TIES WITH KIBS AND SERVICE INNOVATION: THEIR RELATIONSHIP AND THE INFLUENCE OF PERCEIVED INFORMATION QUALITY
A Study among Food Retail Firms in the Netherlands
Author: Niek Bernard Prosper ten Have, № 10642994 Date of submission: July 21, 2014
Study qualification: Master of Science in Business Studies – Strategy Track Institution: University of Amsterdam
First supervisor: Dr. A.S. Alexiev Second supervisor: Dr. B. Lima
TABLE OF CONTENTS
ABSTRACT ... 3
INTRODUCTION ... 4
LITERATURE REVIEW ... 8
Service Innovation ... 9
Ties with KIBS ... 11
Perceived Information Quality ... 13
THEORETICAL FRAMEWORK ... 16
Ties with KIBS and Service Innovation ... 16
Ties with KIBS, Service Innovation, and Perceived Information Quality ... 18
METHODOLOGY ... 20 Research Design ... 20 Independent/Mediating Variable ... 24 Independent/Moderating Variable ... 25 Control Variables ... 28 RESULTS ... 29 Reliability Analysis ... 30 Descriptive Statistics ... 33 Correlation Analysis ... 34 Normality Analysis ... 39 Regression Analysis ... 41 DISCUSSION ... 48
Ties with KIBS and Service Innovation ... 49
Ties with KIBS, Service Innovation, and Perceived Information Quality ... 50
Additional Findings ... 51
Implications for Theory ... 53
Implications for Practice ... 55
Limitations and Future Research Directions ... 57
CONCLUSION ... 59
References ... 61
Appendix A: Tables, Graphs, and Figures ... 67
ABSTRACT
This study examined the relationship between a firm’s ties with knowledge intensive business services (KIBS) and its service innovations, in the context of the Dutch food retail sector. Additionally, this study examined the influence of perceived information quality (PIQ) on ties to KIBS and service innovations. It was hypothesized that, similar to manufacturing firms, service firms would benefit from the broad networks and external knowledge of KIBS,
leading to more service innovations. It was also assumed that a positive managerial perception of the information quality delivered by KIBS would contribute to more extensive ties between firms and these KIBS. Furthermore, was it postulated that PIQ would have a moderating effect on the relationship between a firm’s ties with KIBS and service innovations. Based on data from an internet-mediated questionnaire polling food retail firms in the Netherlands, no evidence was found for a relationship between a service firm’s ties with KIBS (defined in this study as the combined measure of ties with multiple KIBS) and service innovation. However, additional findings did indicate some associations between individual KIBS and specific service-innovation dimensions. Additionally, the results supported the argument that a higher level of PIQ is positively related to a firm’s ties with KIBS. However, no significant
moderating effect was found for PIQ on the relationship between a service firm’s ties with KIBS and service innovation. The results of this study provide worthwhile theoretical and practical contributions to the field of service innovation, KIBS and PIQ, as well as
possibilities for future research.
INTRODUCTION
The service sector is becoming increasingly dominant in developed economies (Tether, 2005). In these developed economies, innovation is seen as the ultimate source of economic growth (Tether, 2005), leading to an increasingly pressing need to understand the factors that create innovations in the service sector. The trend toward the service sector and its innovations is also notable in the current literature, which is pushing this topic into the center of economic policy research (Hipp & Grupp, 2005; Den Hertog, van der Aa, de Jong, 2010; Djellal & Gallouj, 2001; Miles, 2008).
One factor that is deemed to have an effect on service innovations is the ties that a service firm has with knowledge-intensive business services (KIBS). While previous research indicated that KIBS are positively related to innovations in manufacturing firms, little research has been conducted as to whether this relationship holds true with service firms (Hipp & Grupp, 2005). This is because most innovation surveys – an important measurement instrument for ascertaining the determinants of firm innovation – do not specifically focus on the effect of KIBS when investigating firms’ innovations (Tether, 2005). The majority of these surveys ask general questions regarding external information and knowledge sources (Shearmur & Doloreux, 2013). And, surveys that do specifically focus on the effect of KIBS on clients’ innovations are mostly conducted among manufacturing firms. Examples of research on the relationship between KIBS and innovation in a manufacturing context are the works of Shearmur and Doloreux (2013) and Zhang and Li (2010).
Shearmur and Doloreux's (2013) research examined 804 manufacturing firms in Quebec. Their work demonstrated a positive relationship between KIBS and a firm’s innovativeness, and it also showed that the type of innovation depended on the KIBS utilized. This research included fifteen different KIBS and four types of innovations – process, management, product, and marketing.
Zhang and Li (2010) provided another example of research in this same direction. Their study examined 500 new ventures in Guangdong Province, China, and also showed a positive relationship between a combined measure of ties with multiple KIBS and innovation. The KIBS they used in their research were: technology service firms, accounting and financial service firms, law firms, and talent-search firms. Despite the fact that these two research studies took place in different contexts and locations, both results showed that firms’ ties with KIBS are, in general, significantly related to innovations in manufacturing firms.
However, findings on manufacturing firms cannot indiscriminately be applied to service firms. The reason for this is the difference in how service firms operate (Tether, 2005), invest (Hipp & Grupp, 2005), innovate (Miles, 2008), and search for knowledge (Tether, 2005) compared to manufacturing firms. The clearest example of these differences is, of course, the fact that most manufacturing firms are focused on the production of products, while service firms are geared more towards delivering services. It is for this reason that innovations in manufacturing firms are, in general, primarily focused on new core technologies, while innovations in service firms tend to focus on the development of new concepts and procedures (Djellal & Gallouj, 2001). Empirical research by Hipp and Grupp (2005) supported these findings by highlighting the fact that that service firms differ in their innovation patterns as compared to manufacturing firms. Therefore, the dissimilarities between the two types of firms also result in differences in the way in which service firms obtain knowledge as compared with manufacturing firms (Tether, 2005). Both sectors use KIBS to obtain external knowledge, although research by Den Hertog (2000) shows that the KIBS used by manufacturing firms are more focused on information technology (IT) and research services, while service firms are more focused on consultancy. Taken as whole, these differences prompt investigation as to whether the effect of KIBS on innovations in manufacturing firms is similar to the effect of KIBS on innovations in service firms.
