Measuring supply chain orientation using Computer-Aided Text Analysis (CATA):

(1)

Measuring supply chain orientation using

Computer-Aided Text Analysis (CATA):

Disentangling unique and common effects with market

orientation and digital orientation on firm performance

Master thesis

By

Daniela Felser

Supervisor: dr. C. Schlägel Co-assessor: dr. S. Castaldi University of Groningen

Faculty of Business and Economics International Business & Management d.felser@student.rug.nl

(2)

Abstract

Aim This study follows the research call of Min and Mentzer (2004) to enhance the conception of supply chain orientation. It tries to measure supply chain orientation with the help of text content analysis (CATA) (Short et al., 2010; Belderbos, et al., 2017). Furthermore, this research wants to contribute to Hakala’s (2011) view to further research the interplay of three or more strategic orientations of a firm. Therefore, the relationship among supply chain orientation (SCO), market orientation (MO) and digital orientation (DO), and firm performance is analysed.

Theory Research in supply chain orientation has been rising in the last decades (Chen and Paulraj, 2004). Min and Mentzer (2004) developed the first concept to measure SCO. Their concept include 6 dimensions, namely: credibility, benevolence, commitment, norms, compatibility and top management support. Based on this framework, a SCO dictionary is set out in this document.

Methodology The study conducts a text content analysis (CATA) with 960 shareholder letters from 110 manufacturing S&P 500 firms. Time range is from 2009 until 2017. CATA analysis are conducted with the software CATScanner and Yoshikoder. For statistical analyses this research uses panel regressions in Stata and in a further step communality analysis (software R).

Results Relationships between SCO and MO as well as SCO and DO are positive significant. Correlation analysis suggest that SCO is negative correlated to a firm’s performance measured by Tobin’s Q. Communality analysis show that SCO’s unique effect is very high on firm performance, whereas communality coefficients among SCO, MO and DO are very low or even negative.

Discussion It is of great importance to define supply chain, supply chain management and supply chain orientation and set its research scope. In previous studies, most of the time, 6 to 7 dimensions have been used. Still, boundaries are blurred between supply chain orientation, supply chain management and a supply chain as total. Consequently, supply chain oriented research should revise and standardize its classification of subdimensions.

(3)

3

List of Tables

Table 1 – SCO literature review ... 14

Table 2 - SCO dimensions ... 15

Table 3 - Used variables and data source ... 23

Table 4 – Used keywords: Examples ... 25

Table 5 - Descriptive statistics and pairwise correlation matrix ... 29

Table 6 - Regression models - Tobin's Q ... 32

Table 7 - Regression models – robustness test ROA ... 33

Table 8 - Regression models – robustness test ROS... 34

Table 9 - Regression models – robustness test ROI ... 35

Table 10 - Commonality analysis for firm performance ... 36

Table 11 - Details of the commonality analysis for firm performance ... 37

Table 12 – Descriptive statistics ... 48

Table 13 - Dropped keywords ... 49

List of Figures

Figure 1 - Direct supply chain according to Mentzer et al. (2001) ... 10

Figure 2 - Extended supply chain according to Mentzer et al. (2001) ... 10

Figure 3 - Ultimate supply chain according to Mentzer et al. (2001) ... 10

Figure 4 - SC Scope and SCO according to Tucker (2011, p. 22) ... 16

Figure 5 - Variance partition ... 18

Figure 6 - Total words per firm per SH ... 28

Figure 7 - SCO words per firm per SH ... 28

Figure 8 - Boxplots total words ... 29

(5)

5

List of Abbreviations

SCO Supply chain orientation MO Market orientation DO Digital orientation

SCM Supply chain management SC Supply chain

(6)

1. Introduction

Change in business strategies is omnipresent, driven by phenomena such as internationalisation, globalisation and technological transformation in entrepreneurship (Karna et al., 2016; Onetti et al., 2012; Porter, 1986; St. John et al., 2001). One of the main reasons for the change in strategy is competition between firms (Porter, 1999). For this reason, managers seek to identify strategic advantages to improve firm performance (Barney, 1991; Cano et. al., 2004; Porter, 1987). Looking at the economic developments of entrepreneurship, building up business networks and choosing suitable strategies by adapting constant changes in the economy, has become crucial. For example, Ismail et al. (2017) state that the improvement of competitive positions does not only depend on the technologies they adopt. More importantly, they depend on the strategies built up by the firm’s leaders. Therefore, strategic considerations should be demonstrated (Crook et al., 2008; Karna et al., 2016; Ruokonen and Saarenketo, 2009).

The idea that following a specific strategic orientation has an impact on a firm’s financial performance is not new in existing business management literature (Beutel, 2018; Cano et al., 2004; Deutscher et al., 2016). This can be explained by the resource-based view developed by Barney (1991), in which companies use their resources to gain competitive advantages. However, the isolated consideration of a strategic perspective, however, is problematic, since in most cases companies use several strategic orientations at the same time (Cadogan, 2012; Deutscher et al., 2016; Hakala, 2011). Previous studies have combined market orientation with alternative strategic orientations and found that a firm’s performance is increasing when implementing a strategy besides a market-oriented one (Ellis, 2006; Grinstein, 2008; Kohli and Jaworski, 1990; Narver and Slater, 1990). Consequently, this research includes not yet together analysed concepts: supply chain orientation (SCO), market orientation (MO) and digital orientation (DO).

(7)

7

Third concept in this study is supply chain orientation. In general, supply chain orientation is seen as the implementation of strategic considerations, influencing tactical actions in directing the different streams in a supply chain. This occurs across the functions within a firm as well as within functions of firms in the particular supply chain (Min and Mentzer, 2004). As current research in supply chain management face challenges of conceptual discrepancies, fragmentation in literature and lack of generalization, the main focus of this research is on supply chain orientation. Major scholarly interest within the supply chain stream of literature has been rising for a few decades (Chen and Paulraj, 2004). For example, Ketchen and Hult (2007) stated that for organizational success it is of high importance to effectively manage supply chains. More importantly, the first approach to develop and measure supply chain management related scales and verified a positive relationship to firm performance originate from Min and Mentzer (2004). Until now, there have been some other measurement approaches, investigating mediator or moderator effects of supply chain orientation (Min et al., 2007; Patel et al., 2013). Despite previous studies, there is still a research gap for a comprehensive approach to develop and measure supply chain management construct (Chen and Paulraj, 2004). Min and Mentzer (2004) called for a continuously refinement of supply chain measurement scales in order to strengthen the findings. Hence, the underlying study uses text content analysis to offer a different perspective.

Contrary to existing supply chain orientation studies, which mostly have implemented surveys, this study wants to measure supply chain orientation with text content analysis using shareholder letters for more theoretical precision. The big advantage of text content analysis (CATA) is that it has greater reliability than human coders and captures cognitions and beliefs of individuals written in a quick and efficient way (Short et al., 2018). Shareholder letters are used as they were already analysed determining the relationship between a CEO’s level and area of focus and a firm’s performance (Gamache et al., 2015; Short et al., 2018). Market orientation and digital orientation were analysed with the help of CATA and found to correlate positive with firm performance (Beutel, 2018; McKenny et al. 2018; Zachary et al. 2011). Comparing with market orientation and digital orientation, supply chain orientation is a novelty in text content analysis.

