Collaborative Supply Chain Performance Measurement Systems

Master Thesis

Collaborative Supply Chain

Performance Measurement

Systems

A multiple case study on the OTD-process of

manufacturing SMEs in the Swedish lighting

industry

Abstract

Title: Collaborative Supply Chain Performance Measurement Systems. A multiple case study on the OTD-process of manufacturing SMEs in the Swedish lighting industry. Authors: Arvid Svensson, Frida Gustafsson and Julien Guillaume.

Background: Even though the multiple stated benefits of Supply Chain Performance Measurement Systems (SCPMS) to enhance the collaboration, there is a lack of research, especially regarding the presence among Small and Medium-sized Enterprises (SME). Closely associated with the presence of a SCPMS are the challenges. Research on SCPMS and the accompanied challenges have been made in multiple fields and contexts. Yet, SCPMS in the Order to Delivery (OTD)-process between a SME lighting manufacturer and their key customers is missing.

Purpose: The purpose of this thesis is to investigate to what level collaboration regarding SCPMS is present in the OTD-process among the cases studied as well as explain this level of collaboration by studying the challenges of SCPMS. The aim is to contribute to existing literature with an explanatory model that highlights the challenges for a highly collaborative SCPMS with their key customers. This model should also give practical contributions to the case companies.

Method: A multiple case study have been conducted. Qualitative data has been gathered through semi-structured interviews.

Findings and Conclusion: The findings show that the level of collaboration in the SCPMS differs from no presence at all to an almost total high presence in the cases studied. The criteria that were found in previous literature for high collaboration in the SCPMS were overall similar to the practices found in the cases. Most of the challenges found in previous research were able to explain the difference in level of collaboration in the SCPMS. The two major challenges were lack of knowledge and lack of will. Lack of knowledge was present where the will to progress was clear, while also in one case, there was not even a will to progress. Overall, the company with low level and will to progress was most likely to perceive challenges, while the company with the highest level perceived the least challenges.

Keywords

Acknowledgement

This master thesis was written during the spring semester 2019 as a final degree project for the master program: Business Process Control and Supply Chain Management at the Linnaeus University in Växjö, Sweden. Writing this thesis and doing research in the field of Performance Measurement was both exciting and educational, but also challenging for all three researchers during times. The outcome of this thesis, which we are proud of, would not have been possible without benefitting from all the help, advice and support which have been given to us during the process.

We would therefore like to thank our examiner, Helena Forslund, for her detailed feedback and availability through both emails and personal meetings, not only during the master thesis but also during our total time at this program. Furthermore, we would like to thank our tutor Hana Hulthén for contributing to this thesis through productive discussions and helpful ideas.

Writing a master thesis of this nature would not have been possible without empirical data. Therefore, we would like to send our deepest gratitude to all participating companies and respondents who have shown a great interest and provided us with interesting input from the area. It was a true pleasure working with you.

During an academic writing process, feedback is important for improvement. We would therefore like to send a thank you to our fellow students who have given us both constructive feedback and interesting discussions during the seminars.

Table of Contents

List of Figures

Figure 1. Study object (own illustration). 4

Figure 2. Outline of thesis (own illustration). 7

Figure 3. Basic Types of Design for Case Studies (adopted from Yin, 2014, p. 50). 12 Figure 4. Population, target population, sample and individual cases (adopted from

Saunders et al. 2016, p. 124). 14

Figure 5. The research onion (adopted from Saunders et al. 2016, p. 124). 24 Figure 6. The order-to-delivery process with its sub-processes (Forslund et al. 2008, p.

44). 27

Figure 7. The conceptual procedural framework for PMS development (adopted from

Gutierrez et al. 2015, p. 3). 32

Figure 8. The RelReg lifecycle (adopted from Maestrini et al. 2018. p. 2031). 33

Figure 9. Research Model (own illustration). 47

Figure 10. Common OTD-process (own illustration). 71

List of Tables

Table 1. Respondents (own illustration) 16

Table 2. Summary table of the methodological choices (own illustration). 24 Table 3. Summary of logistic metrics (own illustration). 28

Table 4. Benefits of SCPMS (own illustration) 29

Table 5. Operationalization of concepts Chapter 3.1 (own illustration). 31 Table 6. Summary of the design phase (own illustration) 35 Table 7. Summary of the implementation phase (own illustration) 36

Table 8. Summary of the use phase (own illustration) 38

Table 9. Operationalization of the steps in the SCPMS framework (own illustration). 39 Table 10. Summary of challenges for the design phase (own illustration) 42 Table 11. Summary of challenges for the implementation phase (own illustration) 43 Table 12. Summary of challenges for the use phase (own illustration) 45 Table 13. Operationalization of the challenges for a high level of collaboration in the

SCPMS presence (own illustration). 45

Table 14. Respondents at Case A (own illustration). 49

Table 15. Level of collaboration and scoring for Case A (own illustration) 52

Table 16. Respondents at Case B (own illustration). 55

Table 17. Level of collaboration and scoring for Case B (own illustration) 56

Table 18. Respondents at Case C (own illustration). 59

Table 19. Level of collaboration and scoring for Case C (own illustration) 63

Table 20. Respondents at Case D (own illustration). 66

Table 21. Level of collaboration and scoring for Case D (own illustration) 69 Table 22. Summary of collaboration levels and score on all cases (own illustration) 72 Table 23. Summary of the challenges for collaboration in the SCPMS for the studied

List of Abbreviations

EDI Electronic Data Interchange

OTD Order to delivery

PM Performance Measurement

PMS Performance Measurement System

RelReg Relationship Regulator

RQ Research Question

SC Supply Chain

SCPMS Supply Chain Performance Measurement System

1 Introduction

In this chapter the background to the subject is presented. The problem discussion highlights the research gaps, then the purpose of the thesis is explained along with presentation of the research questions. Finally, the outline of the thesis is explained.