Therefore, additional research was required to come to conclusions about the effect of KIBS on service innovations in service firms
So, the first gap indicated was the effect of ties with KIBS on service innovations in service firms. Another gap was found in the literature relating to a concept that might influence the relationship between KIBS and service innovations. Previous literature tested several factors that might influence the relationship between a firm’s ties with KIBS and firm innovation (Shearmur & Doloreux, 2013; Zhang & Li, 2010). Unfortunately, few of these factors were related to the influence of manager heterogeneity. Because individuals, including managers, are heterogeneous by nature (Felin & Foss, 2005), differences in their perceptions can be expected to influence the effect of KIBS. Research by Zhang and Li (2010) that did focus on differences between managers, showed that a manager’s perception towards industry growth can influence the relationship between a firm’s ties with KIBS and firm innovation. The reason for this effect is that the managers’ different perceptions towards industry growth led to variances in the strategic decisions they made regarding the usage of KIBS. Zhang and Li’s (2010) explanation was that a higher level of perceived industry growth reduced the necessity to gain information from KIBS about innovations in other firms. This result emphasizes the possibility that differences in perception have an influence on the relationship between a firm’s ties with KIBS and firm innovation. Additionally, Zhang and Li (2010) acknowledged that it is important for future research to analyze more of these institutional and cultural factors – such as perception – and their influence on the relationship between a firm’s ties with KIBS and firm innovation. One of the institutional factors that is expected to influence the relationship between a firm’s ties with KIBS and firm innovation, is the perception of managers towards the quality of information delivered by KIBS. Previous studies show that the concept of perceived information quality (PIQ) is used in a number of fields (Maltz, 2000), and indicate that a manager’s perception towards information quality can
influence his or her usage of that information (Nicolaou & Mcknight, 2006; Maltz, 2000) and of the external information source (O’Reilly, 1982). However, little research has been conducted regarding specific external information sources. Since firms use KIBS as an external information source to obtain knowledge (Miles & Miles, 2012), it is reasonable to suggest that PIQ has an influence on these factors.
Finding evidence for the influence of PIQ on a firm’s ties with KIBS and service firm innovations is of great importance, as it would enhance the literature on factors that influence service innovations. It would also expand the literature on the influence of PIQ on specific external information sources, thereby adding to the growing research on information-quality research (Ge & Helfert, 2007).
Overall, the aim of this research is to expand upon the relationship between a firm’s ties with KIBS and firm innovations in the service sector, while taking into account the influence of PIQ. Hence, the research question that this paper aims to answer is the following:
To what degree do Knowledge-Intensive Business Services (KIBS) influence service innovations in service firms, and what is the contribution of perceived information quality (PIQ) to this relationship?
In order to answer this question the following sub-questions have been formulated:
1. To what degree are firms’ ties with KIBS related to service innovations in service firms?
2. To what degree do managerial perceptions of information quality towards information obtained from KIBS influence the extensiveness of ties with KIBS?
3. To what degree does PIQ influence the relationship between service firms’ ties with KIBS and service innovations?
The research conducted was quantitative in nature, with data collected through an internet-mediated questionnaire. This questionnaire examined multiple variables, including scales to measure: Ties with KIBS (Zhang & Li, 2010), Service innovation (Janssen, Castaldi, Alexiev & Den Hertog, 2015), and Perceived information quality (Lee, Strong, Kahn, Wang, 2002). Overall, this study provides several theoretical and practical contributions to the literature. First, from a practical point of view, the results of this research are of great importance as they create awareness around managerial perception and its effect on firm innovation. And, with respect to theory, the research expands on previous literature by analyzing the relationship between a firm’s ties with KIBS and innovation in a service context. By analyzing the effect of PIQ, this paper contributes to the further development of institutional understandings of the factors that influence the relationship between a firm’s ties with KIBS and firm innovations.
This paper is structured as follows: First, a critical literature review is provided, covering all of the key concepts involved in this research. Second, a theoretical framework is presented, followed by the methodology. Finally, the empirical results and a discussion section are offered. This section includes both theoretical and practical implications of the research findings, as well as possibilities for future research.
LITERATURE REVIEW
This literature review provides an overview of the main variables utilized in this study – Service innovation, Ties with KIBS, and Perceived information quality. Each of these concepts is discussed in a separate section, starting with the dependent variable – Service innovation.
Service Innovation
Innovation is a broadly used term in the literature, and it has been defined in a variety of ways. Schumpeter (1934) saw innovation as a critical dimension of economic change. As early as the 1930’s, Schumpeter described five types of innovation (1934):
• The restructuring of an industry by the introduction of new forms of competition • The appropriation of new sources of supply of half-manufactured goods or raw
materials
• The improvement of the quality of existing products or the introduction of new products
• Introduction of process methods that are novel to the industry • The opening of new markets
These general innovation types are largely associated with manufacturing and with the development of technologically advanced artifacts (Tether, 2005). This is in alignment with the majority of research on innovation and innovation processes, as most of these studies deal with the manufacturing sector (Gallouj & Weinstein, 1997). Nevertheless, the literature does show an increase in studies that focus on innovation in services (e.g. Tether, 2005; Evangelista, 2000; Gallouj, 2002; Sundbo, 1997; Miles, 2008; Den Hertog et al., 2010). This is of great importance, because services and service firms play an increasingly dominant role in the advanced economies of the world (Tether, 2005). To contribute to this literature, Service innovation is used as one of the variables in this research.
The definition for Service innovation, as used in this study, was adapted from the definitions of Wolpert (2000) and Den Hertog et al., (2010) and reads as follows:
Service innovation entails the pursuing of incremental or radical new service
experiences and solutions, the exploitation of new or potentially disruptive services,
and the introduction of changes into the core concepts or formulas of a business.
To further narrow down this definition, the concept of service innovation has been divided into six dimensions: Service concepts/formulas, Customer interactions, Business partners/value systems, Revenue models, Organizational service delivery systems, and Technological service delivery systems (Den Hertog et al., 2010). The following descriptions are related to each of these dimensions (Den Hertog et al., 2010):
The first dimension, Service concepts/formulas, includes the introduction of a concept or formula that is novel to the market. These new concepts or formulas are generally intangible, such as novel ideas for how to solve organizational problems. But they can also be tangible, particularly when it comes to product delivery (for example, ATMs).
The second dimension is Client interactions. This dimension includes novel ways of interacting with customers, such as by involving them in service design, production, and consumption.
The third dimension includes new Business partners/value systems. This dimension includes new roles for external parties in co-producing the firm’s services as well as the involvement of new partners in the introduction of these services (Janssen et al. 2015).
The fourth dimension relates to new Revenue models. This includes changes in the way revenue is generated and changes in the firm’s financial construction (Janssen et al. 2015).
The fifth dimension concerns new Organizational service delivery systems. This dimension refers to the human resources side of changes in the firm’s service delivery system. It also entails adjustments in the way service workers deliver services to customers, as well as changes in the organizational structure to enable the introduction of new services.
The last dimension relates to changes in the Technological service delivery systems.
This includes the introduction and development of new technologies, such as the introduction of new IT systems.
When a service innovation is implemented, it usually includes more than one of these dimensions (Den Hertog, 2000; Miles, 2008). A good example of an occasion where a combination of these dimensions would be observable is the hypothetical need for a new organizational service delivery system as a firm plans to introduce a new IT system. In this case, multiple dimensions are needed to make the service innovation successful. To obtain knowledge for developing the different types of service-innovation dimensions, some form of knowledge input is required. One form of knowledge input is via the firm’s ties with KIBS (Shearmur & Doloreux, 2013). The role that KIBS have in the creation and delivery of knowledge input and the ties that firms have with these KIBS are further explained in the following section.