The main research question is thus:

(8)

Additionally, there is a research gap for studies investigating the interplay and effects of three or more orientations simultaneously to combine different strategic orientations into an appropriate mix (Hakala, 2011). Ruokonen and Saarenketo (2009) already emphasized the importance of fit between strategic orientations, which are implemented in a firm. Also, Ketchen et al. (1997) found that tactical activities are well suited to explain performance. Yet, the question if various combinations of strategic orientations, or which combinations, result in superior firm performance, is still unanswered (Deutscher et al., 2016). To examine these relationships between market orientation, supply chain orientation and digital orientation, the following two explorative research questions were carried out:

Exploratory research question 1: What are the unique effects of MO, SCO and DO on firm performance?

Exploratory research question 2: What is the relationship among SCO, MO and DO on firm performance regarding its common effects?

Unique effects denote the variance for each concept or orientation included in particular. Beside the differences in one independent variable (MO, SCO, DO) on firm performance, common effects are analysed. They result from additional value, which may be created for firms, when variables sharing significant joint effects at the same time.

This study aims to contribute to research threefold. First, in respect of research on supply chain orientation, to develop a SCO dictionary to measure a firm’s level of supply chain orientation. Second, it intends to test whether there is a positive relationship between supply chain oriented firms and its performance. Third, this research wants to close a gap within literature analysing three concepts and their unique as well as common effects. Looking beyond methodology, there are also important implications for numerous practitioners. Understanding supply chain orientation and its effect on firm performance is of utmost importance for the manager. They not only better but can also impact their strategic decisions in favour of supply chain orientation, they may also react faster to unintended or unforeseen effects caused by implementation of tactical activities.

(9)

9

supply chain orientation’s unique effect is very high on firm performance, whereas communality coefficients among the three concepts are very low or even negative.

The thesis is organized as follows: At the beginning, the theoretical background including the literature review as well as the conceptual framework is given. Therefore, the core conceptual concepts used in this study are elaborated. Additionally, testable hypotheses are formulated and a conceptual model is drafted. They should answer the abovementioned two explorative research questions. Next, the methodological part of this study includes the intended research design and methods that focus mainly on the literature research, the coding technique as well as the analytical processes. This should answer the main research question. Thereafter, the statistical results are presented with relation to the stated hypotheses. Fourth, the discussion includes the interpretation of the regression and commonality analysis and its implications for researchers and practitioners. Finally, limitations are pointed out including future directions of research.

2. Theoretical background

The present section contains the literature review and the hypotheses development as well as the conceptual framework. The literature review includes definitions of the main terms used in this thesis, an overview of SCO conceptualization approaches and the theoretical background provided for the hypotheses development. The hypotheses development consists of the conceptual framework and five hypotheses concerning the relationship between the orientations and firm performance. As to start, to understand SCO and its scope, important terms are explained in the following.

2.1. Main terms

(10)

example, Mentzer et al. (2001) classify supply chain into three degrees of complexity: (1) direct supply chain, (2) extended supply chain and (3) ultimate supply chain (Figure 1 to 3).

The lowest degree of complexity is direct supply chain as shown in Figure 1. It contains a supplier, a company and a customer. They are all involved in the flows of products (upstream and downstream), finances, services and sometimes information (Mentzer et al., 2001).

Figure 1 - Direct supply chain according to Mentzer et al. (2001)

Figure 2 shows the extended supply chain definition, referring to medium level of complexity. All involved suppliers and customers of the direct customers/suppliers in the various flows of a firm are part of the supply chain. These are upstream and/or downstream flows regarding products, services, finances, and/or information. (Mentzer et al., 2001).

Figure 2 - Extended supply chain according to Mentzer et al. (2001)

The highest degree of complexity is linked to ultimate supply chains, which is pictured in Figure 3. They comprise all upstream as well as downstream flows involved in organizations. In other words, all processes of products, services, finances, and information from the ultimate supplier to the ultimate customer (Mentzer et al., 2001).

Figure 3 - Ultimate supply chain according to Mentzer et al. (2001)

(11)

11

at the same time, broadly diversified. To build a bridge to the term supply chain, for each firm involved in SCM there exists a supply chain. To connect it to SCO, an efficient SCM is the result of successful implementation of SCO. In this study, the definition of Min and Mentzer (2004, p. 63). is used: “the systemic, strategic coordination of the traditional business functions

and the tactics across these business functions within a particular company and across businesses within the supply chain, for the purposes of improving the long-term performance of the individual companies and the supply chain as a whole”. SCM is a group of cooperating

firms with the purpose of improving the effectiveness of their business processes and leveraging their strategic orientation (Mentzer et al., 2001).

There is a difference between the SCM perspectives, which focus on enterprise-wide process flow management, and those, which regard SCM as a management philosophy within an enterprise. The authors differentiate between these two points of view by suggesting that the first, inter-organizational perspective was a more precise conceptualization of SCM, while the latter intra-firm phenomenon was an important forerunner, defined as SCO (Esper et al., 2010). Mentzer et al. (2001) divide supply chain management into three classes: (1) a management philosophy, (2) an implementation of a management philosophy and (3) a set of management processes. First, SCM philosophy extends the idea of integrating functions, i.e. the integration of core business operations, processes and departments, across a firm to all members involved in the supply chain (Cooper and Ellram, 1993). Second, previous research defined several activities necessary in order to realize a SCM philosophy. For example, cooperation between members of supply chains is required for SCM (Ellram and Cooper, 1990). Third, some authors focused on managerial processes instead of activities that present SCM (LaLonde, 1994; Ross, 1997). The supply chain management concept by Min and Mentzer (2004) was the first one developing SCM-related scales. The authors distinguish between SCO and SCM. SCO is a management philosophy, and SCM is the total of all outstanding management actions that have been taken to achieve this philosophy (Mentzer et al., 2001). Distinguishing between SCO and SCM makes sense, since SCO is a kind of prerequisite for SCM. Participants who have an SCO deal with the entire chain management, which clearly differs from the pure control of the processes without the commitment of all participants (Diniz and Fabbe-Costes, 2007).

Supply chain orientation. According to Hakala (2011, p. 199), a strategic orientation is defined as follows: “principles that direct and influence the activities of a firm and generate

the behaviours intended to ensure its viability and performance”. In general, SCO is a common

(12)

supply chain and what standards of conduct are required within the company (Esper et al., 2010; Deshpande and Webster, 1989). In a broader sense, effective supply chain management results from successfully implemented SCO (Dhaigude et al., 2015; Min and Mentzer, 2004). While existing strategic management literature agree on the impact of SCO as a management philosophy, the conceptualization has been represented in diverse ways (e.g. Esper et al., 2010). Mentzer et al. (2001, p. 11) define SCO as: “the recognition by an organization of the systemic,

strategic implications of the tactical activities involved in managing the various flows in a supply chain.” This varies from Min and Mentzer (2004, p. 63), who state that SCO is:” the

implementation by an organization of the systemic, strategic implications of the tactical

activities involved in managing the various flows in a supply chain.” In this thesis, SCO

definition developed by Min and Mentzer (2004) is used.

