... 40
4.8 Predictions of random slopes ... 43
4.9 Industry and country level residuals ... 44
Industry level residuals ... 45
Country level residuals ... 47
4.10 Validation ... 49 Face validity ... 49 Statistical validity ... 49 Parameter reliability ... 49 Heteroscedasticity ... 49 Normality ... 50 4.11 Hypothesis testing ... 50 5. DISCUSSION... 51 5.1 Customer centricity ... 51 5.2 Product centricity ... 52 5.3 Industry differences ... 54 5.4 International differences ... 55 5.5 Managerial implications ... 56
5.6 Limitations and directions for future research ... 56
7. REFERENCES ... 58
1. INTRODUCTION
In the current fast changing and competitive business environment, organizations are struggling more than ever to create a competitive advantage and achieve growth (Baldwin 2006; Barney 1991; Malhotra et al. 2005). Long-term success and consistent growth are difficult to achieve but are required by investors (Mcgrath 2012). Organizations can choose to grow by tapping into new markets with existing products, diversifying by moving away from the present product line and market, modifying existing products to new purposes, or an organization can improve on what it is currently doing (Ansoff 1957). The latter is what this research is about; how can organizations improve and grow on their current activities? The nature of organizations and products are changing, and this is forcing organizations to rethink their current strategy in order to achieve growth (Porter and Heppelmann 2014).
How will organizations drive growth? And with which capabilities do firms secure a competitive advantage? According to Treacy & Wiersema (1993), organizations should clearly define their intrinsic nature and choose a value discipline in which they want to excel. Different value disciplines require different operating processes and capabilities (Treacy and Wiersema 1993). Firms can either choose to sell the best product, find the best customer solution or operate with the most competitive total cost (Treacy and Wiersema 1993). A master in one of these value disciplines should be the goal for organizations aiming for business growth. It is of vital importance for long term business survival to create a perfect understanding of which capabilities need to be in place to excel in one of these value disciplines and improve an organization’s performance.
product centricity. Is it always beneficial to become a customer centric organization? It can be argued that a customer centric strategy may not be the best strategy for every firm. Since a product centric strategy allows firms to invest less heavily in customer relationships and be primarily organized around the products, it is believed that remaining organized around products may bring business growth to organizations as well. Therefore, a closer look will be taken at the situations in which it is especially worth to become customer centric, and in which situations it is better to remain product centric. This research aims to quantify the influence of customer centric and product centric capabilities on business growth, and creates a practically useful and academically relevant model which identifies differences between industries and countries.
1.1 Product centricity
Since decades, organizations have primarily focussed on the core products or services they were selling (Fader, 2012). This product centric management philosophy has been prevailing for a long time, however, empirical evidence on the influence of the product concept on business growth remains heavily under investigated.
According to Shah et al. (2006), product centricity finds itself at one side of the customer-product centric continuum. In the current business world, many organizations are considered to be product centric Fader (2012). The product centric approach is primarily transaction oriented instead of relationship oriented (Shah et al. 2006). Product features and advantages are of major importance and the basic philosophy is to sell products to whoever likes to buy it (Mallick, Ritzman, and Sinha 2013; Shah et al. 2006). Fader (2012) argues that the strategic advantage of product centric firms resides in the product expertise, strong product portfolios and a high valuation of the brand compared to the customer. The product centric approach has its roots in the early days of marketing, and has been proven to be very successful in the quick commercialization of new ideas (Mallick, Ritzman, and Sinha 2013). A considerable number of researchers have argued that the product centric era has ended (Fader 2012; Shah et al. 2006). There are reasons to believe this argument, however, there are also convincing cases in which product centricity has proven to be successful in this day and age.
fast changing business environment. These insights necessitate having a closer look at the customer centric approach and possibly require adding a nuance to when customer centricity is beneficial. Reformulation of the currently dominating management philosophy might be necessary.
1.2 Customer Centricity
Since the 90’s of the last century, the focus has moved from the product to the customer (Lamberti 2013; Lemon and Verhoef 2015; Shah et al. 2006). The customer centric management philosophy puts the customer at the centre of the business and highly values the customer concept (Hoekstra, Leeflang, and Wittink 1999). The importance of customer centricity was already emphasized by Drucker and is currently dominating in the marketing literature (1954). In his book The practice of management he argued that “It is the customer who determines what a business is, what it produces, and whether it will prosper”. His early idea that businesses should be looked at through the eyes of the customer has resulted in the adoption of customer centric practices and strategies in private as well as in public sectors (Gummesson 2008).
An increasing customer diversity, highly fragmented markets, wide availability of data and an intensified global competition are all factors that have contributed to this shift towards customer centricity (Ramani and Kumar 2008; Senn, Thoma, and Yip 2013; Shah et al. 2006). The twenty-first century has come with an increasing number of opportunities for customer centric practices (Sheth, Sisodia, and Sharma 2000). Furthermore, customers are increasingly demanding more from firms, and many firms regard this as being a business opportunity (Ramani and Kumar 2008; Sheth, Sisodia, and Sharma 2000). The customer-firm interaction has shifted from a firm-to-customer (one-way) communication towards customer- to-firm (two-way) interaction and value creation (Hoekstra, Leeflang, and Wittink 1999). The customer concept believes that the individual organization should take the individual customer as a starting point and place it at the centre of the business in order to realize superior performance (Hoekstra, Leeflang, and Wittink 1999; Ramani and Kumar 2008).
Verhoef 2015; Shah et al. 2006). In marketing, innovation management, strategic management and IT management research, customer centricity has gained an important place (Hoekstra, Leeflang, and Wittink 1999; Kapoko-Tagne 2013; Senn, Thoma, and Yip 2013; Shah et al. 2006; Verhoef et al. 2009). The majority of customer centricity research has put the emphasis on the antecedents and elements of customer centricity (Hoekstra, Leeflang, and Wittink 1999; Lamberti 2013; Ramani and Kumar 2008; Sheth, Sisodia, and Sharma 2000). Lamberti (2013) has characterized customer centricity as a construct existing of four elements: interactive customer relationship management, customer integration, internal integration and external integration. Furthermore, Shah et al. (2006) created a roadmap for businesses to move from being product centric to becoming customer centric. Besides the academic world, firms have also come to realize that customer centricity has become strategically important for long term survival (Shah et al. 2006). Companies like IBM, Amazon and Best Buy have given customer centricity high priority and gained competitive advantage as a result. Recent developments in research suggest that customer centric strategies have the largest potential for growth (Fader, 2012), however, an empirical study on the actual effects on business growth remain scarce and remains a surprisingly under investigated area. This research provides evidence for a critical but untested assumption of the customer centricity literature - that customer centricity leads to business growth.