1.1 Background

By contributing to many job opportunities and supplying products to large companies, Kumar and Singh (2017) discuss that small and medium sized enterprises (SMEs) are generally considered the backbone of the economies in all countries. They mean that SMEs have benefits since they usually are more responsive due to simpler systems and procedures. SMEs are according to Abrahamsson (2017) firms that are independent non-subsidiary companies that are classified based on number of employees and financial status. In their study, Palomero and Chalmeta (2014) refer to the European Commission (2003) when defining SMEs as businesses with less than 250 employees and an annual turnover below 50 million EUR. Abrahamsson (2017) discuss that SMEs are vital for the economic development and growth, contributing to 80% of the economic growth globally. On the Swedish market, Holmström (2018) states that almost all the companies are SMEs, making it an important part of the Swedish economy. In terms of SMEs in Sweden, Persson and Broman (2012) found that SMEs that are collaborating with their partners can play an important role in the development of SMEs in Sweden.

not compulsory. Due to the slight broader view of SC collaboration, this term and definition will be used for the remainder of this thesis.

To achieve a high level of SC collaboration, and to test and reveal the viability of the strategies of the SC, Gunasekaran et al. (2001) highlight that performance measurement and metrics are needed. Performance measurement is according to Neely et al. (1995, p.81) “the process of quantifying the efficiency and effectiveness of action”, while a Performance Measurement System (PMS) is a set of metrics used for quantifying the effectiveness and efficiency of actions. Further, a performance measure is defined as a metric to quantify either the effectiveness, efficiency, or both, of an action. Hereafter, the term metric will be used when discussing a single performance measure. Sena Ferreira et al. (2012) also highlights the importance to measure and evaluate the performance. Maestrini et al. (2017) adds that performance measurement should be done on a large collection of activities, such as logistics and the collaboration with customers and suppliers to fulfill the objective of a SC. Logistics is according to Prajogo and Olhager (2012) the part of SC collaboration that entails coordinating both the material flow and the information flow from the supplier to the customer. This is about delivering the right products, in the right quantity, at the right place and finally at the right time. Forslund et al. (2008) explains in their article that there are different logistics process and that the order to delivery (OTD) process is one of the most important. They go on to state that the process starts with the customer having a need for the product and ends with the customer receiving the order. In this thesis, logistics will be discussed in terms of the OTD-process and the associated metrics.

measurement systems (SCPMS) will be used in this thesis. Maestrini et al. (2018) define SCPMS as a set of metrics that can be used to quantify the effectiveness and efficiency of processes and relationships that occur between companies. Both Forslund and Jonsson (2007) and Papakiriakopoulos and Pramatari (2010) discuss that performance measurement can be expanded to a wider range of activities on top of the actual measurement. Furthermore, previous research introduces multiple methods and processes on how a SCPMS can be present that will be reviewed and merged. Hence, this thesis will employ a broad view of SCPMS, focusing on the overall framework of measuring the OTD-process collaboratively rather than the details of specific metrics and internal measurements alone.

1.2 Problem Discussion

1.2.1 SCPMS in the lighting industry

An emerging area in the SME research is according to Kumar and Singh (2017) regarding their collaboration with suppliers and customers and how they measure the processes. Pešalj et al. (2018) found that most of the empirical studies that have been carried out regarding SCPMS focused on large companies, and that there is a lack of research addressing how this is used in SMEs. Regarding performance measurement in SCs, Otto and Kotzab (2003) identified logistic processes relevant to research. Forslund and Jonsson (2007) suggest that performance should be managed according to dyadic processes. Also, Forslund (2011) states that SCPMS in dyadic relations is an especially interesting composition to study. In terms of dyadic SC processes, this thesis will target the OTD-process, which by Forslund et al. (2008) is highlighted as important.

companies that supply goods to a base of large wholesalers (D-F). Sena Ferreira et al. (2012) discuss hierarchical relations with one dominating partner that controls the other partners in the network. In those cases, other less powerful partners often follow the decisions from the dominating partner. Therefore, this thesis is made from the perspective of the SME manufacturer, targeting the dyadic relationship in the OTD-process with their key customers that are large wholesalers.

In this thesis, a manufacturer is defined as a company that makes a product through a process involving different components (Business Dictionary, 2019b). The wholesalers are defined as companies that buy products from different manufacturers, warehouses them, and resell to retailers (Business Dictionary, 2019a). Hence, the manufacturer is acting as a supplier and the wholesaler is a key customer. The scope of this thesis is visualized in Figure 1.