2.2. An overview of SCO conceptualizations and previous findings

In general, research in this field has been rising (Chen and Paulraj, 2004). Table 1 provides an overview on existing research on supply chain orientation. SCO orientation was defined by Mentzer et al. (2001) as the acknowledge of the SCM philosophy. SCO implies the systematic, strategic combination as well as coordination of tactical business activities. All different flows in a supply chain are affected, for the aim of improving firm performance in the long term (Mentzer et al., 2001). Min and Mentzer (2004) followed the research call of Mentzer et al. (2001). They were the first who approached to develop and measure supply chain management concepts related scales and verified the positive relationship with firm performance. The SCO framework by Min and Mentzer (2004) consists of six dimensions, which are described below in detail and shown in Table 1. Furthermore, there exist some studies investigating mediation and moderator effects of SCO (Min et al., 2007; Miocevic and Crnjak-Karanovic, 2012; Patel et al., 2013). For instance, Min et al. (2007) analysed the connection between market orientation, SCO, SCM and performance and concluded that MO influences performance through SCO.

(13)

13

and competitive orientation as well as supplier, logistics, operational and value-added orientation. The authors claim that supply chain alignment can serve as a strategic capability for a firm. Their definition is in line with the previous view that "a supply chain" orientation is a management philosophy, and supply chain management is the total of all open management actions that have been taken to realize this philosophy. SCO has proven to have a positive impact on supply chain performance.

By doing a SCO literature review, Esper et al. (2010) extended SCO concept conceptually. They integrated strategic and structural perspectives into SCO. In other words, they extended the SCO concept to areas such as organisational design, human resources, information technology and organisational measurement (Dhaigude and Kapoor, 2015). Organizational design describes emphasis on internal cooperation and inclusion. The area human resources include hiring employees with key qualifications in SCM, focusing on employee satisfaction and implementing leadership structures that encourage learning and cross-functional teams. Information technology is the development of IT capabilities that support internal management integration and exchange of information over the entire supply chain. As fourth area of SCO, they added organizational measurement. That is the establishment of a diagnostic and control framework that improves supply chain perspectives, alignment, learning and innovation (Esper et al., 2010). This perspective is a significant extension of the previously predominant perspective of SCO as an intentional construct (Schulze-Ehlers et al., 2014).

(14)

Table 1 – SCO literature review

Author, year Main findings Dimensions

Mentzer et al. (2001)

Definition of SCM and SCO SCO is a management philosophy

SCM is total of overt actions to realize SCO First construct of SCO and SCM.

7 dimensions

Min and Mentzer (2004)

Extended Mentzer et al., 2001 study, refined the SCO definition and proposed the first scale for SCO and tested the scale with a positive outcome.

6 dimensions, 20 items

Hult et al. (2008)

SCO has proven to have a positive impact on supply chain performance.

6 first-order indicators, 4 balanced scorecard outcomes

Esper et al. (2010)

Created a framework of SCO, SCO is a feature of the strategic and structural vision, extended Min and Mentzer’s concept

4 categories added to Min and Mentzer (2004)

Tucker (2011)

Based on Mentzer et al. (2001), Min and Mentzer (2004), SCO was found to positively influence supply chain operational performance, the support of top management was seen as a precursor of SCO

31 SCO related items were reduced to 3 factors

(15)

15

Table 2 - SCO dimensions according to Min and Mentzer (2004, p. 65)

SCO dimensions Definition Source

Trust Trust consists of credibility and benevolence. Determines cooperation and relationship commitment.1

Achrol (1991)1,

Morgan and Hunt (1994) 1

Trust - Credibility

A firms belief that its partner stands by its word2

fulfills promised role obligationsand is sincere3

Anderson and Narus (1990) 2_,

Dwyer and Oh (1987)3_;

Scheer and Stern (1992)3

Trust - Benevolence

Benevolence is a firm’s belief that its partner is interested in the firm’s welfare4

is willing to accept short-term dislocations5

and will not take unexpected actions that would have a negative impact on the firm6

Deutsch (1958) 4_,

Larzelere and Huston (1980)4

Anderson et al. (1987)5

Anderson and Narus (1990)6

Commitment Commitment is “an implicit or explicit pledge of

relational continuity between exchange partners.”7

Dwyer et al. (1987, p. 19)7

Norms Cooperative norms are “the perception of the joint efforts of both the supplier and distributor to achieve mutual and individual goals successfully while refraining from opportunistic actions.”8

Siguaw et al. (1998, p. 102)8

Compatibility Compatible corporate culture and management techniques of each firm in a supply chain are necessary for successful SCM9

Cooper et al. (1997) 9_,

Lambert et al. (1998)9

Top Management Support

Top management support, which includes leadership and commitment to change, is an important

antecedent to SCM, and theabsence10_{of it is a barrier}

to SCM.11

Lambert et al. (1998)10

Loforte (1993)11

(16)

Figure 4 - SC Scope and SCO according to Tucker (2011, p. 22)

According to these definitions, a firm has an SCO if it can demonstrate the impact of managing the upward and downward flows of products and services, information and finance from origin to destination. If a firm sees the systemic and strategic implications in only one direction, then that firm has no SCO. However, this overview shows that, yet, there is no general agreement on which instruments to use to measure SCO (Dhaigude et al., 2015). The text content analysis used in this study is based Min and Mentzer’s (2004) SCO construct.

2.3. Hypotheses and conceptual framework

Barney (1991) bases the core conceptual construct of this research on the resource-based view. In general, the resource-based view dominates the theoretical perspective in strategic management literature. The view focuses on a firm’s internal resources, capabilities and contents, which are seen as the fundamental drivers of performance (Bharadwaj, 2000; Conner, 1991). Resources are defined as “all assets, capabilities, organizational processes, firm attributes, information, knowledge, etc. controlled by a firm” (Barney, 1991, p. 101). According to RBV, the basis for strategic decisions should be the resources and capabilities, not external environmental conditions. The best business strategy is determined by bringing the bundles of tangible and intangible assets in suitable channel according to firm capabilities (Wernerfeldt, 1984; Barney, 1991). As a result, strategic orientations may create knowledge-based resources and are able to form a competitive advantage (Hult, 2005).

(17)

17

based on resources and therefore could be explained by the resource-based view. Greater benefits of firms not only obtain revolutionary levels, but also involve a higher degree of changes in the organization, such as strategic initiatives, performance measurement practices or reporting mechanisms. However, there is a broad gap between the firm leaders’ intention and the realization of initiatives. That is why strategic considerations should be demonstrated (Ismail et al., 2017).