Ties with KIBS
This section begins with a definition for Knowledge-Intensive Business Services (KIBS), in order to clarify how this concept should be interpreted in this study. The definition used here comes from the research of Den Hertog (2000, p505), which describes KIBS as:
• “private companies or organizations
• relying heavily on professional knowledge, i.e. knowledge or expertise related to a specific (technical) discipline or (technical) functional domain; and,
• supplying intermediate products and services that are knowledge based.”
In this study, the variable Ties with KIBS is researched. This variable will examine the combined measure of ties with multiple KIBS. The KIBS that are combined in this variable
are: IT firms, Marketing service firms, Financial service firms, Other consultancy firms, Law firms, and Accountancy firms
The services that these KIBS deliver are focused on sharing knowledge, consulting, and developing systems for their client firms (Doloreux & Shearmur, 2013; Den Hertog, 2000). These KIBS do this by sharing their experiences, solving problems, training, advising, and connecting different firms with each other (Den Hertog, 2000). Firms are willing to pay for these structured ways of obtaining expertise and knowledge (Doloreux & Shearmur, 2013; Wolpert, 2002), and they use KIBS to broaden their own knowledge base and to find solutions for their problems (Gadrey, Gallouj, & Weinstein, 1995; Den Hertog, 2000; Miles & Miles, 2012). Overall, the function of KIBS can be described as a transferor between general, economy-wide knowledge and firm-specific knowledge (Miles & Miles, 2012). By being a transferor, KIBS can contribute to different forms of innovation (Zhang & Li, 2010).
Research regarding the effect of KIBS on innovations started to receive more attention during the beginning of the 1990’s (Shearmur & Doloreux, 2013). Initially, firms often relied on their internal capabilities to develop innovations, which meant a low need for external knowledge sources (Doloreux & Shearmur, 2013). However, more recent research has shown the positive effect of external knowledge on innovations and has indicated how firms can gain this external knowledge through indirect and direct network ties (Ahuja, 2000; Powell, Koput, & Smith-Doerr, 1996). It was shown that ties with KIBS can broaden a manufacturing firm’s external search scope, which has a positive effect on a firm’s level of knowledge and ability to innovate (Hipp & Grupp, 2005; Shearmur & Doloreux, 2013; Zhang & Li, 2010). Because KIBS are linked to a wide variety of firms, they gain unique network positions that enable them to transfer knowledge from one organization or industry to another (Zaheer & McEvily, 1999; Zhang & Li, 2010).
KIBS are not only innovators on their own – they also function as a transmitter,
facilitator, and source of innovation for their client firms (Doloreux & Shearmur, 2012; Den Hertog, 2000). KIBS are seen as an innovation ‘transmitter’ when they transfer knowledge from one firm, industry, or cluster to another; as an innovation ‘facilitator’ when they assist clients in their innovation processes; and as an innovation ‘source’ when they are co-producers of the innovations inside their clients’ firms (Doloreux & Shearmur, 2013; Den Hertog, 2000; Miles, 2008). Previous research has shown that when the variety of a manufacturing firm’s ties with KIBS increases, the probability of gaining innovations from these KIBS also increases (Zhang & Li, 2010; Shearmur & Doloreux, 2013). However, the usefulness of KIBS is context-specific, and therefore is it important to make a thoughtful decision in terms of which KIBS to utilize. When deciding which KIBS to utilize and in which way, firms depend on several variables. One of these variables is the perception of managers. Differences in managers’ perceptions can have a great impact on the kind of advice that is obtained, as well as the information that is used (Nebus, 2006). For this study, one form of these perceptions is analyzed, namely the perception of the information quality delivered by KIBS. The following section elaborates on this concept of PIQ.
Perceived Information Quality
When firms seek advice, they have access to a variety of internal and external knowledge sources. It is key for managers to decide which knowledge source to use and which information to apply. When deciding which knowledge source is most useful for the firm, managers make several tradeoffs between the expected knowledge value and the cost of obtaining it (Nebus, 2006). Examples of these tradeoffs are provided by the literature, such as: Nebus (2006), which discusses the tradeoffs between expertise and trustworthiness; (Krackhardt & Hanson, 1993), which examines the choice between exploitation and exploration; and (March, 1991), which looks at the tradeoff between cost and information
quality (Haas & Hansen, 2002). As can be seen, when managers decide to use information delivered by KIBS, they consider several of these tradeoffs. This paper focuses on managers’ perceptions of information quality, the criteria linked to PIQ and the effect of this perception on a firm’s ties with KIBS and service innovation.
Managers’ perceptions of information quality, also known as perceived information quality (PIQ), are defined by Nicolaou & McKnight (2006, p335) as the: “cognitive beliefs about the favorable or unfavorable characteristics […] of the exchanged information.” In other words, PIQ can be seen as the positive and negative thoughts a person has towards the characteristics of the exchanged information. PIQ is judged using four general dimensions – intrinsic information quality, contextual information quality, representational information quality, and accessibility information quality (Wang & Strong, 1996; Lee et al., 2002). These dimensions each have several related criteria (Lee et al., 2002), and the importance of each criterion is dependent on the context (Nicolaou & McKnight, 2006). The information quality dimensions, and their related criteria, entail the following (Wang & Strong, 1996; Lee et al., 2002, Kahn, Strong & Wang, 2002):
• The intrinsic information quality dimension is the extent to which information is believable, correct, reliable, and certified free-of-bias. This dimension has four related criteria – believability, reputation, objectivity, and free-of-error (accuracy). • The dimension of contextual information quality indicates the extent to which the
information is relevant, applicable, and helpful for the task at hand. It contains five criteria – value-added, relevance, completeness, timeliness, and appropriate amount of information.
• The representational information quality dimension concerns the extent to which information is easy to interpret, and it includes four criteria – understandability, interpretability, concise representation, and consistent representation.
• The fourth dimension, accessibility information quality, is the extent to which the information is accessible, available, or easily and quickly retrievable. This dimension includes three criteria – accessibility, ease of manipulation (ease of operations), and security.
Together, these four dimensions are useful for ensuring complete coverage of all relevant criteria when defining the concept of PIQ (Lee et al., 2002). A further description of these different criteria is presented in Appendix A, Table 16. The criteria were distributed by Kahn et al., (2002) into two categories, namely product quality and service quality. Service quality consists of the criteria related to the service process with measures that are intangible, such as: timeliness, security, believability, accessibility, ease of manipulation, reputation, and value-added (Kahn et al., 2002). Product quality, on the other hand, contains the criteria related to a product’s features that are tangibly measurable, such as: concise representation, completeness, consistent representation, appropriate amount of information, relevance, and understandability (Kahn et al., 2002). The criteria of both categories – product quality and service quality – are important when seeking to understand a firm’s ties with KIBS, since KIBS are seen as suppliers of both intermediate products and services (Den Hertog, 2000). Taken as a group, all of the criteria of PIQ are combined into one of the variables of this research, namely Perceived information quality, and specifically, mangers’ perceptions of the information quality of information delivered by KIBS. An explanation of the expected relationship between PIQ, a service firm’s ties with KIBS, and service innovation is provided in the following chapter.