Figure 1. Study object. Own illustration.

1.2.2 Research questions

and regions, and how they are shared with SC partners. Salam (2017) also advocates for further research regarding SC collaboration that focus on specific company size and industry type. Furthermore, Maestrini et al. (2017) suggest that research should address to what level SCPMS is present in SC dyads. In the literature review, no studies were found regarding to what level SCPMS frameworks are present in the OTD-process of SME manufacturers in their dyadic relation with their key customers in the Swedish lighting industry.

RQ2: What challenges are present at the cases of SME manufacturers in the Swedish lighting industry for a SCPMS in the OTD-process with their key customers?

1.3 Purpose and Research Questions

The purpose of this thesis is to investigate to what level collaboration regarding SCPMS is present in the OTD-process among the cases studied as well as explain this level of collaboration by studying the challenges of SCPMS. The aim is to contribute to existing literature with an explanatory model that highlights the challenges for a highly collaborative SCPMS with their key customers. This model should also give practical contributions to the case companies. To fulfill the purpose, the thesis seeks to answer the following research questions:

RQ1: What is the current level of SCPMS present in the OTD-process between the cases of SME manufacturers and their key customers within the Swedish lighting industry?

1.4 Outline of the thesis

2 Methodology

In this chapter different methodological choices are presented, and the ones used are highlighted and argued for. The choices taken from the research philosophy to the ethical considerations are discussed. At the end, a summary of the methodological choices is provided.

2.1 Research Philosophy

Research philosophy is according to Saunders et al. (2016, p.124) “a system of beliefs and assumptions about the development of knowledge and the nature of that knowledge in relation to research”. Other terms used in method literature is epistemology, ontology and paradigm (Bryman and Bell, 2017). Saunders et al. (2016) state that research is about making a certain field knowledge going forward, either from a big step or from a small step. However, indifferently to the size of the research, assumptions are generated along the way. These assumptions can be categorized into three categories if we refer to Saunders et al. (2016) works: Ontological assumptions are assumptions which are related to the perception and look at the research object and concern. Also, as a result to that, this kind of assumptions will have an impact on the project research choices. Epistemological assumptions stand for human knowledge and its acceptability, validity and legitimacy. Having assumptions of epistemological nature will influence what is considered legitimate or not for the study. Axiological assumptions are referring to the consideration of the research involved participants, researchers and others, and how these participants in a certain measure are affecting the research processes, and thus the resulting outcome. The researchers require then, depending on what are the expectations, to select a research philosophy which is fitting the best to the needs of their study.

Five major philosophies are usually mentioned in literature on business research: positivism, realism, interpretivism, postmodernism, and pragmatism (Saunders et al. 2016).

2. Realism is explained by Bell et al. (2019) as a position which through reality can be understood if the suitable methods are used. According to Saunders et al. (2016), realism can be divided in direct realism and critical realism. Direct realists believe that the perceptions in the senses are real, while a critical realist argue that what is perceived by the senses is just a picture of the reality and not the reality itself.

3. Postmodernism is according to Bell et al. (2019) an approach where the results are viewed as different versions of reality instead of right or wrong. In postmodernism the focus is on the findings rather than on the data collection. According to Saunders et al. (2016), postmodernism seeks to question the general way of thinking. There is no abstract way to describe the world in a way that is right or true. The right or true way is decided by the collective. Hence, they state that the collective way of thinking might not be right, but it is right for a particular collective at a particular time.

4. Interpretivism is stated to be an alternative to positivism. In interpretivism, there is a difference between people and objects, as well as the social researcher need to understand the subjective meaning of social actions (Bell et al. 2019). As per Saunders et al. (2016), the difference is that people assign meanings to the object. 5. Pragmatism states according to Saunders et al. (2016) that concepts only are relevant when they support actions. In pragmatism, the research starts with a problem and the aim is to generate practical solutions for future practice. The typical methods of pragmatism range from mixed methods, quantitative and multiple qualitative. Hence, pragmatism is suitable for a multiple case study.

2.2 Research Strategies

Saunders et al. (2016) mention that a choice must be made on what sort of research design to use. They continue to explain that the research design is dependent on what sort of questions the study seeks to answers. There are four major types of research design. Those being exploratory, descriptive, explanatory and evaluative. In exploratory studies the focus is on open questions and used to answer questions such as “how” and “what”. Descriptive studies aim at describing the area studied. They continue to explain that descriptive studies are often used together with exploratory studies as a foundation to how it is right now to be able to explore. Descriptive studies are used to answer questions such as “how”, “what”, “when”, “where” or “who”. Evaluative studies focus on evaluating the area studied, and answers questions such as “when”, “where”, “which” or “who”. The fourth one, explanatory research focus on cause and effect relationship between different variables and is used to answer questions such as “how” or “what”. Saunders et al. (2016) also mentions a fifth type of research design, which is combined studies. In this two or more of the previously mentioned research designs are used.

Since the research questions in this thesis are aiming to investigate “what is the current level” and “what challenges”, an explanatory research design has been used. The choice of explanatory research design also builds on that the challenges in the second RQ can be the cause of the “current level of collaboration” in the first RQ. That way, the results from RQ1 are explained by the results in RQ2.

that can be unstructured. At last, quantitative data is more precise and objective whereas qualitative data is deeper and can offer a deeper understanding (Bryman, 2018).