Previous research determined that there is a link between the wording managers use and a firm’s strategic orientation (Barr et al., 1992; D’Aveni and MacMillan, 1990; Short et al., 2010). The managers choice of words calls attention to the area of interest. The attention based view claims that the more frequent and more intense a word is used, the greater is the attention (Abrahamson and Hambrick, 1997; Sapir, 1944). Furthermore, Ocasio (1997) developed the Whorf-Sapir hypothesis, which claims that the executives’ attention leads to firm behavior results. The rising number of SCM measurement attempts (Table 1), and the associated need of a company to be supply chain oriented in order to perform well, may mean that the strategic orientation SCO should be taken more into account (Mentzer et al., 2001; Min and Mentzer, 2004, Hult et al., 2008; Ross, 2013). Min and Mentzer’s six dimensions describe for example the elements of relationships to suppliers, referring to reliability and trust between partners. In general, business relationships to suppliers are entrepreneurial. Aligned with Johansen and Valhne (2003), supplier relationships are important, for example, because of information and knowledge reasons. They also suggested that creating and maintaining partner relationships influence market entry decisions and internationalization. Considering the fact that SCO is positively correlated to firm performance, I hypothesize that:

Hypothesis 1: The SCO of a firm is positively associated with firm performance.

(18)

Exploratory research question 1: What are the unique effects of MO, SCO and DO on firm performance?

On the one hand, unique effects are researched for each included concept. Unique effects denote the variance for each concept or orientation particularly. In other words, the difference in one independent variable (MO, SCO, DO) leads to a certain variance in the dependent variable (firm performance).

Exploratory research question 2: What is the relationship among SCO, MO and DO on firm performance regarding the common effects?

Besides unique effects of strategic orientations on firm performance, common effects are analysed. They result from additional value which may be created for firms, when variables sharing significant joint effects at the same time. In other words, if SCO and DO have a certain direct effect on firm performance, their joint appearance would add the common effect to the amount of both concepts’ unique effects. To make this more vivid, this concept is shown in Figure 5 (Schlägel et al., 2017).

Figure 5 - Variance partition

Variance partition of unique and common effects of strategic orientations on firm performance Total = Total variance explained by strategic orientations

UE1 = unique effects SCO, UE2= unique effects MO, UE3= unique effects DO

CE4 = common effects SCO, MO; CE5 = common effects MO, DO; CE6 = common effects SCO, DO, CE7 = common effects SCO, MO and DO

Addressing the general strategic management question, why do firms perform better than others, I suggest that following a SCO has a unique effect on firm performance. This is based on Christopher (1992), who states that the real battle between firms is not one against the other,

Firm performance

Total

Strategic orientations

UE1 UE2 UE3

CE4 CE5

CE7

CE6

(19)

19

it is more about its supply chain against the other company’s supply chain. That is why I claim that, in the long run, well-maintained and trust-based relationships with suppliers have an influence on a firm’s performance. Furthermore, I suggest mentioning partners, suppliers or retailers in a shareholder letter, has an extraordinary influence on a firm’s competitive advantage. This refers to resource-based view, where it is claimed that resources set up in the right way, lead to competitive advantage (Barney, 1991). Ismail et al. (2017) assume that it is not only about a firm’s resources, but rather which strategy a firm implements to make use of those resources.

The resulting hypothesis is:

Hypothesis 2: SCO has a unique effect on firm performance, which cannot be explained by DO and/or MO.

(20)

Hypothesis 3: There are common effects between SCO and MO on firm performance.

In the following Hypothesis, strategic implications regarding DO and SCO are put together. These two constructs go hand in hand, considering that through digital developments, for example disruptive technologies, an ongoing change is happening in supply (Karna et al., 2016). As a consequence, the need for new functions in a firm or restructuring units become more and more crucial for business success (Ismail et al., 2017). Through technological developments and intentions such as the Internet of Things, firms may feel increasing pressure and expectations from various sides. This may represent the rising intention to follow a digital orientation, as well as to digitalize resources to value creation and capture of the firm (Drnevich and Croson, 2013). Previous studies presented the positive relationship between building up digital resources and firm performance. Leisching et al. (2017) argued that digital business strategy has an impact on the organizational capability to generate market information and improves value creation and capture. Beutel (2018) determined a positive relationship between digital orientation and firm performance. He defines digital orientation as a firm’s strategic orientation to enhance the usage of digital technologies in services and products for the customer. It also includes the aim to increase digitalisation of internal and inter-firm processes as well as infrastructure to gain a competitive advantage. Moreover, developments in supply chain lead to a growing amount of big data in supply chains. Because of that, the demand for enhanced sophisticated solutions for processing is becoming more important. Whether it is about the automation of processes or the usage of artificial intelligence in supply chain area such as demand forecast. Processes are digitalised, and process developers/managers are always looking for simplification of processes. In fact, AI investments have led to benefits for many companies. This can be seen as a further step of realisation of both digital and supply chain orientation (Deloitte Insights, 2018).

As a result, the following hypothesis states:

Hypothesis 4: There are common effects between SCO and DO on firm performance.

(21)

21

SCO does not analyse competitors or look at the competitive environment, but rather concentrates on the supply chain and relationships to suppliers. Johanson and Vahlne (2003) proposed that the creation and reinforcement of foreign partnerships with customers and suppliers is a key factor in determining the type of international entry and expansion. As a third, digital orientation focuses on the technological needs and developments in a firm. This includes products, services and technology the firm decide to provide to the market. Implementing all three concepts together would thus have an impact on firm performance. The argument is based on Hult and Ketchen (2001), stating that strategic orientations of the firms are capabilities that may interact and create competitive advantages together. As previous studies show that one should pay attention to the combined effect of various strategies, I hypothesize the following relationship:

Hypothesis 5: There are common effects among SCO, MO and DO on firm performance).

3. Methodology

Current research in the field of SCO face challenges of conceptual discrepancies, fragmentation in literature and lack of generalization. The underlying study uses text content analysis to offer a different perspective. Annual fundamental data and market data on listed companies is collected from the Wharton Research Data Service (WRDS) Compustat. The findings are complemented by a penal regression analysis, which analyses the relationships of chosen orientations and performance of a firm. Finally, commonality analysis is conducted to find responses to the two exploratory research questions.

3.1. Data sample

(22)

sample. For this reason, the included firms have to be listed in the S&P 500 for the whole study period. Firms that do not meet this requirement are opted out. The final data sample included 960 shareholder letters of 110 manufacturing S&P500 firms.

Variables. Data was derived for each firm by using firm tickers. Table 3 shows all variables used in the data sample. Because of high data availability, Compustat data was used as secondary data. As a final step of CATA (see section 3.2.), its outputs were matched with firm specific financial data derived from WRDS.