THEORETICAL FRAMEWORK
This chapter describes the hypotheses, as well as the expected relationships between the different variables. Starting with the relationship between a service firm’s ties with KIBS and service innovation. Followed by the influence of PIQ on the relationship between a service firm’s ties with KIBS and service innovation. An overview of all hypotheses is provided in Figure 1.
Figure 1. Conceptual Model
H1
H2 H3
Ties with KIBS and Service Innovation
Research shows that managers use different sources of information for the completion of their projects (Nebus, 2006). Initially, most of these information sources were located inside the firm, as firms relied mainly on internal knowledge to create innovations and to find solutions for problems (Doloreux & Shearmur, 2013). More recent research indicates that external knowledge is also an important factor when it comes to innovation (Kline & Rosenberg, 1986; Chesbrough, 2003). The external resources used by firms bring in external knowledge and information that is complementary to the firms’ internal knowledge (Sammarra & Biggiero, 2008). The complementary effect of external knowledge enables firms to learn and to extend
Ties with
Knowledge Intensive Business Services
Service Innovation:
- Service concepts/formulas - Customer interactions
- Business partners/value systems - Revenue models
- Organizational service delivery systems - Technological service delivery systems Perceived Information Quality
Control variables: - Firm size
- Firm age - Firm type
their knowledge bases (Lazzarotti & Manzini, 2009). So, both internal and external knowledge are seen as important components for firms seeking to innovate. External knowledge enables firms to obtain new skills, information, and knowledge (Sammarra & Biggiero, 2008), while internal knowledge enables firms to assimilate and transform this external knowledge into new processes, services, and products (Eisenhardt & Martin, 2000). As knowledge becomes an increasingly significant component in a firm’s ability to compete, firms are becoming more knowledge-intensive. An increase in the employment of more highly educated employees in the European Union shows this trend of industries and services towards knowledge intensification (Hipp & Grupp, 2005). Firms in both sectors are restructuring their businesses, intensifying their IT utilization, and specializing in their core businesses (Hipp & Grupp, 2005). Taken as a whole, these aspects enhance firms’ needs to attract external knowledge. One way to obtain this external knowledge is by forming ties with KIBS. KIBS have broad networks, which enable them to combine and distribute knowledge (Den Hertog & Bilderbeek, 1997).
Research has shown the positive link between external knowledge provided by KIBS and the innovation level of manufacturing firms (Den Hertog, 2000; Shearmur & Doloreux, 2013; Zhang & Li, 2010). Despite this relationship, little research has been conducted towards the effect of KIBS on innovations in service firms – a gap described earlier in this paper. Due to clear differences between service and manufacturing firms, the effect of manufacturing firms’ ties with KIBS on innovations cannot indiscriminately be generalized to service firms (Hipp & Grupp, 2005). Some of the disparities between these two types of firms include differences in the way they: operate (Tether, 2005), invest (Hipp & Grupp, 2005), innovate (Miles, 2008), and search for knowledge (Tether, 2005). Regardless of these differences, both type of firms make use of KIBS to gather external knowledge. And both types of firms are expected to profit from the broad networks of these KIBS, which can
enhance a firm’s innovativeness (Zhang & Li, 2010). Nevertheless, no research was found that showed statistical evidence for the positive relationship between a service firm’s ties with KIBS and service innovation. Regarding this potential relationship, the following hypothesis was established:
Hypothesis 1: A service firm’s ties with KIBS will be positively related to service innovation.
To test this hypothesis, each of the six dimensions of service innovation are analyzed - Service concepts/formulas, Customer interactions, Business partners/value systems, Revenue models, Organizational service delivery systems and Technological service delivery systems. Based on previously reported results, the hypothesis for each of these dimensions is similar – a service firm’s ties with KIBS will be positively related to all six service-innovation dimensions. Additional support for the above hypotheses is that service innovations usually contain more than one service-innovation dimension when implemented (Den Hertog, 2000; Miles, 2008). So, when KIBS have a positive effect on one service-innovation dimension, it can be expected that other dimensions would be influenced as well. Therefore, even when KIBS do not have a direct effect on each service-innovation dimension, it is still very likely that there will be an indirect positive effect.
Ties with KIBS, Service Innovation, and Perceived Information Quality
Managers can differ substantially in their perceptions and ways of behavior. This paper focuses on managers’ perceptions of the quality of information delivered by KIBS. As research shows, managers’ perceptions of information quality can influence their intent to use certain information (Nicolaou & Mcknight, 2006; Maltz, 2000) and the frequency with which they utilize external information sources (O’Reilly, 1982). A higher PIQ leads to a greater intent to use the provided information (Maltz, 2000) or source (O’Reilly, 1982), and a lower 18 | P a g e
PIQ leads to a less frequent use of the specific information source (O’Reilly, 1982). In an ideal situation, managers are expected to use those information sources that are perceived to offer the highest quality information (O’Reilly, 1982). This research by O’Reilly (1982) was conducted using several information sources, with only one of the elements being external information sources. The concept of external information sources is very broad, and no further research was found that specifically focused on the relationship between PIQ and a firm’s ties with KIBS. For that reason, the effect of PIQ on a firm’s ties with KIBS has been researched in this study. And, because KIBS are seen as external information sources, the results of O’Reilly (1982) created the expectation that managers will use KIBS more frequently when the PIQ of these KIBS is higher. In short, a higher PIQ is expected to result in more extensive ties with KIBS.
Hypothesis 2: A higher PIQ towards information delivered by KIBS will be positively related to firms having more extensive ties with KIBS.
In addition to the expected relationship between PIQ and a firm’s ties with KIBS, PIQ is also expected to influence the relationship between a firm’s ties with KIBS and service innovation. As described, external sources – such as KIBS – can expand service firms’ knowledge bases by providing them with external information and knowledge that complement their internal knowledge (Sammarra & Biggiero, 2008). This external knowledge and information is seen as an important factor in firms’ ability to innovate (Kline & Rosenberg, 1986; Chesbrough, 2003). Therefore, when firms do not use the external information provided by KIBS, it might result in a decrease in their innovativeness.
A factor that can influences the intent to use information is managers’ perceptions of the information quality (Maltz, 2000). A lower PIQ leads to a lower intent to use the information, and vice versa (Maltz, 2000). Therefore, when the information quality as delivered by KIBS is perceived as low, the knowledge and information delivered by these 19 | P a g e
KIBS is expected to be used less often. Not using this external information and instead putting more effort into creating internal knowledge will have consequences for the innovativeness of the firm (Powell et al., 1996). The reason for this is that not using the external knowledge and information provided by KIBS is expected to lessen the firm’s opportunity to broaden its network and knowledge base, both of which are important factors in innovation (Zhang & Li, 2010). The expectation is that a higher PIQ will lead to a stronger relationship between a firm’s ties with KIBS and service innovation, and that a lower PIQ will lead to a weaker relationship between KIBS and service innovation. Based on these expectations, the following hypothesis was formulated:
Hypothesis 3: The positive relationship between a firm’s ties with KIBS and service innovation will be stronger when managers’ PIQ of the information delivered by KIBS is higher.