Since this thesis focus on words rather than numbers, on the participant’s perceptions instead of the researchers and to have a deeper understanding about the studied subject, and generate theory, a qualitative research strategy has been used. Backman (2016) explains that advantages with a qualitative method are that it is less strict and thereby giving the researcher more flexibility and a less strict framework to adapt to.

2.3 Research Approach

The research approach used is according to Saunders et al. (2016) depending on how and when the theory is collected in the research project. Bryman (2018) explains that there are two common research approaches, those being deductive and inductive. Both Bryman and Bell (2017) and Saunders et al. (2016) add a third approach within business research that has become popular in recent time; the abductive approach. A deductive approach starts off with a theoretical base of knowledge that will guide the research. In a deductive approach hypothesis/concepts are formed based on the theory and later examined empirically. In an inductive approach on the other hand the theory is the result of the empirical study (Bryman, 2018; Bryman and Bell, 2017; Saunders et al. 2016). An abductive approach is explained by Bryman and Bell (2017) as an approach that avoids the limitations that both deductive and inductive approaches can face. The researcher moves back and forth between using an inductive and deductive approach (Saunders et al. 2016). The researcher starts off with a problem and then alternates between theory and empirical study (Bryman and Bell, 2017; Saunders et al. 2016).

The conducted research is based on a deductive approach. This since the theoretical framework was first built up, and then the concepts found in the theory was operationalized for (see Chapter 3). These concepts were used as a base for the gathering and analysis of the empirical material.

2.4 Research Design & Method

study, ethnography, action research, grounded theory and multiple others. Among those, surveys and experiments are designs that are based on statistics and more suitable in quantitative studies (Saunders et al. 2016). Of the qualitative designs, case study research is a common method in the field of business. It is suitable when the purpose is to understand a complex phenomenon. Case-studies enable the researcher to focus on individual cases, while retaining a real-world perspective, as when studying organizational and managerial processes. Since the phenomenon that is researched in this thesis, SCPMS and the challenges of doing it collaboratively, should be considered complex and associated with both organizational and managerial processes, a case study design is used for the research.

A single case study is suitable when the researched phenomenon can be considered rare or specific for the single case studied (Yin, 2014). SC collaboration, PM, SCPMS, and the associated challenges cannot be considered specific for any type of industry nor company. Instead, multiple case studies (two cases or more) can be considered advantageous compared to single case studies. Multiple case studies are not as vulnerable and leads to better analysis since multiple cases are analyzed and can be compared to each other. By comparing cases, the theoretical results are strengthened compared to if findings are done in one single case only. The holistic approach is recommended when the case should be seen as a whole, while the embedded case study is suitable when logical sub-units can be included (Yin, 2014). Based on the recommendations in method literature, this research consist of four individual cases, with holistic data collection where each case was analyzed holistically as a whole.

Bryman (2016) stated that research method refers to the used technique in order to collect the necessary data required for the analysis, and the different administration instruments it can include in order to achieve it. In the case of qualitative case studies, semi-structured interviews are the center part of the research method and design. The use of semi-structured interviews is further discussed in Chapter 2.6.

2.5 Population & Sampling

2.5.1 Population

to the target population. Those conclusions would result in biased or incorrect results (Saunders et al. 2016). The relationship between population, target population, sample and individual cases is visualized in Figure 4.

Figure 4. Population, target population, sample and individual cases. Adopted from Saunders et al. (2016, p. 275).

2.5.2 Sampling

Non-probability sampling can be done in four major ways. Quota sampling, which is mainly used for structured interviews in survey studies. Volunteer sampling which can be divided in snowball sampling where one case is found and helps to find additional cases, and self-selection sampling where each case has a desire to be included in the study. The fourth one is haphazard sampling, where the most common method is convenience sampling. In convenience sampling, cases are chosen because they are easily available. The samples are chosen with no obvious connection to the research question. Fifth is the Purposive sampling, which is based on the judgement of the researcher to choose cases that enables to answer the research questions and fulfill the purpose. To answer the research questions of this thesis, a purposive sampling has been used. The purposive sampling has been used to find the manufacturers that were acting as cases in the thesis. Based on the target population, companies that were considered representative for research questions were found and chosen companies.

Case Respondent Reference Interview method Case A Business area manager A-BA Phone call

Case A Logistics manager A-LM Phone call Case B Sales and customer service B-S Site visit Case B Logistics manager B-LM Site visit Case B Market manager B-MM Phone call Case C Supply chain manager C-SC Site visit Case D Founder D-F Site visit Case D Logistics/Warehouse manager D-L Site visit

Table 1. Respondents. Own illustration.

Saunders et al. (2016) discuss that the number of interviews that needs to be conducted in a non-probability sample depends on the homogeneity of the target population. The larger the homogeneity, the fewer interviews are required. According to Borgès da Silva (2001), the sample size in a qualitative study depends on the length of the interviews and the feasibility of it. For a general study, Saunders et al. (2016) suggest between 5 and 25 semi-structured interviews. Furthermore, Strauss and Corbin (1998) discuss saturation as the point where additional data collection does not add anything new to the purpose, and Borgès Da Silva (2001) states that even the largest studies are not exceeding 50 interviewees.

challenges. Additionally, the data collection on Case C included specific presentations and documents on their metrics and measurements that gave the researchers a clear overall view of Case C.