Dependent variable. The dependent variable is firm performance. Firm value is compounded by Tobin's Q as this value has found broad acceptance throughout text content analysis research (Belderbos, 2017; Beutel, 2018; Short et al., 2010). Another reason is that market value, Tobin’s q, is able to capture both short-term and long-term prospects. Moreover, it is a frequently used variable to measure a firm’s performance. There are many different options to calculate Tobin’s q. In this thesis, the formula in Table 3 is used because of data availability. For robustness tests ROA, ROS and ROI is used. In Table 3, the calculation of these variables is shown. The data has been matched with Tobins Q as well as the robustness test variables out of the Compustat database via a time lag.

Independent variables. As independent variables SCO for supply chain orientation, MO for market orientation and DO for digital orientation is used. Similar to Belderbos et al. (2017) and Short et al. (2010) ratios of the variables were created to account for differences in the length of the letters.That is the number of CATA output words per variable divided by total words of shareholder letter.

(23)

23 Table 3 - Used variables and data source

Dependent variables Definition Source

Tobin’s Q Logarithm of Tobin’s Q =

(total assets+ (Common Shares outstanding* Price close annual-fiscal) – common ordinary equity-total)/ total assets

Compustat

Tobin’s Q (t+1) One year future value of logarithm of Tobin’s Q Compustat

Dependent variables for robustness tests

ROI Logarithm of net income over total invested capital Compustat ROA Logarithm of net income over total assets Compustat ROS Logarithm of net income loss over total revenue Compustat

Independent variables

SCO number of SCO words divided by total words of SH letter CATA output MO number of MO words divided by total words of SH letter CATA output DO number of DO words divided by total words of SH letter CATA output

Control variables

Size Logarithm of total assets Compustat

R&D intensity R&D expenses divided by total assets Compustat

Age Logarithm of firm age Compustat

Tobin’s Q (t-1) One year lagged value of logarithm of Tobin’s Q Compustat

Industry Clustered standard error according to the first two digits of SIC code Compustat

3.2. CATA analysis

(24)

infrastructure, digital governance, digital processes and digital skills (Beutel, 2018). To measure SCO, the study uses a dictionary developed by its author. The dictionary is based on the supply chain construct of Mentzer et al. (2004). In the next paragraphs further explanation and the development, procedure is described.

SCO Dictionary. To generate a dictionary capturing firms’ SCO a mixture of deductive and inductive approach was used. First, the theoretical background of SCO and SCM was researched. Reading through the articles and examining the core conceptual framework by Min and Mentzer (2004) used in this thesis, the author got an idea of conceptualizing SCO and SCM approaches. Simultaneously, definitions for SCO were analysed. My understanding of a supply chain refers to extended supply chain defined by Mentzer et al. (2001). Theory suggested SCO with various number of subdimensions (Hult et al., 2008; e.g. Tucker, 2011). Based on broadly used Min and Mentzer’s (2004) definition of SCO with its six subdimensions, an exhaustive keyword list to capture each subdimension was created.

The keyword list was validated in the next step. Words were dropped if they were too general for text content analysis and scored too many hits. For instance, ‘trust’ can be used in many different ways. Another reason for dropping keywords was ambiguity, for example, the term: ‘reduce complexity’. As the probability that, for example, the subdimension benevolence appears as a word in any letter, descriptive words or similar words have been included in the dictionary. Moreover, keywords and terms were excluded when they were not used in any letter (see Table 13 in appendix). Additionally words, which could be associated with the orientation, synonyms and words with the same root, were added (Short et al., 2010). Besides a top-down approach to code qualitative data, a bottom-up approach was conducted (Belderbos et al., 2017; Short et al., 2010). Keywords and terms describing and mentioning supply chain orientation were derived from the data. By following an inductive approach, 10% of the shareholder letters were manually coded.

(25)

25 Table 4 – Used keywords: Examples

Used keywords: Examples of Contexts of Deductive and Inductive Keywords

Confidence and trust Last year I said we would create value in the short and long term by improving the performance of our innovative

core, making the right capital allocation decisions, earning respect from society, and continuing to promote an

ownership culture of confidence and trust.

Continuous improvement We have put in place a variety of new growth engines, made continuous

improvements in our operations, and further developed our customer

relationships in ways that will serve ADM and, ultimately, you, our stockholders very well going forward.

Distribution channel We are a leader in manufactured plumbing products, with access to broad

distribution channels worldwide.

Distribution network Our distribution network expanded with the addition of more than 300 new distribution outlets including many new ParkerStore retail locations. Enterprise resource planning We had a difficult year in Japan, particularly in Hips, Knees and Trauma &

Extremities, largely due to issues related to the implementation of an

enterprise resource planning system early in the year.

Partners Part of the work we did in 2016 was re-examining the market to evaluate how we compare with our competitors and better understand the changing needs of our customers and partners.

Partnerships Our long, proud history, our commitment to our social purpose, our global

partnerships and our outstanding portfolio of assets give us a powerful

advantage as we compete in the ever-evolving market for energy. Retailer Most notably, the retail environment continued to be cautious throughout

2011 due to retailers concerns about consumer spending in the uncertain global economic environment.

Six sigma This was a result of our focus on program execution and productivity initiatives driven by Raytheon Six Sigma and operational improvements. Supplier This impacted suppliers, retailers and distributors, some of whom went out of

business and many of whom significantly reduced inventory levels to conserve cash.

Supply chain We have invested to increase our capacity, strengthen our supply chain and improve our systems and processes.

Transform our company To achieve this vision, we are moving from a position of strength to

transform our company for the future.

Workflow we are helping define the future of work and enabling printing beyond paper with new technologies that will disrupt the market and change the way we think about workflows and information processes

(26)

Table 5 Content analysis keyword list for SCO Keyword list for SCO

business improvement process(es), chainwide, channel network, channel relationship, collaborating with our partner(s), commitment to our valued customers, compatibility, confidence and trust, continued trust and confidence, cooperative relationships, coordinated over the supply chain, coordination across all suppliers, create long-term value for all stakeholders, cross-functional team, dealer network, demand forecast , demand planning, distribution capabilities, distribution capability, distribution channel(s), distribution network(s), distributor network, enterprise resource planning, expanded network of channel partners, extending our distribution network, global network of dealers, growing our network of channels, implementation of our transformation plan, implementation of process improvements, improvement in operating, improvement process(es), integrate operations, longlasting and trustworthy relationships with customers, long-term supply chain relationships, loyal and strong dealer network, negotiate a collaboration, network integrity, network or innovation partner, network or innovation partners, organizational structure, organizational structures, partner channels, partnerships collaboration, preserve the trust, process innovation, process optimization, processes reliability, product storage, productivitiy improvement, reduced channel complexity, reduced formal organizational structures, relationship around the globe, reliability of supply, reliable supplier(s), reliable supply, resource allocation(s), resource plan, resource planning, retail distribution, strategic partner(s), strategic partnerships, supplier*, supply chain, supply, system improvement plan, teamwork leadership and commitment, top-to-bottom alignment, transform our business, transform our company, transform our competitive abilities, transportation network, trusted partner, upstream, value for our operations, vendor

CATA. Two software programs were used for computer-aided text analysis. CATscanner outputs show the amount of total words, total characters as well as the amount of hits for the dictionary used. Boxplots and tables of this software analysis are shown in the results section (see section 4.1). Moreover, Yoshikoder was used for context analysis to get a feeling of the usage of the words. This software is able to do conduct concordances and thus, word contexts could be determined. In Table 5, you can find examples of the words added to the dictionary including an example sentence.