METHODOLOGY
Research Design
This research used quantitative data collection to form conclusions. The research technique used was an internet-mediated questionnaire. The questionnaire was used to measure the variables of this research – Service innovations, Ties with KIBS, and Perceived information quality. The questionnaire was also used to measure the characteristics of the firm – Firm size, Firm age, and Firm type – and the characteristics of the manager – Age, Gender, Organizational position, Educational level, Working experience, and Strategic decision-making involvement. In line with previous research, Firm size, Firm age (Shearmur & Doloreux, 2013; Alexiev, Jansen, Van den Bosch & Volberda, 2010), and Firm type (Zhang & Li, 2010) were used as control variables. The reason for using quantitative data is that previous qualitative research has already analyzed the above-mentioned variables
independently, while this research examines the relationship between the variables.
Empirically, this paper will test these relationships using a sample comprised of retail firms located in the Dutch food retail sector. The food retail sector in the Netherlands is known for being highly competitive and dynamic (Quix & Van der Kind, 2012). This requires retailers to become experts in continuously renewing their services in order to respond to fading retail concepts (Burns, Enright, Hayes, McLaughlin, & Shi, 1997; Quix & Van der Kind, 2012). The fading of these retail concepts is caused by several developments, such as changes in consumer demand, information systems improvements, and an increase in the intensity of competition (Burns et al., 1997; Quix & Van der Kind, 2012). For this reason, retailers are required to continuously innovate in order to stay ahead of competition. Ties with KIBS are deemed to play an important role in Dutch retailers’ search for innovation. Thus, the Dutch food retail sector provides a rich context in which to test the proposition that a service firm’s ties with KIBS can contribute to its service innovations.
To indicate which KIBS are frequently used by retail firms in the Netherlands, two retail experts, Dr. Laurens Sloot and Frank Quix, were asked for their advice. Both of them indicated that consulting, IT, and marketing firms play significant roles in retail and are relevant in the Netherlands, which is in line with the research of Den Hertog (2000). In addition, financial firms, law firms, and accounting firms will also be included in this research, as these are commonly used KIBS by firms in general. These KIBS were also researched in previous studies dealing with manufacturing firms (Zhang & Li, 2010; Shearmur & Doloreux, 2013). The combined measure of a firm’s ties with these six KIBS result into one variable, which is the independent and mediating variable, Ties with KIBS. Additionally, all of the criteria of PIQ are combined to create the independent and moderating variable, Perceived information quality. For the dependent variable of Service innovation, each of the six service-innovation dimensions are analyzed separately, to indicate if the
influence of a firm’s ties with KIBS and PIQ is different for each innovation dimension. Overall a similar effect is expected for each of these six service-innovation dimensions – Service concepts/formulas, Customer interactions, Business partners/value systems, Revenue models, Organizational service delivery systems and Technological service delivery systems.
Research Sample
As mentioned above, the data for this research will be collected from food retail firms located in the Dutch retail market. It is important to randomly select the retailers to prevent biases. It can, for instance, be expected that retailers located near Amsterdam make more use of KIBS, due to a higher density of KIBS near those retailers. To avoid these biases, a random selection of 961 retail firms was selected from a list of small- to large-sized retailers. Bureau van Dijk, one of the world’s leading comprehensive sources of business information and financial data on companies worldwide (including retail firms in the Netherlands), compiled this list (Bureau van Dijk, 2014). The reason for analyzing small- to large-sized firms is that an empirical analysis that merely focuses on medium or larger firms would miss out on a majority of relevant firms and could therefore skew the results (Preissl, 1997). The questionnaire was designed in Dutch and English and pretested by two retail managers to ensure both the correct usage of terminology and validity.
Potential respondents were contacted in advance by telephone in order to assess their willingness to cooperate with the research. Those that were willing to cooperate – 366 firms – received an additional email including the link to the questionnaire. In total, 118 of these firms filled in the questionnaire, resulting in a response rate of 12.28 percent of the total sample, a figure slightly lower than the average response rate for internet-based surveys in the Netherlands (Deutskens, De Ruyter, Wetzels, & Oosterveld, 2004). A reasonable cause for this difference is the fact that the respondents for this research had to be related to a firm – in
this case, a food retail firm.
Five of the respondents missed crucial data, and were therefore removed from the analysis, resulting in N = 113. These respondents had the function of director, manager, or highly involved employee; 22 years of average work experience; and a mean on strategic decision-making involvement of 6.4 on a 7-point scale (1 = very uninvolved; 7 = very highly involved) (Zhang & Li, 2010). These results suggest that the respondents were experienced and knowledgeable about the topics being researched. Frequency analysis, reliability analysis, normative analysis, correlation analysis, regression analysis, and additional analyses were conducted to analyze the different variables of this research and their underlying relationships.
Dependent Variable
– Service Innovation (Janssen et al, 2015) (6 dimensions, 14 items)
Den Hertog et al., (2010) separated service innovation into six dimensions – Service concepts/formulas (3 items, a = 0.88), Customer interactions (2 items, a = .775), Business partners/value systems (3 items, a = 0.74), Revenue models (1 item), Organizational service delivery systems (2 items, a = 0.75), and Technological service delivery systems (2 items, a = 0.78). Most of the time, these dimensions are related to each other, which implies that a change in the service concept might require a change in the service delivery system (Miles, 2008). To more clearly distinguish innovativeness among retail firms, it was decided to analyze all six dimensions of service innovation. Because this research seeks to determine the effect of ties with KIBS on innovations in service firms, focusing on one specific innovation dimension could have biased the results.
The instrument used to collect the data is the measurement instrument of Janssen et al., (2015) which includes 14 statements that are useful for measuring all six dimensions of service innovation. Each of these 14 statements was rated on a 7-point Likert scale (1 =
strongly disagree, and 7 = strongly agree). An example of one of the statements used to measure a particular dimension of service innovation is the following: “Our organization developed new channels for communication with her customers” (Janssen et al., 2015). Table 1 provides further descriptions of the statements included in the questionnaire.
Independent/Mediating Variable
- Ties with KIBS (Zhang & Li, 2010) (6 items, a = 0.74)
To measure Ties with KIBS, respondents were asked to indicate on a 7-point Likert scale (1 = very little, and 7 = very extensive) the degree to which their firms had close relationships with: (1) IT firms, (2) Marketing service firms, (3) Financial service firms, (4) Table 1. Dependent Variable – Service Innovation
№ Control variable Items
1. Service
concepts/formulas
• Our organization developed new (service) experiences or solutions for customers.
• We combined existing services into a new formula.
• We developed a new way of creating value for ourselves and our customers.