The results in this thesis are based on case studies made with four companies within the Swedish lighting industry. As stated by both Saunders et al. (2016) and Yin (2014), a case study of this magnitude is too small to serve as a sample of the size that is needed to represent a large population. Hence, it is important to highlight that the results in this study are based on four individual cases. Quantitative research with statistically based methods is needed to generate results that can be considered generalizable for the total lighting industry in Sweden.

2.6 Data Collection Methods

When doing case study research, it is an advantage to use semi-structured in-depth interviews (Bryman, 2016). This research sought to explain to what level a SCPMS is present and understand what challenges that are the reasons for the current level are. Hence, it is important to have a deep understanding on the reasons for the decisions made by the cases studied, which is characteristic for semi-structured interviews.

When doing qualitative research with interviews, Cassell and Symon (2004) state that they should be done with a low level of structure. For this research, semi-structured interviews were considered most suitable. With semi-structured interviews, the respondent had a larger ability to affect the content of the interviews. Semi-structured interviews demanded the researcher to actively follow the answers and to be prepared with follow-up questions on interesting topics to reach depth in the answers (Alvehus, 2013). Compared to structured interviews, where standardized questionnaires are used, and unstructured interviews where no questions are prepared, semi-structured interviews are done with an interview guide with a list of specific themes that should be covered in the interview (Cassell and Symon, 2004; Bryman and Bell, 2017). When developing the interview guide and preparing the data collection, knowledge have been gathered from three different sources in accordance with Cassell and Symon (2004):

2. Personal knowledge: The personal knowledge of the researchers was the basis of the purposive sampling and the formulating of the interview questions.

3. Experience in the field and information gathered from the preparatory work: Information were gathered through preparation interviews with two of the respondents.

As can be seen in Table 1 three of the interviews were done over the phone and five were done through personal visits. Due to the distance the two interviews with Case A were done over the phone, and one of the interviews with Case B was also done over the phone since the person was not at the office during our visit. Bryman (2011) explains that this is a good way of conducting interviews when the geographic distance is long, or to save time. He further explains that the benefits with phone interviews are that they are more time efficient and that the respondent is not affected by the presence of the researcher. But there are also negative aspects with doing phone interviews, that being that the interviewer and respondent cannot see each other’s body language and expressions and might through this miss important signals. This was not considered a problem since both interviews in Case A were done over the phone and the respondents talked about similar topics and validated what the other person had said. Also, the phone interview with the B-MM in Case B validated the answers that we had from the site visit and the respondents from Case B talked about similar things.

2.6.1 Findings Validation

To further increase the quality and the level of the thesis, additional interviews were made in the last step, when the results were done. In accordance with Yin (2014), a review with the respondents validated the findings in the thesis. All cases and respondents were asked to participate in this review. The focus was to have a review from at least one respondent at each case that was considered most suitable. From the eight respondents included in this thesis, the following reviewed the findings:

 Case A. Logistics Manager. A-LM. A-LM was chosen for a follow up interview due to a great overall knowledge and view of the OTD-process in Case A. A-LM is also included in the current work of developing metrics for internal use in the company.

culture. B-MM was chosen due to a large interest in measurement and a clear will for the company progress on that aspect.

 Case C. Supply Chain Manager. C-SC. C-SC is the head of all SC activities in Case C, including the SCPMS and the internal measurement. Hence, C-SC had all the knowledge needed for great follow-up discussions.

 Case D. Founder. D-F. As founder of the company, D-F is included in all processes and in the development of the business. Hence, D-F is responsible for the development of measurements.

The secondary data in terms of the theory used for this thesis were collected through a literature review. In the literature review we focused on peer reviewed articles. When searching for relevant articles, keywords related to the subject of the thesis were used. We have also used some consultancy reports to find background information about the lightning industry since this is not a very well research area. Also, some webpages were used to find up to date information and numbers on the lightning industry.

2.7 Data Analysis Methods

There are multiple methods for doing analysis in qualitative research (Yin, 2014). Since the data used in qualitative research derives from interviews which contain large amount of unstructured text material, it is not clear how it should be analyzed. Compared to quantitative analysis, there are no clear rules on how analysis should be done in qualitative studies (Bryman, 2016). Generally, qualitative analysis includes; examining, categorizing, tabulating, testing and recombining the results to generate empirical findings (Yin, 2014). In terms of case studies, Yin (2014) suggests five possible techniques for data analysis described here after:

1. Pattern matching: Comparing patterns found in the data with patterns that were predicted in advance.

2. Explanation building, which is a type of pattern matching for explanatory studies. Explanation building seeks to build an explanation about the case by using advanced procedures to explain causal links of “how” and “why” something happens.