1.3 Influence of product centric and customer centric capabilities on business growth
1.4 Industry differences
On average, the influence of product and customer centric capabilities might be valuable, but it is even more insightful to look at the situational factors in which one or the combination of both sets of capabilities may lead to even greater success. Shah et al. (2006) already suggested that the benefits of customer centricity might vary between industries. This research is an answer to their suggestion for future research. Many companies are facing major difficulties in fully utilizing the customer centric concept, could the relative value of customer centricity be larger in one industry compared to another? Consequently, it seems logical that becoming customer centric is not the appropriate growth strategy for every business, despite the increasing availability of individual-level data and researchers promoting customer centricity. Apparently, differences between firms in the influence of customer centricity on business growth exist, which means that customer centricity should not be seen as a one-size-fits-all approach.
The reason, and a probable explanation for these differences, is based on the resource-based view framework. The resource-based view suggests that organizations’ resources should be valuable, rare, imperfectly imitable and not substitutable in order to have the potential of creating sustained competitive advantage (Barney 1991). Firm resources can only be valuable if their strategies exploit opportunities or neutralize threats (Barney 1991). This indicates that certain resources are valued differently by organizations in one industry compared to organizations in another industry due to the differences in opportunities and threats faced by them. This heterogeneity should be taken into account, hence a thorough investigation will be done on the moderating effects of different industries.
1.5 Country differences
Which organizations benefit more from customer centric capabilities than product centric capabilities? Are organizations in one industry less suitable for customer centric capabilities than other industries? Or are customer centric capabilities in an individualistic country less successful than customer centric capabilities in collectivistic countries? Answers to these questions are prerequisites for a sustainable competitive international strategy but have been largely under investigated, which makes it of major importance to do research on. The aim of this research is to provide academics and practitioners with a solid understanding of situational factors in which it is more beneficial to adopt customer centric or product centric capabilities in order to grow the organization. The central question of this research is: “To what extent do
customer centric and product centric capabilities influence business growth, and what are the differences across industries and countries?”
1.6 Approach
2. THEORETICAL FRAMEWORK
The aim of this study is to provide insights into two specific strategies for growth. We investigate the influence of product centric and customer centric capabilities on business growth and take a closer look at how differences between countries and industries affect this. In this chapter we discuss different product and customer centric capabilities that are included in this study. To start with, customer centric capabilities are discussed and customer centricity hypotheses are constructed. Following that, product centric capabilities are discussed followed by a description of the related hypotheses.
2.1 Business growth
The primary objective of every company in every industry and every country is to make a profit and maximize this profit as long as possible (Fader 2012). Steady and predictable business growth is an important goal for organizations and its shareholders. Several academics have provided frameworks that can be used in making strategic decisions aimed at supporting growth objectives (Richardson 2008). For instance, Ansoff (1957) provided a renowned framework distinguishing four growth strategies that combine existing and new markets with existing and new products and services. Business growth can not only be achieved in different ways, but also be measured in different ways. Several indicators and formulas have been used in empirical growth studies with revenue as a very common measure in academic literature (Achtenhagen, Naldi, and Melin 2010). As an illustration, an extensive review of 35 articles that have been published in nine leading strategy and organization journals found that 83% of the studies used revenues as a concept for growth (Achtenhagen, Naldi, and Melin 2010). Despite the usefulness of looking at revenue growth for one specific firm, it is more valuable to look at a firm’s performance compared to competition. Paying attention to reference points in a firm’s environment, such as competitors, influence organizational learning and strategic decision making (Blettner et al. 2013). The combination of revenue growth with competition as a reference point provides academics and practitioners with valuable insights in relative performance. Therefore, current revenue growth compared to competition will be used as a central concept in this research.
2.2 Product centricity
the opportunity to expand by tapping into new markets or by adjusting existing products and convince customers that the product has improved significantly. Operational excellence, as Treacy & Wiersema (1993) have defined it, brings markets with a combination of quality, price, and ease of purchase. According to Fader (2012), Apple is the perfect example of a typically product centric firm which is evidently successful in its approach. Millions of iPhones, MacBooks, iPods and iPads are sold across the world with increasing revenues year by year (Fader 2012). In this line of reasoning, it is expected that product centric capabilities are positively related to business growth. Data on three product centric capabilities were available from the Insights2020 questionnaire, and are included in this research based on their theoretical relevance. This theoretical relevance is discussed in the following sections.
Strong focus on core products/services
The basic philosophy behind product centricity is to sell products to whoever wants to buy them (Shah et al. 2006). Products can be sold by emphasizing product features and advantages in their differential positioning towards the customer (Sharp and Dawes 2001). However, besides differentiation, increasing the distinctiveness of the products/services might pose additional benefits (Sharp 2010). Distinctiveness goes beyond differentiation and allows firms to charge higher prices without heavily investing in features. A strong focus on core products/services might lead to more customer utility and better utilisation of firm resources (Sharp and Dawes 2001). A strong focus on the core products/services does not stop at the benefits of differentiation or distinctiveness but can also be seen across the entire organization. By having performance metrics in place that measure the number of new products, profitability per product and market share by product, firms may have cost-efficiency benefits (Shah et al. 2006). Product profit centres, product managers and product sales teams take a central role in the organizational structure in order to effectively manage the product portfolio (Mallick, Ritzman, and Sinha 2013). It is expected that firms can be significantly successful by creating a strong focus on products/services due to differentiation, distinctiveness and efficiency benefits. Therefore, we hypothesize:
Hypothesis 1a: The level of focus on the core product/service has a positive relationship with business growth.
Transactional customer relationship
relationship is defined as “a buying-selling agreement where participants conduct business for a specific time period according to terms generally outlined in a standard contract” (Whipple, Lynch, and Nyaga 2010, p. 507). According to Li & Nicholls (2000), a transactional relationship has a short-term orientation, with clear beginnings and endings, does not consider mutual interests, with a low level of relational involvement. Due to this low level of relational involvement, firms do not have to invest in gaining an understanding of the customer, customized service and long term commitment (Whipple, Lynch, and Nyaga 2010). Reasoning from a cost saving perspective, a transactional relationship is likely to result in business growth. Therefore, the following hypothesis can be formulated:
Hypothesis 1b: The level of transactional customer relationships has a positive relationship with business growth.
New product development based on technological innovations
For product centric firms, technological innovations are very important in new product development (Shah et al. 2006). These technological innovations can take place either at the product or at the process side (Hong, Kim, and Cin 2015). Product technological innovations can be defined as “new product developments involving major advances in product technology” (Calantone, Vickery, and Dröge 1995, p. 216). These are internalized into a product or service and are likely to create new deliverables and value for the customer. Product technological innovations allow firms to fight competition in the market by product differentiation (Hong, Kim, and Cin 2015). Process technological innovations can be defined as “Innovations that relate to how to operate a business” (Hong, Kim, and Cin 2015, p. 877). These innovations are typically in the field of efficiency and allow businesses to create products at lower costs. Product and process technological innovations are highly related when executed in the right way (Calantone, Vickery, and Dröge 1995). Lower costs and product improvements are likely to be valued more by customers and are therefore regarded as having a positive influence on business growth. As a result, we state the following hypothesis:
Hypothesis 1c: The level of new product development based technological innovation has a positive relationship with business growth.