2. Customer
interactions
• Our organization developed new channels for communicating with her customers.
• The way we have contact with our customers is renewed. • We changed the task distribution between ourselves and
our customers.
3. Business
partners/value systems
• The role of external parties in producing our services is renewed.
• We involved new partners in the delivery of our services.
4. Revenue models • By introducing new services we changed the way we
generate revenues.
• The way we get paid (financial construction) is altered.
5. Organizational
service delivery systems
• We changed our organization in order to produce our new services.
• Our production of new services requires new skills from our employees.
6. Technological
service delivery systems
• Technology plays an important role in the renewed production of our services.
• We renewed our service offerings by new or different use of ICTs.
Note: source: Janssen et al., (2015)
Other consultancy firms, (5) Law firms, and (6) Accountancy firms. This measurement instrument is similar to the one used by Zhang and Li (2010), who researched the effect of manufacturing firms’ ties with KIBS on product innovations in such firms. Because the items can be generalized to service firms, no adaptations were needed to use these items.
After conducting the reliability analysis, these six items were combined to an average to create the independent/mediating variable – Ties with KIBS. In total, the variable Ties with KIBS includes 6 items with a Cronbach’s alpha of 0.74. A review of these items is reported in Table 2.
Independent/Moderating Variable
- Perceived information quality (24 items, a = 0.96)
To capture all of criteria related to information quality it is helpful to utilize four distinct information-quality dimensions – intrinsic information quality, contextual information quality, representational information quality, and accessibility information quality (Lee et al., 2002). These dimensions, as described in the literature review section of this paper, are useful in defining PIQ. But when it comes to measuring information quality, it is better to divide the criteria according to the product and the service performance model for information quality (PSP/IQ) (Kahn et al., 2002; Lee et al., 2002). Lee’s et al., (2002). The 2 x 2 PSP model Table 2. Independent/Mediating Variable – Ties with KIBS
№ Independent/Mediating
variable
Items
1. Ties with KIBS • Indicate the extent to which your venture
had contact/close relationships with the following firms in the past three years:
(1) IT firms, (2) marketing service firms, (3) financial service firms, (4) other consultancy firms, (5) law firms and (6) accountancy firms.
Note: source: Zhang & Li (2010)
consists of four quadrants: ‘product quality’ versus ‘service quality’, and ‘conforms to specifications’ versus ‘meets or exceeds customers expectations’ (Table 3). These quadrants divide the information-quality criteria into four categories: Sound information, Dependable information, Useful information, and Usable information (Kahn et al, 2002) – which are used in this research to analyze PIQ.
Table 3. The PSP/IQ model Conforms to Specification Meets or Exceeds Consumers Expectation Product Quality Sound Information • Free-of-error • Concise Representation • Consistent Representation Useful Information • Appropriate Amount • Relevancy • Understandability • Interpretability • Objectivity Service Quality Dependable Information • Timeliness • Security Usable Information • Believability • Accessibility • Ease of Manipulation • Reputation • Value-Added Note: source: Lee et al., (2002)
Previous research has indicated that the criteria in each of these categories show a high internal consistency – soundness (a = 0.94), dependability (a = 0.83), usefulness (a = 0.93), and usability (a = 0.94) (Lee et al., 2002). For that reason, it was decided to combine the criteria for each category into one item. By doing so, fewer items were needed to measure PIQ, which increased the likelihood that respondents would fill in the complete questionnaire. An example of a categorical item with compiled criteria is the one used for the category Usable information, which states the following:
Usable information is information that is believable and accessible, has a good reputation for quality, and is easy to combine with other information.
When rating information from the following sources, how usable do you say the information usually is?
Respondents had to rate the items for PIQ on a scale of 0 to 10 (0 = not at all, 10 = completely). This is similar to the scale used by Lee et al., (2002), who introduced a measurement instrument that fits the requirements of this research.
Overall, the PIQ variable consisted of four items per KIBS resulting in 24 items. These 24 items were combined into one independent/moderating variable – Perceived information quality –, which will enable this research to determine the effect of PIQ on the relationship between a firm’s ties with KIBS and service innovation. An overview for each of the items related to PIQ is found in Table 4.
Control Variables
Following previous research, this study controlled for various variables (Zhang & Li, 2010; Shearmur & Doloreux, 2013; Alexiev et al., 2010). The first control variable is Firm age (i.e., the number of years since a firm’s founding), because it is indicated in previous research that older firms produces less innovations than younger firms (Hansen, 1992; Shearmur & Doloreux, 2013). The second control variable is Firm size (the natural log of the number of Table 4. Independent/Moderating Variable – Perceived Information Quality
№ Independent/Moderating
variable
Items
1. Perceived information quality
• Sound information is information that is complete, free-of-error, has a concise representation and a consistent
representation. When rating information from the following sources*, how sound do you think the information usually is? • Information that is sufficiently current to your work and
protected against unauthorized access, is known as
dependable information. When rating information from the following sources*, how dependable do you think the information usually is?
• Information that is relevant, understandable, interpretable, objective, and of an appropriate amount, is known as useful information. When rating information from the following sources*, how useful do you think the information usually is? • Usable information is information that is believable and
accessible, has a good reputation for quality, and is easy to combine with other information. When rating information from the following sources*, how usable do you think the information usually is?
Notes: - *(1) IT firms, (2) Marketing service firms, (3) Financial service firms, (4) Other consultancy firms, (5) Law firms, (6) Accountancy firms.
- source: adapted from Lee et al.,(2002)
full-time employees in 2014). This control variable is included, because larger firms might lack the flexibility to pursue innovations (Alexiev et al., 2010). The third control variable is Firm type (franchise firm = 0, and independent firm = 1). Following Zhang and Li (2010), this control variable was used, as independent firms are expected to be less obligated to follow rules when it comes to innovating, which could cause differences in how independent firms innovate compared with franchise firms. Items for these control variables are reported in Table 5.
RESULTS
This chapter of the paper describes the research results. First, the reliability analyses are reported for the variables with multiple items. The Cronbach’s alpha is analyzed for the six dimensions of the dependent variable: Service innovation (Janssen et al., 2015) and for the items of the independent/moderating and independent/mediating variables: Perceived information quality and Ties with KIBS (Zhang & Li, 2010). In addition, a factor analysis was conducted to further analyze which items could be combined into individual variables.
After the reliability analysis, an overview is reported of the descriptive statistics. The Table 5. Control Variables
№ Control variable Items
1. Firm size • How many employees does your firm currently have in FTEs (full-time equivalent)?
2. Firm age (years) • What is the year of founding of your firm? 3. Firm type • Franchise firms (Plus franchise, Jumbo franchise,
Albert Heijn franchise, Coop franchise, Other franchises)
• Independent Firms
descriptive statistics section highlights the most noteworthy data found by this research. A broader description of these statistics, including the less notable statistics, is found in Appendix A, Table 17 and Table 18.