4. Logic models, which is used for operationalizing a complex chain of events over time to study the change.

5. Cross-case synthesis, which is a technique that is only applicable on multiple case studies. Cross-case synthesis is relevant for studies of two or more cases. In cross-case synthesis, every cross-case is treated as a separate study with the purpose of finding cross-case conclusions. One method is to create word tables that categorize the findings from each individual case. Hence, similarities and differences can be analyzed between the cases (Yin, 2014).

Due to the explanatory research strategy of this thesis, it employs explanation building as a data analysis method. This is mainly due to the nature of the second research question. By explanation building, the challenges found in the second research question could explain the answer from the first research question and enabled for the creation of models. Furthermore, since this is a multiple case study, cross-case synthesis was conducted to enable for a comparison of the individual cases. Hence, the individual cases were categorized, compared in terms of differences and similarities. Then, cross-case conclusions could be made. To compare the cases in terms of to what level a SCPMS is present, the cases were analyzed according to the levels low, medium and high. The criteria for low and high collaboration are argued for in table 8, while the criterion for medium level is added in the analysis. The medium level was needed in some cases where the cases were doing more than the criterion for low collaboration but did not reach the level for high collaboration. To simplify the comparison, the levels low, medium and high are assigned with figures: low = 1; medium = 2; high = 3. By assigning figures to the levels, an overall average score (mean value) on each case could be calculated. That way, the total level of SCPMS presence could be decided and compared among the cases. The score will then be rounded off to the closest level of low, medium or high.

2.8 Research Quality

Two common concepts for research quality is validity and reliability (Bryman and Bell, 2017; Saunders et al. 2016; Yin, 2014). Reliability is about having consistency in the research so that the operations are demonstrated, and the study can be replicated with the same results (Yin, 2014).

The validity is according to Yin (2014) divided in three major parts: internal validity, construct validity and external validity.

1. Internal validity: concerns when research seeks to explain how and why an event leads to another. Hence, the internal validity is critical when doing an explanatory study. It is important to make sure that the event studied leads to the effect studied (Yin, 2014). As in this thesis, it is important to make sure that the challenges perceived by the companies are based on the phase in the SCPMS framework studied, and not some other event. This risk is considered by asking follow-up questions regarding the challenges on the question for each phase in the SCPMS framework, clearly connecting the challenge to the phase. Furthermore, pattern matching in terms of explanation building is done for the analysis, considering the risk of rival explanations.

2. Construct validity is a matter of identifying the right operational metrics for the concepts used in the research. The phenomena studied must be defined in terms of specific concepts, which must be measurable. To ensure construct validity, this research included thorough operationalization of all-important concepts. The concepts have been referred with definitions and criteria for what was considered collaborative or not were added. To further ensure construct validity, the findings and results of this thesis were reviewed with one of the respondents on each case. The respondents had the opportunity to discuss and correct the interpretations made, which also enabled additional findings in accordance with Yin (2014). 3. External validity: tests for knowing whether the results in a study are generalizable

The general objective with reliability is that same findings and conclusions should be made if another researcher performs an identical study again. Qualitative research with semi-structured interviews reflects the reality at the point of the interview. Hence, they are not supposed to be replicable. The value in the qualitative method is the flexibility and ability to explore complex phenomenon. Therefore, it is not reasonable to replicate that kind of non-standardized research (Saunders et al. 2016). Another target of reliability is to reduce the errors and biases in the study (Yin, 2014). Reliability is developed for quantitative studies and can be much more difficult to achieve in qualitative studies. Anyhow, there are some threats to reliability in qualitative research that should be considered:

 Participant error: any factor that might affect the performance of the participant in the interview. E.g. that the participant is in a hurry (Saunders et al. 2016). To avoid participant errors, the interviews in this research were planned long in advance on the premises of the participant.

 Participant bias: factors that can lead to false responses. E.g. if the interview is conducted in an open space, where others can hear the answers (Saunders et al. 2016). The interviews made physically in this thesis were exclusively made in private conference rooms and offices. When doing phone interviews, the researchers were in a private room and made sure that the participant were in a comfortable setting. Hence, actions were made to avoid participant bias.

 Researcher error: factors that affect the researcher’s interpretation of the answers. E.g. lack of preparation or tiredness which leads to misunderstandings (Saunders et al. 2016). To avoid researcher error, the interviews were prepared with interview-guides with multiple follow-up questions for different answers. Furthermore, all three researchers were attending the interviews and following the answers.

 Researcher bias: factors that can lead to bias in the researcher’s recording of responses. E.g. that the researcher’s subjective view is in the way of accurate recording and interpretation of the responses (Saunders et al. 2016). To avoid researcher bias, the researchers had in mind to always stay objective. All three researchers were attending to the interviews. The interviews were recorded and analyzed both jointly and separately among the researchers. Furthermore, the empirical data were reviewed and approved on by the respondents.

2.9 Ethics

When conducting research and gathering empirical material it is very important to make ethical considerations (Bryman, 2018; Saunders et al., 2016). Both Bryman (2018) and Saunders et al. (2016) state that it is important that the researchers explains to the participants what the thesis is about and give the participants the opportunity to ask questions (Saunders et al. 2016). The respondents should also be notified on how the research will be conducted, how the answers will be handled (Bryman, 2018) as well as giving them time to think about if they want to participate or not (Saunders et al., 2016).