3.3. Analytic approach

Statistical analysis. This study is based on panel data, referring to cross-sectional and time series data sample (Batalgi, 1981). In this study the following statistical analyses were made: (1) panel regressions with Tobin’s Q, (2) robustness tests with ROA, ROI and ROS as well as (3) communality analysis.

(27)

27

Before doing a panel regression in Stata, certain assumptions have to be met. First, after Hausman test was run, the model was detected to be fixed effects (Hausmann, 1978). Second, by the use of clustered standard errors in the panel regression, all error terms are uncorrelated and have a constant variance. That is why heteroskedasticity and autocorrelation issues can be neglected (Brooks, 2014).

Communality analysis. Multiple regression is able to explore complex relationships. However, the interpretation of several linear regression results with regard to multicollinearity is a challenge. Communality analysis offers an interpretation level for regression coefficients that cannot be identified by the analysis of structure and function coefficients alone. More precisely, commonality analysis takes the total explained variance (R²) from the set of potential regressions, splits up regression effects into unique and common amount of explained variance assignable to independent variables. Unique effects display how much variance is represented by a particular variable, while common effects explain common variance of groups of variables. (Nimon, 2010; Pedhazur, 1997; Ray-Mukherjee et al., 2014; Thompson, 2006). For this reason, communality analysis is the most appropriate tool to partition variance (R²) (Mood, 1969; Seibold and McPhee, 1979). Based on this ability and in order to answer the explorative research questions, a commonality analysis was carried out as a second step of the analysis. For the execution the statistical program R and the package ‘yhat’ was used (Nimon and Oswald, 2013).

4. Results

In the following section CATA outputs of the SCO dictionary are presented. Furthermore, descriptive statistics as well as a correlation matrix are included. This is followed by the description of hypotheses tests, the results of the Tobin’s Q penal regression as well as robustness test results. At the end of this section, the results of communality analysis are included.

4.1. CATA output

(28)

per year. In this case, most of the firms rarely use SCO words in their shareholder letter. However, some firms can be named which use considerably more SCO words than others: PCAR (372), SHW (331), BA (207), PBI (262), BLL (250), SWK (233), PEP (216), XRX (216), FTI.1 (215), CF (208). The top 10 firms had a total of 2600 SCO counts, which corresponds to about 20% of all SCO words counted. In comparison, the top 10 firms with the lowest SCO score are as follows: NOC 26; WY 26; ABT 34; LEG 36; ROP 39; TXN 39; GILD 43; BCR 44; PKI 45; BF.B 46. They have a total of 378 SCO words counted, representing roughly 3% of the total number of SCO words.

Figure 6 - Total words per firm per SH

Figure 7 - SCO words per firm per SH

Figure 8 and 9 show boxplots referring to the CATA SCO output. Figure 8 describes the number of total words. Interestingly, the length of the letters to shareholders remains almost the same during this period. There are some outliers, however. Looking at Figure 9, the words mentioned in the letters to shareholders have risen slightly over the past four years (2014-2017).

Total Words

(29)

29

Figure 8 - Boxplots total words Figure 9 - Boxplots SCO

4.2. Descriptive statistics and pairwise correlation matrix

Table 5 presents arithmetic means, standard deviations, and pairwise correlations between dependent, control and independent variables.

Table 5 - Descriptive statistics and pairwise correlation matrix

Variables Count Mean SD 1 2 3 4 5 6 7 8 9 10

1 Tobin's Q 1301 .70 .40 2 Tobin's Q (t-1) 1203 .69 .39 -.99* 3 ROA 1301 .08 .06 -.12* -.63* 4 ROI 1301 .14 .11 -.11* -.62* -.55* 5 ROS 1301 .11 .10 -.03* -.46* -.77* -.54* 6 SCO 1320 .02 .01 -.11* -.32* -.15* -.04* -.05* 7 MO 1320 .04 .01 -.00* -.20* -.07* -.04* -.00* -.63* 8 DO 1320 .01 .01 -.09* -.09* -.26* -.14* -.15* -.37* -.38* 9 Size 1301 9.66 1.11 -.16* -.23* -.16* -.08* -.02 -.13* -.13* -.04* 10 R&D Int. 1301 .03 .04 -.17* -.24* -.11* -.20* -.04 -.11* .05 -.05* .78* 11 Age 1307 3.82 .81 -.12* -.11* -.07* -.13* -.06 -.05 -.07* -.12* -.40* .26* *_{p < 0.05}

In the correlation matrix (Table 5), ratios of SCO, MO and DO are used. Time range is from 2009 until 2017. Positive coefficients would indicate a rise of future effective firm value when SCO is enhanced. Contrary, negative coefficients would be linked to a decrease in future firm value when supply chain oriented implications are rising. SCO correlates negatively significant with Tobin’s Q (-.11), Tobin’s Q(t-1) (-.32) as well as ROA (-.15). Further above I suggested

(30)

Tobin’s Q and ROA. Consequently, the correlation analysis rejects hypothesis 1 saying that SCO is positively linked to firm performance.

Tobin’s Q correlates positive with ROA, ROI and ROS (.63, .62, .46, respectively). The remaining relationships among each of the four variables are also linked positively significant. This finding can be linked to the fact that all four variables describe firm performance. In other words, if one of the four variable increases, the other one would rise as well. The highest positive significant correlation exists between SCO and MO (.63). This is in line with Grinstein (2008), suggesting that firms perform better when not only active in market-oriented strategies. Therefore, hypothesis 3, SCO and MO have a positive relationship with firm performance, is supported. Moreover, SCO correlate positively significant with DO (.37). Generally, DO exhibits only positive coefficients. More precisely, DO correlates positively significant with all variables except Tobin’s Q(t-1). The highest one refers to SCO as mentioned above, followed

(31)

31 4.3. Hypotheses tests

Table 6 reports the empirical results of Tobin’s Q panel regression models. The regressions include time fixed effects and control for firm-level characteristics. Standard errors are shown in parentheses. Model 1 to 3 follow the same time range set up, from 2009 to 2017. The first model includes only the control variables. In Model 2, the ratio of SCO is added. Model 3 covers all three ratios of SCO, MO and DO. Model 4 includes the control variables, all strategies and the logarithm of Tobin’s Q(t-1) in the time range from 2008 until 2017. Instead

of logarithm of Tobin’s Q (Model 1 to 4), Tobin’s Q (t+1) is used as dependent variable in Model

5 and 6. Both models cover the time range from 2009 until 2018. Model 5 includes the control variables and the three ratios. Additionally to this, Model 6 includes Tobin’s Q as independent variable.