2.3 Customer Centricity
power to make as much money from them as possible and to find more customers like them” (Fader 2012, p. 86). Firms are customer centric when they are able to discover customer needs and develop solutions that satisfy these needs (Lamberti 2013). All decisions should start with the customer and businesses should be mainly relationship-oriented in order to be customer centric. In order to operationalize customer centricity “A structure organised around customer needs (one unit fulfils the needs of a unique customer segment) rather than around functions and products” (Lee et al. 2015, p. 251). The organization should be externally oriented and make employees customer advocates (Shah et al. 2006). Although this literature addresses several associated facets, in essence, every definition includes that the customer is placed at the centre of the business. Therefore, the following definition of customer centricity will be taken as a starting point for this research: “when an organization puts the customer at the heart of
everything it does” (Galbraith 2005). This definition gives the customer a central place and
indirectly implies that a customer centric business organizes everything around this customer.
Voice of the customer
In order to identify customer needs, it is crucial to listen to the voice of the customer (Griffin and Hauser 1993; Lam and Mayer 2013). The importance of this has been recognized by many firms (Goffin et al. 2012). Taking the voice of the customer into account in all key business decisions refers to the market/customer orientation literature (Blocker et al. 2011). Customer orientation can be regarded as “the organization-wide generation of market intelligence pertaining to current and future customer needs” (Kohli and Jaworski 1990, p. 6). In order to uncover these current and future customer needs, it is crucial to listen to the voice of the customer (Blocker et al. 2011). In order to operationalize the voice of the customer, the entire organization should be aware of the importance of the customer’s voice and business activities and decisions should be adjusted accordingly (Goffin et al. 2012; Griffin and Hauser 1993; Zondag and Ferrin 2014). It is of high importance that the voice of the customer is fully integrated in all firm activities (Sheth, Sisodia, and Sharma 2000). For example in the case of the supply chain, in order to make supply chain management customer centric, customer demands should be integrated with supply data as the basis for strategic and operational decision making (Zondag and Ferrin 2014). Considering the voice of the customer across in all business decisions has been regarded as a crucial factor in the creation of successful customer centric firms for several reasons (Hoekstra, Leeflang, and Wittink 1999; Ramani and Kumar 2008).
First, it is likely that customers will experience benefits due to the organizational alignment in the form of customer satisfaction (Galbraith 2005). The current generation of customers is expecting more and more from firms in terms of personalization (Baldwin 2006; van Doorn and Hoekstra 2013). It becomes harder for firms to distinguish from the vast amount of suppliers in many industries, and firms are looking for innovative solutions to this problem (Mcgrath 2012). When listening to the voice of the customer, it is possible for firms to, for example, unveil the type of relationship the customer is looking for with the organization (Avery, Fournier, and Wittenbraker 2014). By taking this into account in all key business decisions, firms are likely to establish relationships that are highly valued by their customers -resulting in effective customer relationship management (Reinartz, Krafft, and Hoyer 2004). This deep understanding of the customer contributes to a credible customer relationship and allows fulfilling the promise of customer centricity.
may feel that they contribute to the firm in the form of knowledge and experiences which increases their sense of ownership with the firm (Jaakkola and Alexander 2014). A sense of ownership is the extent to which customers feel accountable to a firm and actively involve in the organization (Ramani and Kumar 2008). This gives the organization the opportunity to stay close to their main source of business. Furthermore, when the customers start to have positive cognitive associations with the firm due to the fact that a firm takes their needs into account, customers might engage in influencing behaviour (Jaakkola and Alexander 2014). This influencing behaviour happens when customers use knowledge or experiences of a firm to affect the perceptions of other people’s preferences and knowledge about the specific firm (Kumar et al. 2010). This might be an important factor contributing to business growth. Therefore the following hypothesis is formulated:
Hypothesis 2a: The level of taking the voice of the customer into account in all key business decisions has a positive influence on business growth.
Customer-oriented organizational culture
An organizational culture is defined as: “The pattern of shared values and beliefs that help individuals understand organizational functioning and thus provide them norms for behaviour in the organization” (Chakravorti 2011, p. 144). An organization that is aiming to be customer centric is regarded to have a very distinct organizational culture with an agreed-upon set of shared values which puts the customer first across the entire organization (Shah et al. 2006). Everything and everyone in the organization is organized around the customer, with business objectives stated in terms of the customer (Galbraith 2005; Hoekstra, Leeflang, and Wittink 1999). A customer oriented organizational culture is a culture in which the customers’ interests are the main concern and the firm sets priorities around their customers (Galbraith 2005). Firms should be constantly trying to learn from their customers and every employee is expected to participate in thinking about what is right for the organization and its customers in order to create a customer-oriented organizational culture (Hoekstra, Leeflang, and Wittink 1999). Consequently, managing the customer’s interest becomes a cross-functional activity (Hoekstra, Leeflang, and Wittink 1999; Peltier, Zahay, and Lehmann 2013). A shared vision on how customer information is managed contributes to internal alignment on how to deal with customer information (Peltier, Zahay, and Lehmann 2013).
regarded to be beneficial for firms, however, the real value and financial consequences for firms with customer-oriented organizational cultures remains unclear until now.
Notwithstanding, in the field of organizational learning and CRM practices, Peltier et al. (2013) showed a link between a customer-oriented organizational culture and performance through improved customer data quality and a stimulation of intra-organizational learning. Furthermore, through a strong focus on the customer and their value, it is easier for firms to determine the potential value for the firm in the future (Kumar et al. 2010). A customer-oriented organizational culture is likely to discover heterogeneity in value and preferences among customers which allow a firm to set priorities in which customers to serve (Reinartz, Krafft, and Hoyer 2004). Furthermore, firms can create a competitive advantage by gaining an edge over their competitors in retaining customers and in positioning themselves to take advantage of environmental changes (Porter and Heppelmann 2014; Porter and Millar 1985). Based on this, careful management and monitoring of customers has become more important in sustaining a competitive advantage (Bellou 2007).
Furthermore, a study by Lee, Yoon, Kim & Kang (2006) shows that a market-oriented culture affects firm performance by emphasizing the importance of the customer across the entire organization. This takes the responsibility for the customer beyond the marketing department and emphasizes the importance for the entire organization (Lemon and Verhoef 2015). In line with this reasoning, it is expected that a customer-oriented organizational culture contributes to business growth. Therefore, the following hypothesis is formulated:
Hypothesis 2b: The level of being a customer oriented organizational culture has a positive effect on business growth.
New product development born out of customer needs
New product development activities linked to business performance have been researched a lot in the academic world (Calantone, Vickery, and Dröge 1995; Crosby and Johnson 2006; Weber, Weggeman, and Van Aken 2012; Yoo et al. 2012). A high level of customization and product flexibility allows firms to meet special customer needs and positively influence business performance (Calantone, Vickery, and Dröge 1995). In the new product development process, innovation and growth cannot be successfully realized without investing in customer R&D (Arussy 2006). Therefore, it is expected that new product development based on an understanding of customer needs will result in business growth. The following hypothesis is formulated:
Hypothesis 2c: The level of new products or services born out of understanding customer needs has a positive influence on business growth.