The section following the descriptive statistical analysis describes the correlation analysis. The correlation analysis gives a detailed overview of correlations between the dependent, independent/moderating, independent/mediating, and control variables. It also describes the mean, standard deviation, and minimum and maximum of each of these variables. The minimum and maximum are reported, because different scales are used to measure the variables. This section includes further clarification on the most significant correlations.
The last two sections of this chapter describe the normality analysis and the regression analysis. The normality analysis provides information about the skewness and kurtosis of the dependent, independent/mediating, and independent/moderating variables. Additionally, Appendix A, Table 19 and Graphs 1, 2, 3, 4, 5, 6, 7, and 8 provide an overview of the variables’ details, their skewness and kurtosis, and the related histograms.
The final section, as mentioned, describes the regression analyses. Regression analyses are used to test the hypotheses. In this research, the process regression analysis of Hayes (2008) was used, which enabled combined moderation and mediation testing. This moderated mediation model is reported in Hayes (2008) under model number 74.
Reliability Analysis
This research used eight multiple-item variables to measure the dependent, independent/mediating, and independent/moderating variables. The dependent variable (with six dimensions) is Service innovation, the independent/moderating variable measures Perceived Information quality, and the independent/mediating variable measures Ties with
KIBS.
The dependent variable Service innovation used one to three statements on the questionnaire to measure each of the six dimensions. Five of these dimensions had a Cronbach’s alpha between 0.74 and 0.88. This is a Cronbach’s alpha above 0.70, meaning that the statements measuring these dimensions had a good internal consistency (Pallant, 2010). Factor analysis indicated that for the dimension Revenue models (a = 0.58), one of the two questionnaire statements should be removed to construct a measurement variable for Revenue models. Additionally, one questionnaire statement was removed from the service-innovation dimension, Customer interactions, in order to increase the Cronbachs’s alpha for this variable from 0.74 to 0.77. After removing these two statements, six variables for innovation could be computed: Service concepts/formulas, Customer interactions, Business partners/ value systems, Revenue models, Organizational service delivery systems and Technological service delivery systems. The eigenvalues and percentage variances for each of these variables are found in Table 6: Rotated Component Matrix – 12 Items of Service Innovation.
Table 6. Rotated Component Matrix – 12 Items of Service Innovation
Component 1 2 3 4 5 6
Service concepts/formulas - Item 1 0.86
Service concepts/formulas - Item 2 0.83 0.32
Service concepts/formulas - Item 3 0.84
Customer interactions - Item 1 0.42 0.83
Customer interactions - Item 2 0.35 0.76 0.33
Business partners/value systems - Item 1 0.82 Business partners/value systems - Item 2 0.85
Revenue models - Item 2 0.96
Organizational service delivery systems - Item 1 0.87
Organizational service delivery systems - Item 2 0.32 0.47 0.64 Technological service delivery systems - Item 1 0.31 0.68 0.30 Technological service delivery systems - Item 2 0.30 0.74
Eigenvalues 5.50 1.37 1.08 0.92 0.68 0.60
Percentage Variance 45.85 11.41 9.00 7.68 5.66 5.03
Number of test measures 3 2 2 2 2 1
Notes: - Loadings =>0.30
- Each item consists out of one statement
The independent/moderating variable Perceived information quality consisted of 24 items, with a mean of 0.53 and a Cronbach’s alpha of 0.96. Factor analysis further reported an eigenvalue of 13.22 for the first component, with a percentage variance of 55.07. According to Pallant (2010), this indicates internal consistency and reliability for the different items measuring PIQ. Therefore, the different items composing PIQ could be computed into one variable.
The six items measuring the independent/mediating variable were also computed into one variable – Ties with KIBS. The items of the variable Ties with KIBS had a mean of 0.32 and indicated internal consistency, with a Cronbach’s alpha of 0.74. This figure is above 0.70, meaning that the items of this variable had a good internal consistency (Pallant, 2010). Removing items did not increase the Cronbach’s alpha. Therefore the items used to measure Ties with KIBS could be computed into an individual variable. An overview of the reliability analysis of each of the variables is reported in Table 7: Reliability Analysis - Variables.
Table 7. Reliability Analysis - Variables
Variable Level Cronbach’
s Alpha
Cronbach’s Alpha based on Standardized Items N Service Innovation Service concepts/formulas 0.88 0.89 3 Customer interactions 0.78 0.78 2
Business partners/value systems 0.74 0.74 2
Revenue models - - 1
Organizational service delivery systems
0.75 0.75 2
Technological service delivery systems 0.78 0.78 2 Ties with KIBS* - 0.74 0.74 6 Perceived information quality - 0.96 0.96 2 4 Note: * The KIBS used in this research are: IT firms, Marketing service firms, Financial service firms, Other consultancy firms, Law firms, and Accountancy firms.
Descriptive Statistics
This research analyzed 18 variables. Six of these variables are the six dimensions of service innovation as dependent variables: Service concepts/formulas, Customer interactions, Business partners/value systems, Revenue models, Organizational service delivery systems, and Technological service delivery systems. Two other variables are the independent/mediating and independent/moderating variables – Ties with KIBS and Perceived information quality. There are also three control variables – Firm age, Firm size, and Firm type. The other five variables were measured to provide further descriptions of the respondents.
In total, 961 firms were contacted by telephone and asked if they were willing to complete this study’s questionnaire. Of these 961 firms, 366 of them replied positively that they were willing to receive the questionnaire. In total, 118 of these possible respondents replied to the email, resulting in a response rate of 12.28 percent. Five of these respondents filled in “Not Applicable” for the items related to the variable Ties with KIBS. These respondents were not willing to give this information or did not make use of KIBS, and therefore these five respondents were removed from the dataset, resulting in a usable response rate of N = 113. To assess the non-response bias, the responding firms were compared with the non-responding firms (Zhang & Li, 2010), resulting in no great differences among firm type or size. To ensure the quality of the data, the respondents were monitored in terms of their Position in their firm, Work experience, and Strategic decision-making involvement. The results showed that 96.5 percent of the respondents had the Position of manager or director. The mean for Work experience in the food sector was 22 years. And the Strategic decision-making involvement resulted in a mean of 6.44 on a 7-point scale (1 = very uninvolved, 7 = very highly involved). These results indicate that the respondents were experienced and knowledgeable about the topics of this study (Zhang & Li, 2010). A complete overview of the
statistics for all the variables is found in Appendix A, Table 17 and Table 18.
From a total of 113 respondents, 111 filled in their Firm type – 18 (16.2%) of the firms were independent firms, 23 (20.7%) of the firms were Plus franchise, followed by Jumbo franchise with 22 (19.8%) firms. Both Albert Heijn franchise and Coop franchise accounted for 12 (10.8%) firms, and the largest group, with 24 (21.6%) firms, was the group Other franchises. These results indicated that 83.8 percent of the firms were franchises and that 16.2 percent were independent firms.