As adjusted R2 increases the more control variables are included, there is a gradient apparent from Model 1 to 3. Adjusted R2 is the highest in Model 4 (.63), which indicates that 63% of the variation in Tobin’s Q (y) is due to variation in SCO (x). Model 6 shows the second highest adjusted R2 value (.52). Looking at the control variables, we can see that size correlates negative significant in all 6 Models. Age is negative as well, but only slightly and not significant. R&D intensity correlates slightly negative with the dependent variable, only in Model 6 significant negative (-1.09). Number of observations is the highest in Model 4 (1180). In Model 5 and 6 the number of observations is 962, respectively.

(32)

Table 6 - Regression models - Tobin's Q

(1) (2) (3) (4) (5) (6)

Variable Tobin’s Q Tobin’s Q Tobin’s Q Tobin’s Q Tobin’s Q (t+1) Tobin’s Q (t+1)

Age -.09 -.09 -.10 .09 -.04 .01 (.16) (.16) (.15) (.08) (.08) (.07) Size -.37*** _-.37*** _-.38*** _-.26*** _-.31*** _-.09* (.05) (.05) (.05) (.05) (.07) (.05) R&D Intensity -.10 -.05 -.18 -.52 -1.20 -1.09** (1.96) (1.93) (1.98) (1.15) (1.04) (.39) Tobin's Q .59*** (.03) Tobin's Q (t-1) .58*** (.03) SCO 1.88 2.13 2.00** _3.35*** _2.06** (1.45) (1.56) (.79) (1.11) (.80) MO -.56 -.28 1.48* _1.80*** (.97) (.46) (.73) (.58) DO 3.53 2.15* _4.35* _2.25 (2.29) (1.04) (2.37) (1.84) Time period 09-17 09-17 09-17 08-17 09-18 09-18 N 974 974 974 1180 962 962 Adj. R² .39 .39 .40 .63 .31 .52

Standard errors in parentheses

(33)

33 4.4. Robustness tests

For robustness purposes the entire regression is conducted with ROA (Table 7), and ROS (Table 8) and ROI (Table 9). This shall clarify whether the results of Tobin’s Q are representative. The results in Table 7, 8 and 9 show that the indifferences of the signs between SCO correlations and firm performance have disappeared. Whether for SCO, nor for MO any significance is shown in all three robustness tests. The relationship between DO and Tobin’s Q shows statistical significance in all three robustness tests.

Table 7 - Regression models – robustness test ROA

(1) (2) (3) (4) (5) (6)

Age -.28 -.28 -.27 -.22 -.37 -.25 (.26) (.26) (.26) (.19) (.22) (.21) Size -.46*** _-.46*** _-.48*** _-.41*** _-.17 _-.07 (.15) (.15) (.14) (.12) (.13) (.11) R&D Intensity 3.19 3.20 2.94 2.54 4.79 5.72* (2.31) (2.33) (2.25) (1.52) (3.72) (2.70) Tobin's Q .13** (.05) Tobin's Q (t-1) .17*** (.04) SCO .60 -.26 -.36 -3.54 -4.49 (3.35) (4.22) (3.60) (4.27) (4.40) MO 1.01 .75 2.36 2.39 (2.58) (2.44) (1.43) (1.55) DO 9.45** _8.61** _.86 _.16 (4.14) (3.86) (2.97) (3.39) Time period 09-17 09-17 09-17 08-17 09-18 09-18 N 936 936 936 1093 931 906 Adj. R² .08 .07 .08 .12 .04 .06

*_{p < 0.1,}**_{p < 0.05,}***_{p < 0.01}

(34)

Table 8 - Regression models – robustness test ROS

(1) (2) (3) (4) (5) (6)

Age -.35 -.35 -.35 -.31 -.39 -.30 (.25) (.25) (.25) (.18) (.23) (.22) Size .04 .04 .02 .08 .10 .12 (.17) (.17) (.17) (.17) (.18) (.15) R&D Intensity 1.08 1.10 .83 1.19 5.10 6.26** (1.24) (1.23) (1.15) (1.45) (3.80) (2.85) Tobin's Q .11** (.04) Tobin's Q (t-1) .14*** (.04) SCO .93 .80 .86 -2.74 -3.77 (2.71) (3.59) (3.16) (3.32) (3.32) MO -.13 -.59 .69 .78 (2.31) (2.24) (1.36) (1.59) DO 8.39* _7.98* _-1.71 _-1.74 (4.26) (4.20) (3.05) (3.50) Time period 09-17 09-17 09-17 08-17 09-18 09-18 N 936 936 936 1093 931 906 Adj. R² .01 .01 .02 .03 .01 .02

*_{p < 0.1,}**_{p < 0.05,}***_{p < 0.01}

In Table 8, regression results for ROS are described. In Model 3 and 4, DO is significant (Model 3: 8.39, Model 4: 7.98). SCO correlates negative with Tobin(t+1) in Model 5 (-2.74) and Model

(35)

35 Table 9 - Regression models – robustness test ROI

(1) (2) (3) (4) (5) (6)

Age -.29 -.30 -.29 -.31 -.45 -.31 (.30) (.30) (.29) (.24) (.32) (.28) Size -.33 -.33 -.36* _-.27 _-.07 _.03 (.19) (.19) (.19) (.16) (.18) (.13) R&D Intensity 5.47** _5.50** _5.18** _4.21** _6.47 _6.81** (2.14) (2.15) (2.06) (1.53) (3.81) (2.72) Tobin's Q .18*** (.05) Tobin's Q (t-1) .21*** (.05) SCO 1.47 .72 .65 -.62 -1.84 (3.19) (4.23) (3.84) (3.83) (4.12) MO .78 .37 1.17 1.23 (2.85) (2.67) (1.20) (1.28) DO 11.04** _9.82** _-1.10 _-2.72 (4.48) (4.11) (3.68) (3.96) Time period 09-17 09-17 09-17 08-17 09-18 09-18 N 936 936 936 1093 931 906 Adj. R² .08 .07 .08 .13 .06 .08

*_{p < 0.1,}**_{p < 0.05,}***_{p < 0.01}

The independent variable DO shows significance in Model 3 (11.04) and Model 4 (9.82) for ROI (Table 9). SCO shows positive correlations with Tobin’s Q in Model 2 (1.47), Model 3 (.72) and Model 4 (.65). Yet, the coefficient is not significant.

4.5. Communality analysis

(36)

suppressor predictor variables (Kraha et al, 2012). SCO, MO and DO variable show small unique effects, 7.3%, 2.6% and .8%, respectively. Contrary to expectation and previous studies using communality analysis, the common effects of this work are slightly negative (SCO: 3.7%; MO: 2.6%; DO: .5%).