2.4 Industry differences
capabilities that focus more on the emotional, social and relational benefits of the brand/product, it is assumed that firms in the Fast Moving Consumer Goods industry are more likely to see revenue growth. On the contrary, firms in B2B industries are likely to value transactional relationships. This is what makes it highly probable that product centric capabilities seems to be more appropriate for growth in this specific industry context. Due to these different perceptions of value in B2B and B2C contexts, it can be expected that this plays an important role in the influence of customer and product centric capabilities on revenue growth between industries. Therefore, we state the following hypothesis:
Hypothesis 3: The effect of product centric and customer centric capabilities on business growth differs between industries.
2.5 International differences
value caring for others. The Uncertainty Avoidance dimension can be defined as “The extent to which people feel threatened by uncertainty and ambiguity and try to avoid these situations” (Hofstede 1991, p. 113). This dimension includes the need for well-defined rules and regulations that stipulate the actions to be taken.
The cultural dimensions of Hofstede have been linked to international relationship marketing in a study by Samaha et al. (2014), and they concluded that the effectiveness of relationship marketing differs greatly between countries. They conclude that this high degree of heterogeneity across countries requires a fine-grained perspective on a firm’s international marketing strategy. These differences between cultural dimensions are also expected to influence business growth in the field of customer and product centric capabilities since relationships are an important component of both the customer and product centric approach (Blocker et al. 2011; Hoekstra, Leeflang, and Wittink 1999; Senn, Thoma, and Yip 2013; Shah et al. 2006). The former aims to establish relationships and the latter is solely transaction oriented.
Despite the fact that Hofstede has used his dimensions mainly for human resource management, it is becoming increasingly popular in marketing and business research (Soares, Farhangmehr, and Shoham 2007). The dimensions can be utilized to infer cultural characteristics (Hofstede, Hofstede, and Minkov 2010; Orr and Hauser 2008; Soares, Farhangmehr, and Shoham 2007). Since a country’s culture can be regarded as a major force which influences people’s perceptions and behaviours, it is likely that both approaches are valued differently across cultures and countries as a result (Samaha, Beck, and Palmatier 2014). Therefore, we hypothesize:
Hypothesis 4: The effect of product centric and customer centric capabilities on business growth differs between countries.
3. METHODOLOGY
This chapter describes the model building methodology that we used to derive the results. Several times will be referred to the Appendix for further details on in-depth technical elements or visualizations. Section 3.1 discusses the data collection procedure and the characteristics of the sample. Section 3.2 considers the variables that have been selected to test the hypotheses with. Section 3.3 discusses data preparation and final section 3.4 shortly explains the data analyses.
3.1 Data collection and sample characteristics
Data is collected through a global marketing leadership initiative of Millward Brown Vermeer focussing on driving customer centric growth. This research initiative is recognized as Insights 2020 and is based on a quantitative survey including 2668 business participants from across the entire organization in 17 industries coming from 88 countries. Furthermore, 350 vision interviews with business, marketing and insights & analytics leaders were held. Millward Brown Vermeer initiated Insights 2020 in cooperation with ESOMAR, the Advertising Research Foundation, LinkedIn, Korn Ferry and Kantar. Variables used in this research are a subset of the total questionnaire containing numerous questions on customer centricity.
The original database consisted of respondents that are part of agencies that advise other businesses as well. Since respondents from agencies are likely to answer questions related to their specific company differently because of their advising role towards other companies, they should not be compared with non-agency respondents. Therefore, the agency respondents (Media (Strategy & Buying), Advertising (Strategy/Creation), Consulting, Market Research and Big data/Analytics) have been removed from the dataset.
3.2 Variable selection
In order to test the hypotheses of this research, it is of major importance to choose the right measures. In the following section, the variables that were selected for testing the hypotheses are discussed. The questions that are selected from the Insights 2020 questionnaire can be found in Appendix 5. A summary of these variables can be found in table 1.
Revenue growth
Insights 2020 included four variables on an organization’s performance versus competition that were asked to respondents. These include Revenue Growth, Marketing Return on Investment, Brand Equity and Profitability. Respondents had to rank themselves on how they perform on these variables compared to competition (1 = much worse than competitors to 7 = much better than competitors). For this research, Revenue Growth compared to competition is used because this can be regarded as an indication for an organization’s current and future performance compared to competition.
Individual level variables
TABLE 1
Descriptive statistics key variables
Variable name Description N Minimum Maximum Mean Std.
Deviation
RevenueGrowth Revenue growth compared to competition
2668 1 7 4.73 1.31
PC_ProductFocus Strong focus on the core product/service
2668 1 5 3.97 .73
PC_TransRel Dummy variable indicating transactional relationship
2668 0 1 .11 .31
PC_NPD_TechInn Degree of new products or services based on technological innovation
2668 1 7 3.58 1.31
CC_VoiceOfCustomer Degree of taking the voice of the customer into account in all key business decisions
2668 1 5 3.29 .99
CC_Culture Degree to which customer centricity is truly embraced by all functions in the entire organization
2668 1 5 3.06 1.02
CC_NPD_CustNeeds Degree of new products or services based on new understanding of customer’s needs
2668 0 100 65.90 21.11
B2C Degree of B2C 2668 1 5 2.36 1.31
PDI Score on Power
Distance dimension
2557 11 100 52.43 19.48
IND Score on Individualism
dimension
2557 6 91 64.11 27.05
MAS Score on Masculinity
dimension
2557 5 100 54.48 18.05
UAI Score on Uncertainty
Avoidance dimension
2557 8 100 53.78 19.71
Industry Industry identifier 2668 1 17 NA NA
Country Country identifier 2668 1 88 NA NA
Valid N 2557
Industry level variable
The industry variable measuring the degree of direct contact with the consumer (B2C) is measured on the individual level in the Insights 2020 questionnaire but included with an average value per industry. The variable is measured on a five point scale with value 1 indicating an industry with a number of business to business companies, and value 5 indicating a high number of business to consumer companies (1=B2B, 5=B2C).
Country level variables
on Hofstede’s website. The data matrix used for this research has been updated in August 2015. The national cultural dimensions are measured on a scale from roughly 0 to 100 with 50 as the mid-level (Hofstede, Hofstede, and Minkov 2010). This mid-level indicates whether a country scores low or high on a particular scale.
Industry identifier
A variable identifying the 17 industries in which respondents completed the questionnaire is used. The industries range from a wide variety of industries e.g. Technology, Fast Mover Consumer Goods and Retail. A complete list of the industries that are included in this research, can be found in Appendix 4.