The sizes of these firms were divided according to the number of full-time equivalent (FTE) workers, in line with the European scale for small- and medium-sized firms. The firms were divided as follows: 24 (21.2%) were less than 10 FTEs, 60 (53.1%) firms had 10 to 49 FTEs, 25 (22.1%) had 50 to 199 FTEs, and the smallest group, with more than 200 FTEs, accounted for 4 (3.5%) firms.
It is also notable that the Age of the respondents had a mean of 41.78 years and a standard deviation of 10.50 years. It is also noteworthy that the Gender of the respondents was divided unequally, with 104 (92%) male respondents and 9 (8%) female respondents.
Correlation Analysis
To analyze the relationships between the variables used, this research made use of the Pearson product-moment correlation coefficient. An overview of the results of this bivariate correlation analysis is found in Table 8: Correlation Matrix. In this table, the significant correlations are marked with one or two asterisks, depending on if the correlation is significant at the 5% level (*) or at the 1% level (**). As the data of the two categorical controlling variables, Firm age and Firm size, were obtained as continuous variables, they were also implemented as continuous variables in the correlation analysis. The other categorical control variable, Firm type, was converted into a dummy variable (0 = franchise
firm, and 1 = independent firm), to enable a correlation analysis with each level of these three variables. This resulted in 11 variables, which are listed in Table 8.
The following section of this chapter elaborates on the most important correlation coefficients that have a p-value of 0.05 or lower. First, a description is given of the correlations between the six dimensions of the dependent variables. The next correlations described are the correlations between the dependent variables and the independent/moderating variable. This is followed by the correlations between the dependent variables and the independent/mediating variable. Then, the significant relationship between the independent/moderating and independent/mediating variable is reported. The following section concludes with the correlations of the main variables with the control variables. The correlations were analyzed for N = 113 respondents.
Except for the relationship between Revenue models and Customer interactions, all dependent service-innovation variables show correlations with each other. This is in line with the findings of Miles (2008) and Den Hertog (2000), which indicated that there is a relationship between the different service-innovation dimensions. The correlation matrix reported the following results:
Service concepts/formulas was strongly and positively related to the other service-innovation variables: Customer interactions (r = 0.49; p < 0.01), Business partners/value systems (r = 0.35; p < 0.01), Revenue models (r = 0.28; p < 0.01), Organizational service delivery systems (r = 0.48; p < 0.01), and Technological service delivery systems (r = 0.51; p < 0.01). Accordingly, Customer interactions showed a strong positive relationship with Business partners/value systems (r = 0.36; p < 0.01), Organizational service delivery systems (r = 0.48; p < 0.01), and Technological service delivery systems (r = 0.51; p < 0.01). Regarding Business partners/value systems, a weaker positive correlation was shown for its relationship with Revenue models (r = 0.23; p < 0.05), and a stronger positive correlation was
found for its relationship with Organizational service delivery systems (r = 0.46; p < 0.01) and with Technological service delivery systems (r = 0.59; p < 0.01). A strong positive relationship was found for Revenue models with the following two dependent service-innovation variables: Organizational service delivery systems (r = 0.29; p < 0.01) and Technological service delivery systems (r = 0.27; p < 0.01). The final correlation found in this research between the dimensions of the dependent service-innovation variables was the correlation between Organizational service delivery systems and Technological service delivery systems, which had a strong positive correlation (r = 0.65; p < 0.01).
With regards to the correlations between the dependent variables and the independent/moderating variable, Perceived information quality, most relationships had a p-value below 0.05. The relationships between Service concepts/formulas and Perceived information quality (r = 0.21; p < 0.05) and between Organizational service delivery systems and Perceived information quality (r = 0.20; p < 0.05) were positive, although slightly weaker compared with the other correlations that had a p-values lower than 0.05. The dependent variables that had a stronger positive correlation with Perceived information quality were the following: Customer interactions (r = 0.28; p < 0.01), Business partners/value systems (r = 0.350; p < 0.01), and Technological service delivery systems (r = 0.36; p < 0.01)
When looking at the relationship between the independent/mediating variable, Ties with KIBS, and the other independent/moderating and dependent variables, a strong positive correlation (with p < 0.01) is shown for all relationships, except for the relationship with Revenue models, which showed no correlation. The relationships that Ties with KIBS had with these variables had the following correlation coefficients: Service concepts/formulas (r = 0.42; p < 0.01), Customer interactions (r = 0.37; p < 0.01), Business partners/value systems (r = 0.35; p < 0.01), Organizational service delivery systems (r = 0.37; p < 0.01), Technological service delivery systems (r = 0.40; p < 0.01), and Perceived information quality (r = 0.504; p <
0.01).
Furthermore, some interesting correlations were found between the dependent variables and the control variables. Service concepts/formulas showed a weak negative correlation with Firm age (r = -0.20; p < 0.05). Customer interactions also showed this weak negative correlation with Firm age (r = -0.24; p < 0.05). Additionally, Customer interactions had a weak negative correlation with Firm type (r = -0.20; p < 0.05). And Perceived information quality reported a negative correlation with the Firm age (r = -0.21; p < 0.05). In terms of relationship between the independent/mediating variable, Ties with KIBS, and the control variables, one correlation was observed (p < 0.05). This was the weak negative correlation with Firm age (r = -0.19; p < 0.05). An overview of all these correlations, as well as additional correlations, is found in Table 8: Correlation Matrix.
Table 8. Correlation Matrix
Normality Analysis
To analyze if the dependent, independent/moderating, and independent/mediating variables were normally distributed, a normality analysis was conducted. The following section describes the mean, trimmed mean, skewness, and kurtosis of each of these variables, as well as the Kolgomorov-Smirnov statistic. The full set of the normality analysis per variable, including the related histograms, is found in Appendix A, Table 20 and in Graphs 1, 2, 3, 4, 5, 6, 7, and 8.
Four of the dependent service-innovation variables – Service concepts/formulas, Customer interactions, Organizational service delivery systems, and Technological service delivery systems – had means ranging from 4.52 to 5.00 and trimmed means from 4.60 to 5.10. The trimmed means have filtered out the 5 percent most extreme scores, and the small difference between the mean and the trimmed mean shows that the extreme scores had only a minor impact on the mean. Furthermore, the dependent service-innovation variable Business partners/value systems had a mean of 3.89 and a trimmed mean of 3.92. And the sixth dependent service-innovation variable, Revenue models, had a mean of 3.35 and a trimmed mean of 3.30. The minor differences between the means and the trimmed means indicate that the extreme scores had no significant impact on the means.
The independent/moderating variable, Perceived information quality, was measured on a 10-point scale and resulted in a mean of 6.58 and a 5% trimmed mean of 6.70. The independent/mediating variable, Ties with KIBS, was measured on a 7-point scale and resulted in a mean of 3.89 and a 5% trimmed mean of 3.91. So, for both of these variables, the extreme cases had a minor impact on the mean. Due to the minor impact of extreme cases on the means of each of these eight variables, it was decided to not to remove extreme cases from the dataset.
When looking at the negative skewness (-1.47) and the positive kurtosis (2.07) of