In Table 11, the proportions of variance uniquely associated with SCO, MO and DO are pictured. The right side of the column gives the percentage of explanatory power of the particular unique or common effect in relation to the explained total variance.The amount of SCO variance is more than any other partition, representing 94.8%. With this result, hypothesis 2, SCO has unique effects on firm performance, can be supported. The only positive common effect analysed is MO and DO (8.5%), which matches with its correlation relationship. Regressing SCO with MO and DO resulted in negative commonality coefficients. Besides two out of three negative commonality coefficients in second order communalities, the third-order effect is negative as well (-9.1%). Reasons for negative communality coefficients is so called suppression or predictors which influence each other in the opposite way (Pedhazur, 1997). This indicates that for example SCO confounds the power of MO and vice versa. Some researchers claim that negative communalities equate to zero (Frederick 1999). Others argue that adding the confounding variable would indicate its magnitude (i.e., variance explained) (Capraro and Capraro, 2001). As a consequence, there is no evidence for hypothesis 5, which encompasses all three concepts and is therefore negligible in this horizon. Yet, beta weights of communality analysis match with pairwise correlations values regarding the signs.

Table 10 - Commonality analysis for firm performance

Predictor Beta r rs rs² Unique Common

SCO -.38 -.19 -.68 .467 .07 -.04

MO .23 .00 .00 .00 .03 -.03

DO .10 .06 .22 .05 .01 -.01

Total .51 .11 -.07

(37)

37

Table 11 - Details of the commonality analysis for firm performance

Predictor set Communality %

Unique to SCO .07 .95 Unique to MO .03 .34 Unique to DO .01 .11 Common to SCO, MO -.03 -.33 Common to SCO, DO -.00 -.06 Common to MO, DO .01 .09 Common to SCO, MO, DO -.01 -.09

Total .08 1,00

5. Discussion

Before discussing the findings of the analysis of previous chapters, keep in mind the purpose of this study: First, the aim was to measure SCO by using CATA. Second, to test whether there is a positive relationship between supply chain-oriented strategy and firm performance. Third, to proof empirically if strategic orientations used in this study have unique and/ or common effects. Generally, findings suggest that it is of great importance to define SC, SCM and SCO and set its research scope. As mentioned before, the measurement of supply chain is problematic as there are no conventional measurement devices such as income statements or balance sheets for individual businesses (Bowersox et al., 2002). Therefore, it is problematic to find applicable scope of SCO for study purposes.

(38)

other. Moreover, SCO correlates positively significant with DO. This supports hypothesis 4, which indicates that changing industry environments through digitalisation and ongoing change in supply chains (Karna et al., 2016) should be considered in a firm strategy. Hypothesis 5 can be neglected, as the joint effects are minimal or negative among the three concepts. As an explanation, one could argue that that the conceptual approaches that combine organizational orientation and business performance are more complex than a direct link (Lonial and Carter, 2015). Greer and Ford (2009, p. 51) argued that: “The numerous

boundaries that span organizations in supply chain relationships make it difficult to develop and transfer the proper behaviours necessary for effective change.” Moreover, supply chain

relationships are costly in maintenance (Burt, 1992; Oviatt and McDougall, 2005). The same applies to the implementation of MO (Kohli and Jaworski, 1990) and DO (Mithas et al., 2013). As a result, each firm must be careful in maintaining relationships with its partners, as it is cost-intensive and time-consuming.

These findings are in line with Esper et al. (2010). They argue that SCO implies a high level of alignment between the strategy and the structure of the firm. Fit is decisive for business success (Chandler, 1962; Rumelt, 1974). The strategic perspective as such is not sufficient, as it demands structural assistance for successful implementation. SCO is an internal business strategy in which companies operate in the global market in a supply chain environment. To operate efficiently in this environment, adequate strategic and structural fit is required (Esper et al. 2010). To link a negative influence of SCO on firm performance with negative correlation between SCO and firm performance, the following sentence from Stryker corporation’s shareholder letter (2014, p. 5) is outlined: “We had a difficult year in Japan, particularly in

Hips, Knees and Trauma & Extremities, largely due to issues related to the implementation of an enterprise resource planning system early in the year”. This may indicate that an

asymmetry between supply chain changes and a firm’s general strategy may have negative influence on its performance. Probably SCO needs more time to flourish and show positive influences on performance. As a consequence, SCO may be a long-term focused, well-considered strategic orientation.

5.1. Implications for theory and research

(39)

39

of supply chain clearly and to point out, how broad or narrow SCO is seen. As there are several approaches to capture SCO and definitions of SCM has varied from author to author. Therefore, this study adds to existing theory by providing more insights into supply chain, its strategic orientation and in a further step, its impact on firm performance.

Second, this research has shown that the determinants examined are partly positively linked to firm performance. Supply chain oriented and market oriented firms tend to have superior firm performance, followed by supply chain oriented and digital oriented firms. Hence, the focus should be on newly examined strategic orientations related to firm performance, as they might act as important determinants of firm performance (Deutscher et al., 2016; Hakala, 2011). Hence, the combination of SCO with alternative orientations might be investigated in further in theory. The underlying study could be used as a basis for further text content analysis conceptualizing supply chain.

5.2. Implications for practice

Generally, the results are also an important take-away for managers. Implementing SCO successfully, meaning the strategic orientation aligns with the structure of the firm, leads to effective SCM, and in a further step to higher firm performance. It is essential for a firm to have reliable suppliers who guarantees on-time delivery and maximum product quality. For this reason among others, a strategy should be pursued to maintain and develop these relationships. The first implication of this finding for SCO practitioners is to understand that managing a supply chain is an on-going continuous process, not a discrete project or program.

Because of disruptive technologies, for example, an ongoing change is happening in supply (Karna et al. 2016). In addition, words or terms used in shareholder letters are changing over time. That is why continuous monitoring of developments is necessary.

(40)

Third, an important implication for practicioners is that strategic orientations should not be seen as sole drivers, but rather managers should engage in the interactions of different types of strategic orientations. So far, not enough attention has been paid to the links and combinations of key drivers along the entire supply chain, nor to their alignment with the competitive conditions of an industry and the unique value proposition of each organisation. In other words, it is less about the resources a company owns that contribute to its success than about how the company uses those resources. The outcome of the two-fold analysis has proven that the isolated approach of focusing on a single strategic direction has limited impact on business performance.

5.3. Limitations and future direction

Doing text content analysis with shareholder letters brings along some limitations. First, some limitations regarding the data sample selection. The data source shareholder letters vary in length, design and availability. Firms publish the letters voluntarily, that is why in some cases they were not available. This in turn means that we only know about firms who publish their shareholder letters. It could be possible that SCO plays a different role in smaller firms, rather than S&P 500. Moreover, we do not know about other sectors as we were looking at manufacturing S&P 500 firms. As the data sample only contains US companies, this study faces national bias. We do not know about international perspectives. Future research could overcome this limitation by including shareholder letters from different size, industries and/or countries.

Measuring supply chain orientation using Computer-Aided Text Analysis (CATA):