Country identifier
Data is collected from 88 countries across the entire world. Ranging from Uganda to Panama and from South Korea to Germany. The exact countries that are included in this research can be found in Appendix 2 and the distribution of the countries across regions can be found in Appendix 3 showing that 23,6% of the sample is derived from North America, 11,8% from Middle and South America, 35,7% from Europe, 3,3% from the Middle East, 3,3 from Oceania and 6,1% from Africa. The country identifier variable identifies the country in which the respondent’s company is based.
3.3 Data analysis
Modelling methodology
Choosing an appropriate modelling methodology is of vital importance for the reliability and validity of the final model. The data in this research consists of individuals that can be grouped into higher-level units, being countries and industries. Due to this complex data structure, multilevel modelling is a suitable method to explore and analyse the data (Leckie and Bell 2013).
structure requires a statistical method suitable for hierarchical data and makes a hierarchical multilevel model a convenient method.
Furthermore, the individual respondents belong to combinations of countries and industries; they are nested within countries and are separately nested within industries. Industries and countries are not typically nested within one another, as neither all individuals in the same industry live in the same country, nor do all individuals from the same country operate in the same industry. In this case, countries and industries are crossed with each other with the possibility that every individual belongs to a combination of industry and country. The model in this research will take into account that the individuals are nested within the cells of a two-way cross-classification of industries and countries.
Estimation technique
Estimating the cross-classified multilevel model is done in the program MLwiN. This program uses the Markov Chain Monte Carlo estimation technique. A method that is used for calculating numerical approximations for Bayesian statistics and makes it possible to compute large multilevel models that require integrations of large numbers of parameters (Spiegelhalter et al. 2002).
Model comparison
Model comparison is based on a measure of fit (Spiegelhalter et al. 2002). In order to compare models that are estimated by MCMC, the Deviance Information Criterion (DIC) is used (Leckie and Bell 2013). The DIC is based on a trade-off between the fit of the model and the data and the corresponding complexity of the model. This is based on Akaike’s Information Criterion (AIC) but has the additional benefit of comparing non-nested models. Comparable to AIC, the DIC punishes the goodness of fit with the complexity of the model (the number of parameters). Model comparison is based on the differences in DIC between models; the absolute size of DIC is regarded to be unimportant. A decrease of the DIC value indicates an increase in model fit.
Intra-class correlation coefficients (ICCs)
intra-class correlation coefficient calculation. The ICC calculation for the highest level in a three level model looks as follows:
𝑐𝑜𝑟𝑟(𝑅𝐺𝑖, 𝑅𝐺𝑖′) =
𝜎𝑢(3)2 𝜎𝑢(3)2 + 𝜎
𝑢(2)2 + 𝜎𝑒2
where 𝜎𝑢(3)2 is the variance of the country level residuals, 𝜎𝑢(2)2 is the variance of the industry level residuals and 𝜎𝑒2 is the variance of the individuals level residuals.
3.4 Model building process
TABLE 2 Model building steps STEP
1. Data exploration end preparation Correlations
Missing data Outliers
2. Null-model
Two-way cross classified model Single level model
Comparison
Intra-class correlation
3. Country and industry interaction classification 4. Individual level explanatory variables
5. Industry level explanatory variables 6. Country level explanatory variables
4. RESULTS
Before starting the analysis, it is of major importance to get a better understanding of the key variables of this research.
4.1 Data exploration and preparation
Correlation
First, a correlation matrix is created. The correlation matrix in Appendix 7 shows the correlations between the variables at the individual level and every combination appears to be significantly correlated. However, when investigating the values of the correlations, they appear to be relatively low. According to Vaus (2002), a correlation above 0.70 is considered to be very strong and above 0.90 nearly perfect. The values in the correlation table of Appendix 7 shows no values above the value 0.70 which indicates that the correlations are within an acceptable range. A probable explanation for the significant but relatively low correlations is the large sample size of this research. The significance of these low correlations can therefore be omitted and pose no problems for this research (Vaus 2002).
Missing data
In order to perform statistically valid analyses, it is important to closely look whether missing data exist and investigate the nature of these missing data (Leckie and Bell 2013). The country identifier missed a value for one individual. This individual is removed from the dataset because it is not possible to run the analyses on individuals that cannot be assigned to a specific country. The individual and industry level variables show no missing data. However, on the country level, every Hofstede dimensions shows 111 missing values. This missing data is not missing at random because the missing values exist due to the fact that Hofstede has not yet created scores on the national culture dimensions for several countries. This regards the following 28 countries: Afghanistan, Algeria, Angola, Anguilla, Bolivia, Cambodia, Cyprus, Dominican Republic, Egypt, Ghana, Honduras, Jordan, Kenya, Lebanon, Lithuania, Macau, Mozambique, Namibia, Nigeria, Oman, Puerto Rico, Qatar, Saudi Arabia, Senegal, Sri Lanka, Sudan, Ukraine and the United Arab Emirates.
Outliers
As can be seen from the boxplots of revenue growth across industries and countries in Appendix 8, several outliers can be identified. However, the outliers are all within the range of the variable and are therefore considered to have realistic and valuable values. Since the variance within countries and industries are of major interest for this research, the outliers will be retained in the dataset.
4.2 Null model
The analysis starts with a two-way cross-classified variance components (model 1) for revenue growth. At this stage, no explanatory variables have been added to the model, which results in a null two-way cross-classified model. This model only includes an intercept, separate random effects for each of the two crossed-classifications, and an observation-level residual error term. The total revenue growth variation will be decomposed into individual variance components for the crossed-classifications and the observation-level. This provides a description of the variance components related to the country and industry level. The null model, expressed in a classification notation is composed as follows:
𝑅𝐺𝑖= 𝛽0+ 𝑢𝑐𝑜𝑢𝑛𝑡𝑟𝑦(𝑖) (3) + 𝑢𝑖𝑛𝑑𝑢𝑠𝑡𝑟𝑦(𝑖)(2) + 𝑒𝑖 𝑢𝑐𝑜𝑢𝑛𝑡𝑟𝑦(𝑖)(3) ~𝑁(0, 𝜎𝑢(3)2 ) 𝑢𝑖𝑛𝑑𝑢𝑠𝑡𝑟𝑦(𝑖)(2) ~𝑁(0, 𝜎𝑢(2)2 ) 𝑒𝑖~𝑁(0, 𝜎𝑒2)
where 𝑅𝐺𝑖 is the revenue growth compared to competition of individual i, 𝛽0 is the mean score across all individuals, 𝑢𝑐𝑜𝑢𝑛𝑡𝑟𝑦(𝑖)(3) is the effect of individual i’s industry, 𝑢𝑖𝑛𝑑𝑢𝑠𝑡𝑟𝑦(𝑖)(2) is the effect of individual i’s country, and 𝑒𝑖 is the residual error term.
In order to examine whether the multilevel model fits the data better than a simpler single-level model, the null hypothesis that there are no higher level effects at all should be tested (Leckie and Bell 2013). By means of the Likelihood ratio (LR) test statistic, the null joint hypothesis can be examined:
LR = (-2log𝐿0) - (-2log𝐿1)
where 𝐿0 is the likelihood value for the single-level model and 𝐿1 is the likelihood value for the cross-classified model.