Saunders et al. (2016) also pointed out that an agreement between the researchers and the participants needs to be in place before gathering the data about how the recording will take place, if the participants want to be anonymous or not, and how the researchers will ensure the confidentiality of the participants. Sekaran and Bougie (2016) also highlights the confidentiality aspect, by explaining that the researchers can explain to the respondents how the answers from the data gathering will be used.

When the potential respondents were approached the first time, they obtain information about what the thesis is about and how a potential interview would be conducted. They also had the opportunity to ask questions and think about if they wanted to participate or not, which is in accordance with Bryman (2018) and Saunders et al. (2016). Before the interviews took place, the respondents were asked if it was okay if the interviews were recorded, which all of them agreed to. Some of the respondents required anonymity of the company, so all the cases in this thesis will be anonymous, including the respondents.

2.10 Summary of Methodological Choices

Figure 5. The research onion. Adopted from Saunders et al. (2016, p.124).

Methodology Choice Research Philosophy and

Paradigms

Pragmatism

Research Strategies Explanatory research Research Approach Deductive approach

Research Design & Method Qualitative research with semi-structured interviews Population and Sampling Non-probability with self-selection sampling Data Collection Methods Semi-structured interviews

Data Analysis Methods Explanation building and cross-case synthesis

Research Quality Joint interview attendance, separate interview transcriptions, data validation with the respondents

Ethics Protection of interviewees and case companies through anonymity

Table 2. Summary table of the methodological choices. Own illustration.

2.11 Research Working Process

3 Theoretical framework

In this chapter the theoretical framework for this thesis is presented. First, the context of the lighting industry is described. Then, previous literature on SCPMS in the OTD-Process is presented. Thirdly, Frameworks for SCPMS are explained and the one used in this thesis is presented. This part ends with an operationalization table. Then, the challenges that are present in previous research are presented. This part also concludes with an operationalization table.

3.1 The Lighting Industry

The lighting industry is by Gupta (2010) described as one of the most important industries within retailing. According to Candelon et al. (2015), the lighting industry can be defined as an industry composed of three main segments, with a market revenue around 112 billion dollars worldwide. In their study, the market repartition is shared with the general lighting representing 78% of the revenues, the automotive lighting for 19%, and the backlighting representing 3% of the revenues. The composition of lighting companies studied in this thesis is discussed in Chapter 2.5 and is delimited to the general lighting. General lighting consists of both point and linear lamps, consumer/household luminaires and professional luminaires and systems (Candelon et al. 2015).

Gupta (2010) discusses that the lighting industry traditionally has been conservative but that the customer base has become more demanding in the more recent years. Also, the SCs of lighting companies are generally quite similar to other retail SCs, consisting of manufacturers, wholesalers, large and small retailers both as retail chains and individual ones. Particularly in the Swedish lighting industry, there are multiple SME manufacturers that supply products to a small base of large wholesalers (D-F). Compared to many other retail industries, such as fashion clothing or toys, Wong et al. (2006) discuss that the products in the lighting industry have quite a stable demand and longer product lifecycles of 4-6 years.

3.2 SCPMS in the Order to Delivery Process

collaboration as a partnership to reach common goals and gain mutual benefits. To know the success of SC collaboration, Min et al. (2005) highlight the importance of measuring the activities. Their study suggests the development and use of common metrics that can measure the collaborative activities and identify weaknesses that need to be managed in

the collaboration. One way of managing the collaboration can according to Verdecho et

al. (2009) be through the use of a SCPMS. A SCPMS can help the companies to define and collect useful information from the collaboration. As a part of integrating SC processes, Prajogo and Olhager (2012) highlight SC collaboration and that it entails coordinating operational activities such as the material flow from the supplier to the customer. They further mention that the OTD-process is about delivering the right products, in the right quantity, at the right place and finally at the right time. To conclude it they explain that integration in a SC is about collaboration between the functions of the different actors in the OTD-process. As previously stated by Forslund et al. (2008) the OTD-process is one of the most important logistics processes. This process consists of four sub-processes, those being order process (customer), delivery process (internally at the supplier), transportation process (from supplier to the customer) and goods receipt process (customer). The entire order to delivery process is shown in figure 6.

Figure 6. The order‐to‐delivery process with its sub‐processes. Forslund et al. (2008, p.

44)

OTD performance metric: Reference:

Lead-time Forslund (2011); Forslund (2014)

On-time delivery Forslund (2011); Forslund (2014); Theodoras et al. (2005); Gunasekaran et al. (2004); Zaied et al. (2016)

Service level Forslund (2014)

Product availability Theodoras et al. (2005); Zaied et al. (2016)

Quality of delivered goods Theodoras et al. (2005); Gunasekaran et al. (2004); Zaied et al. (2016)

Order completeness Theodoras et al. (2005) Delivery service Forslund (2011) Logistic cost Forslund (2011) Flexibility of service systems to meet

customer needs

Gunasekaran et al. (2004) Effectiveness of invoicing Gunasekaran et al. (2004) Percentage of urgent deliveries Gunasekaran et al. (2004) Number of faultless deliveries noted

invoiced Gunasekaran et al. (2004) Percentage of error free customer

receipts Zaied et al. (2016)

Table 3. Summary of logistic metrics. Own illustration.