= 9015). The cross-classified model (Model 1) is statistically better performing than the single-level model (Model 0). This indicates that the respondents do not act independently and grouping at the higher level seems appropriate.
The coefficient in the fixed part of the model is the intercept that can be interpreted as the average revenue growth of the total number of respondents (Average Revenue Growth = 4.7) and is significantly different from zero (p=0.000).
TABLE 3 Null-model comparison
Model 0 Model 1
Estimate S.E. Estimate S.E.
Fixed Part Constant 4.733*** 0.025 4.700*** 0.081 Random Part Level: Country cons/cons 0.122 0.046 Level: Industry cons/cons 0.040 0.024 Level: Respondent cons/cons 1.716 0.047 1.628 0.045 -2*loglikelihood: DIC: 9.015.104 8.915.145 pD: 2.003 44.484 Units: Country 88 Units: Industry 17 Units: Respondent 2668 2668 Units: constant 1 *p <.10, **p <.05, ***p <.01
Intra-class correlation coefficients (ICCs)
The variance components of the previously estimated model can be interpreted by means of the intra-class correlation coefficient (ICC). The ICC calculates the implied correlation of the individuals within the different levels (Leckie and Bell 2013). The calculations can be found in Appendix 9.
industries have a value of .400. The industry ICC is .209 and shows the correlation between individuals from the same industry but located in a different country. Furthermore, the ICC between individuals who are located in the same country and who are active in the same industry appears to be .609, explaining nearly 61% of the response variation. This indicates that the largest homogeneity is seen for individuals who both are located in the same country and work in the same industry. On the contrary, the relatively low value for the industry ICC shows that the values between industries are less similar than at the country level.
TABLE 4
Intra-class correlation coefficients
Level ICC
Country 0.400
Industry 0.209
Individuals nested in common country and industry combination 0.609
4.3 Country and industry interaction classification
In the previous model, industry and country effects were assumed to be additive. However, it is possible that an interaction effect resides between the variance of both levels. A new variable was created which represents the unique effect of an industry and country combination.
This random interaction effect between industries and countries was added to the two-way cross-classified variance components model, being a fourth classification. The model has been extended as follows: 𝑅𝐺𝑖= 𝛽0+ 𝑢𝑐𝑜𝑚𝑏𝑖𝑛𝑎𝑡𝑖𝑜𝑛(𝑖)(4) + 𝑢𝑐𝑜𝑢𝑛𝑡𝑟𝑦(𝑖)(3) + 𝑢𝑖𝑛𝑑𝑢𝑠𝑡𝑟𝑦(𝑖)(2) + 𝑒𝑖 𝑢𝑐𝑜𝑚𝑏𝑖𝑛𝑎𝑡𝑖𝑜𝑛(𝑖)(4) ~𝑁(0, 𝜎𝑢(4)2 ) 𝑢𝑐𝑜𝑢𝑛𝑡𝑟𝑦(𝑖)(3) ~𝑁(0, 𝜎𝑢(3)2 ) 𝑢𝑖𝑛𝑑𝑢𝑠𝑡𝑟𝑦(𝑖)(2) ~𝑁(0, 𝜎𝑢(2)2 ) 𝑒𝑖~𝑁(0, 𝜎𝑒2)
A comparison of the null-cross classified model (Model 1) with a new model including the random interaction (Model 2) can be seen in Appendix 10. The additional random effect reduces the DIC by 20 points, indicating that the non-additive model would be the preferred model based on model fit. However, the parameter simulations have difficulties in converging to stable parameters (see Appendix 11). This indicates that the model has difficulties with estimating the model. The database does not include sufficient information, which makes the model too difficult to estimate. A larger dataset is needed in order to make it possible to include this interaction classification. Even though it seems theoretically interesting to inspect the interaction classification, it is not possible with the data in this research.
4.4 Individual level explanatory variables
The six capabilities of the model have been added to the model with fixed parameters. By including these explanatory variables, the model attempts to account for the expectation that the capabilities substantially differ in their effect on revenue growth. The basic random-intercept model (Model 3) can be specified follows:
𝑅𝐺𝑖= 𝛽0+ 𝛽1 𝐶𝐶_𝑉𝑜𝑖𝑐𝑒𝐶𝑢𝑠𝑡𝑜𝑚𝑒𝑟𝑖+ 𝛽2 𝐶𝐶_𝑇𝑟𝑢𝑙𝑦𝐸𝑚𝑏𝑟𝑎𝑐𝑒𝑑𝑖+ 𝛽3 𝐶𝐶_𝑁𝑃𝐷𝐶𝑢𝑠𝑡𝑁𝑒𝑒𝑑𝑠𝑖 + 𝛽4 𝑃𝐶_𝑃𝑟𝑜𝑑𝑢𝑐𝑡𝐹𝑜𝑐𝑢𝑠𝑖+ 𝛽5 𝑃𝐶_𝑁𝐷𝑃𝑇𝑒𝑐ℎ𝐼𝑛𝑛𝑖+ 𝛽6 𝑃𝐶_𝑇𝑟𝑎𝑛𝑠𝑅𝑒𝑙𝑖 + 𝑢𝑐𝑜𝑢𝑛𝑡𝑟𝑦(𝑖)(3) + 𝑢𝑖𝑛𝑑𝑢𝑠𝑡𝑟𝑦(𝑖)(2) + 𝑒𝑖 𝑢𝑐𝑜𝑢𝑛𝑡𝑟𝑦(𝑖)(3) ~𝑁(0, 𝜎𝑢(3)2 ) 𝑢𝑖𝑛𝑑𝑢𝑠𝑡𝑟𝑦(𝑖)(2) ~𝑁(0, 𝜎𝑢(2)2 ) 𝑒𝑖~𝑁(0, 𝜎𝑒2)
significant but has a different effect than hypothesized (𝛽= -.157, p<.05). The DIC has dropped with 251.764, from 8915.14 to 8663.376, which is a considerable improvement in model fit.
The between-country variance is .083, the between-industry variance .021 and the individual respondent variance is 1.481. Compared to the estimates of model 2, the explanatory variables explain 31.97% at the country level, 47.5% at the industry level and 9.03% at the individual level. The explanatory variables appear to explain more variance at the industry and country level compared to the individual level.
4.5 Industry level explanatory variable
At the industry level, an explanatory variable is added. The basic random-intercept model (Model 4) can be specified as follows:
𝑅𝐺𝑖= 𝛽0+ 𝛽1 𝐶𝐶_𝑉𝑜𝑖𝑐𝑒𝐶𝑢𝑠𝑡𝑜𝑚𝑒𝑟𝑖+ 𝛽2 𝐶𝐶_𝑇𝑟𝑢𝑙𝑦𝐸𝑚𝑏𝑟𝑎𝑐𝑒𝑑𝑖+ 𝛽3 𝐶𝐶_𝑁𝑃𝐷𝐶𝑢𝑠𝑡𝑁𝑒𝑒𝑑𝑠𝑖 + 𝛽4 𝑃𝐶_𝑃𝑟𝑜𝑑𝑢𝑐𝑡𝐹𝑜𝑐𝑢𝑠𝑖+ 𝛽5 𝑃𝐶_𝑁𝐷𝑃𝑇𝑒𝑐ℎ𝐼𝑛𝑛𝑖+ 𝛽6 𝑃𝐶_𝑇𝑟𝑎𝑛𝑠𝑅𝑒𝑙𝑖 + 𝛽7 𝐵2𝐵_𝐵2𝐶𝑖+ 𝑢𝑐𝑜𝑢𝑛𝑡𝑟𝑦(𝑖) (3) + 𝑢𝑖𝑛𝑑𝑢𝑠𝑡𝑟𝑦(𝑖)(2) + 𝑒𝑖 𝑢𝑐𝑜𝑢𝑛𝑡𝑟𝑦(𝑖)(3) ~𝑁(0, 𝜎𝑢(3)2 ) 𝑢𝑖𝑛𝑑𝑢𝑠𝑡𝑟𝑦(𝑖)(2) ~𝑁(0, 𝜎𝑢(2)2 ) 𝑒𝑖~𝑁(0, 𝜎𝑒2)
The results of this model (Model 4) can be found in Appendix 13. The fixed part parameter for industry degree of B2C/B2B appears to be significant (p<.10) and negative. The between country variance is now slightly smaller than in the model without explanatory variable at the industry level. Even though the model did not improve (indicated by the increase of the DIC by 1), the addition of the B2C/B2B variable gives valuable and significant insights into the This indicates that every capability except for the product centric capability NPD based on
technological innovation has a positive significant influence on revenue growth.
Additionally, larger inequalities between the influences of the different capabilities exist on the industry level, rather than the country and individual level.
influence on revenue growth. Therefore, the industry level variable will be retained in the model.
4.6 Country level explanatory variables
At the country level, the four Hofstede dimensions have been added as explanatory variables. The model can be expressed mathematically (Model 5) as follows:
𝑅𝐺𝑖= 𝛽0+ 𝛽1 𝐶𝐶_𝑉𝑜𝑖𝑐𝑒𝐶𝑢𝑠𝑡𝑜𝑚𝑒𝑟𝑖+ 𝛽2 𝐶𝐶_𝑇𝑟𝑢𝑙𝑦𝐸𝑚𝑏𝑟𝑎𝑐𝑒𝑑𝑖+ 𝛽3 𝐶𝐶_𝑁𝑃𝐷𝐶𝑢𝑠𝑡𝑁𝑒𝑒𝑑𝑠𝑖 + 𝛽4 𝑃𝐶_𝑃𝑟𝑜𝑑𝑢𝑐𝑡𝐹𝑜𝑐𝑢𝑠𝑖+ 𝛽5 𝑃𝐶_𝑁𝐷𝑃𝑇𝑒𝑐ℎ𝐼𝑛𝑛𝑖+ 𝛽6 𝑃𝐶_𝑇𝑟𝑎𝑛𝑠𝑅𝑒𝑙𝑖 + 𝛽7 𝐵2𝐵_𝐵2𝐶𝑖+ 𝛽8 𝑃𝐷𝐼𝑖+ 𝛽9 𝐼𝑁𝐷𝑖+ 𝛽10 𝑀𝐴𝑆𝑖+ 𝛽11 𝑈𝐴𝑉𝑖+ 𝑢𝑐𝑜𝑢𝑛𝑡𝑟𝑦(𝑖) (3) + 𝑢𝑖𝑛𝑑𝑠𝑡𝑟𝑦(𝑖)(2) + 𝑒𝑖 𝑢𝑐𝑜𝑢𝑛𝑡𝑟𝑦(𝑖)(3) ~𝑁(0, 𝜎𝑢(3)2 𝑢𝑖𝑛𝑑𝑢𝑠𝑡𝑟𝑦(𝑖)(2) ~𝑁(0, 𝜎𝑢(2)2 𝑒𝑖~𝑁(0, 𝜎𝑒2
The results in Appendix 14 show that the only dimension that appears to be significant is masculinity (𝛽= .005, p<.05). As can be seen in the table 5, the sample size has decreased from 2668 to 2557. This is due to the 110 cases that have missing values on the Hofstede dimensions. MlwiN automatically omits these cases with missing values.
TABLE 5 Model comparison
*p <.10, **p <.05, ***p <.01
Model 0 Model 1 Model 2 Model 3 Model 4 Model 5
4.7 Random slopes of individual level variables
In order to check whether the influence of the individual level variables differs between countries and industries, the individual level variables have been given random coefficients. The previous models allowed the effects of the explanatory variables to be similar for each country and industry. This assumption will be adjusted by allowing the individual level variables to have varying intercepts and slopes across countries and industries. This has been done by including the individual variables with random intercepts and slopes and can be expressed mathematically as follows:
𝑅𝐺𝑖 ~𝑁(𝑋𝐵, Ω) 𝑅𝐺𝑖= 𝛽0𝑖+ 𝛽1𝑖 𝐶𝐶_𝑉𝑜𝑖𝑐𝑒𝐶𝑢𝑠𝑡𝑜𝑚𝑒𝑟𝑖+ 𝛽2𝑖 𝐶𝐶_𝑇𝑟𝑢𝑙𝑦𝐸𝑚𝑏𝑟𝑎𝑐𝑒𝑑𝑖+ 𝛽3𝑖 𝐶𝐶_𝑁𝑃𝐷𝐶𝑢𝑠𝑡𝑁𝑒𝑒𝑑𝑠𝑖 + 𝛽4𝑖 𝑃𝐶_𝑃𝑟𝑜𝑑𝑢𝑐𝑡𝐹𝑜𝑐𝑢𝑠𝑖+ 𝛽5𝑖 𝑃𝐶_𝑁𝐷𝑃𝑇𝑒𝑐ℎ𝐼𝑛𝑛𝑖+ 𝛽6𝑖 𝑃𝐶_𝑇𝑟𝑎𝑛𝑠𝑅𝑒𝑙𝑖 + 𝛽7 𝐵2𝐵_𝐵2𝐶𝑖+ 𝛽8 𝑃𝐷𝐼𝑖+ 𝛽9 𝐼𝑁𝐷𝑖+ 𝛽10 𝑀𝐴𝑆𝑖+ 𝛽11 𝑈𝐴𝑉𝑖+ 𝑢𝑐𝑜𝑢𝑛𝑡𝑟𝑦(𝑖) (3) + 𝑢𝑖𝑛𝑑𝑢𝑠𝑡𝑟𝑦(𝑖)(2) + 𝑒𝑖 𝛽0𝑖= 𝛽0+ 𝑢0,𝑐𝑜𝑢𝑛𝑡𝑟𝑦(𝑖) (3) + 𝑢0,𝑖𝑛𝑑𝑢𝑠𝑡𝑟𝑦(𝑖)(2) + 𝑒0𝑖 𝛽1𝑖= 𝛽1+ 𝑢1,𝑐𝑜𝑢𝑛𝑡𝑟𝑦(𝑖) (3) + 𝑢1,𝑖𝑛𝑑𝑢𝑠𝑡𝑟𝑦(𝑖)(2) 𝛽2𝑖= 𝛽2+ 𝑢2,𝑐𝑜𝑢𝑛𝑡𝑟𝑦(𝑖)(3) + 𝑢2,𝑖𝑛𝑑𝑢𝑠𝑡𝑟𝑦(𝑖)(2) 𝛽3𝑖= 𝛽3+ 𝑢3,𝑐𝑜𝑢𝑛𝑡𝑟𝑦(𝑖) (3) + 𝑢3,𝑖𝑛𝑑𝑢𝑠𝑡𝑟𝑦(𝑖)(2) 𝛽4𝑖= 𝛽4+ 𝑢4,𝑐𝑜𝑢𝑛𝑡𝑟𝑦(𝑖)(3) + 𝑢4,𝑖𝑛𝑑𝑢𝑠𝑡𝑟𝑦(𝑖)(2) 𝛽5𝑖= 𝛽5+ 𝑢5,𝑐𝑜𝑢𝑛𝑡𝑟𝑦(𝑖) (3) + 𝑢5,𝑖𝑛𝑑𝑢𝑠𝑡𝑟𝑦(𝑖)(2) 𝛽6𝑖= 𝛽6+ 𝑢6,𝑐𝑜𝑢𝑛𝑡𝑟𝑦(𝑖) (3) + 𝑢6,𝑖𝑛𝑑𝑢𝑠𝑡𝑟𝑦(𝑖)(2)
Changing the coefficients from fixed to random has significantly contributed to the model, indicated by a significant (p = .021) joint chi square test. This indicates that the variance of the intercepts and slopes are significantly different from zero. Even though, the DIC value has remained approximately at the same level, the information provided by the random coefficients of the individual level explanatory variables significantly differ from zero. For this reason, the random coefficients will be retained in the model. The reason for not including an interaction term between countries and industries can be found in section 4.3.
The information provided by the random intercepts and slopes for countries and industries significantly differs from zero. This indicates that the capabilities have a different influence on revenue growth per country and industry. This difference in behaviour across countries and industries is in alignment with hypothesis 3 (The effect of product centric and customer
TABLE 6 Final Model Results
Fixed Part Parameter
Constant 4.457*** B2B_B2C -.020 Power Distance -.002 Individualism -.002 Masculinity .006* Uncertainty Avoidance .002 Random Part PC_ProductFocusa .272*** PC_NPDTechInna .013 PC_TransRela -.115 CC_VoiceCustomera .118** CC_Culturea .041 CC_NPDCustNeedsa .013*** Level: Country Constant .059 Level: Industry Constant .057 Level: Respondent Constant 1.450 Level: countryXindustry Constant -2*loglikelihood: DIC: 8315.892 pD: 109.343 Units: Country 61 Units: Industry 17 Units: Respondent 2557
Notes:𝒂 Slope estimates vary randomly across countries and industries (joint chi-square test, p <.05); *p < .10, **p
< .05, ***p < .01
Lastly, new products based on services born out of understanding customer needs appears to have a positive significant influence on revenue growth (𝛽= .013, p < .01). This confirms hypotheses 1a ,2a and 2c.
When investigating the model results and covariance matrix, it can be seen that every capability has received a separate intercept and slope per industry and country. A closer look at the different slopes of the significant capabilities provides valuable insights. First, a strong focus
on the core product/service born out of understanding customer needs has mean of .272
(standard error .047). The individual slopes for countries and industries vary around .272 with .007 (standard error .008) at the country level and with .003 (standard error .002) at the industry level. Second, taking the voice of the customer into account in all key business decisions has a mean of .118 (standard error .046), this is only slightly higher than the model with a single slope. However, the individual slopes for countries and industries vary around this mean with a variance of .005 (standard error .005) for the country level and .006 (standard error .006) for the industry level. Lastly, new products or services born out of understanding customer needs has a mean of .013 (standard error .006), indicating a slight increase compared to the single slope model. The individual slopes for country and industries vary around this mean with a variance of .001 (standard error .000) at the industry level
Besides de slopes, the intercepts of the individual lines differ across industries and countries as well. It can be seen that the mean (intercept) of the random coefficients model has changed from 4.489 to 4.457. This is a minor change and indicates that the country and industry combination slopes vary about this mean with a variance of .059 (standard error .044) at the country level and a variance of .057 (standard error .027) at the industry level. At the individual level, scores have a variance of 1.450 around the country and industry slopes.
With the addition of random intercepts and slopes, the final model is complete and the following hypotheses can be accepted:
- Hypothesis 1a (A strong focus on the core product/service has a positive relationship
with business growth)
- Hypothesis 2a (Taking the voice of the customer into account in all key business
decisions has a positive influence on business growth)
- Hypothesis 2c (New products or services born out of understanding customer needs
has a positive influence on business growth).
4.8 Predictions of random slopes
In order to discover patterns of industry and country lines, a closer look is taken at the predictions of the random slopes that can be found in figure 2.
As can be seen from the graphs of Strong Focus on Products/Services, the values stay relatively close to each other. However, the difference between a high and a low score on Strong Focus
on Products/Services across industries and countries, is relatively large and positively adds to
the mean Revenue Growth when scoring higher than roughly 3.5 on this variable. This means that differences between countries and industries are rather small but the effect differs majorly between a high and a low score on Strong Focus on Products/Services.
The variable voice of the customer has a larger variance at smaller revenue growth compared to larger revenue growth. This large variation at lower scores on voice of the customer indicates that there is a larger variation between the negative effects between industries and countries on the mean of this variable. However, the relative variance across countries appears to be slightly smaller than across industries. Something that is surprising in the graph of the variable Voice
of the Customer, is an outlying country that appears to be scoring slightly lower on all values.
This appears to be The Netherlands. Overall the variances on this variable are relatively small ranging from -.36 to .33.
For the variable NPD based on customer needs, the variances are much larger. Especially at low scores on this variable, differences are large at both the industry and country level. Firms rating higher on Revenue Growth see smaller differences between countries and industries than firms rating lower on Revenue Growth. Even though the fact that the higher end of this variable stays closer together the differences appear to be larger than voice of the customer. Furthermore, the difference between high and low scores on NPD based on customer needs, appears to be largest of all variables, ranging from -1.7 to 1,9. This means that a low score on NPD based on
customer needs (e.g. 1), the mean revenue growth will be lowered with -1.7.
The largest variation between the effectiveness of industries and countries can be found at the customer centric capability NPD based on customer needs. Furthermore, this variable shows to have the largest difference in influence on Revenue Growth as well. The smallest variation between industries and countries appears at the product centric capability strong focus on