3.2.1 Benefits of SCPMS

on different horizons of time that can be achieved from a well-established SCPMS, which is presented in Table 4.

Time

horizon Short-term Medium-term Long-term Benefits Increased planning capability,

improved customer service, shorter order-cycles, reliable delivery, assets utilized, reduced inventory, increased cash flow.

Product variety, effective product life cycles, time to market, reduced overhead cost and increased flexibility.

Increased market share increased human resource capability, increased customer service and reduced overhead costs.

Table 4. Benefits of SCPMS. Own illustration.

Forslund (2011) mentions benefits with using a SCPMS in a dyadic relationship, those include understanding the customers’ needs, tailored service, and the collaboration can lead to a win-win situation for both actors, increased performance ex, service, quality and reduced cost, mutual efforts to solve problems, decrease the bullwhip effect and finally an increased effectiveness.

3.2.2 Defining Levels of Collaboration

SC collaboration can according to previous research be done in multiple ways (Harland et al. 1996; Jahre and Fabbe-Costes, 2005). As mentioned in Chapter 1.2 the aim of this thesis is to study collaboration in the dyadic relations of the case companies. Jahre and Fabbe-Costes (2008) discuss that dyadic collaboration can be done in three different ways:

1. Collaboration between the focal company and its downstream partners, e.g. their customers, which is the collaboration studied in this thesis.

2. Collaboration between the focal company and its upstream partners, e.g. their suppliers.

3. Collaboration with both upstream and downstream partners, done separated.

chain and its collaboration benefits, the companies involved must be linked electronically altogether. The extent of collaboration can according to Danese (2013) be defined according to the degree that the company is exchanging information and the degree to which partnership relations exist with the upstream partner.

Bagchi and Skjoett-Larsen (2003) categorize the level of collaboration as either low, medium or high. Depending on how aligned, what sort of system and to what extent information is shared with actors along the SC determines the level of collaboration. They continue to explain that companies that have none or just a few inter-firm connections and inter-organizational collaboration as well as just sharing a few or none OTD-activities are regarded as having a low level of collaboration. For it to be classified as a high level of collaboration Bagchi and Skjoett-Larsen (2003) continues to explain that the companies need to collaborate on a high level when it comes to OTD-activities, that they are working in teams that stretches outside the company's boundaries, and that they have teams to commonly plan and measure the processes.

When it comes to the level of collaboration in SCPMS, Bagchi and Skjoett-Larsen (2003) classifies a low level of collaboration as, companies that only use metrics in some parts of the OTD-process concerning delivery and inventory, medium level of collaboration as companies that use some common measurements in certain areas, for example lead time, logistics cost and service levels. A high level of collaboration is achieved when measurements concerning processes along the partners are used, and that the performance of these processes are shared with the partner.

To decide the level of the SCPMS in terms of collaboration on the cases in this thesis, criteria are presented in Chapter 3.3.4 and Table 9. Table 9 includes criteria for both low and high level of collaboration.

3.2.3 Operationalization

Concept Definition Operationalization Order to delivery

(OTD) process The OTD-process include the ordering, delivery, transportation and goods receipt between the manufacturer and their customer (Forslund, 2008). The flow of material and information from the manufacturer to the customer. Which is about delivering the right products, in the right quantity, at the right place and finally at the right time (Prajogo and Olhager, 2012).

-The process of receiving, deliver, transport, and final delivery of goods to the customer, including the information flow. Level of

collaborative performance measurement

The level of collaboration depends on to what extent collaborative metrics are used and shared in the SC (Bagchi and Skjoett-Larsen, 2003) and how frequent information is shared (Danese, 2013; Bagchi and Skjoett-Larsen, 2003). High collaboration is considered when information is shared electronically (Forslund, 2015; Archer et al. 2008)

-Metrics are used

collaboratively and shared. -Information is shared on a regular basis with electronic tools if collaboration is high. Metric An actual tool for quantifying the efficiency,

effectiveness, or both, of an action (Neely et al. 1995).

-Actions in the processes are assigned with a metric. Supply Chain

Performance Measurement System (SCPMS)

A set of metrics that can be used to quantify the effectiveness and efficiency of processes and relationships that occur between companies (Maestrini et al. 2018). Including the process of jointly implementing the framework.

-The partners establish a framework with metrics on how to jointly measure the performance of the logistic processes between them.

Table 5. Operationalization of concepts Chapter 3.1. Own Illustration.

3.3 Frameworks for SCPMS

Gutierrez et al. (2015) mention that there are multiple different frameworks for general performance measurement systems (PMS). Bourne et al. (2000) proposes that general PMS development consists of three phases. It starts off with the design phase, which involves starting from the strategy of the company and from that decided what the key objectives to be measured are, and after this to decide on the metrics to be able to measure the key objectives. The next phase is to implement the PMS, which includes to collect, process, analyze and distribute the information needed for continuously using the metrics. The final phase is the use of the performance measurements. This is a common basis for developing a PMS. Gutierrez et al. (2015) present a similar framework, with the addition of assessment (see Figure 7).

Figure 7. The conceptual procedural framework for PMS development. Adopted from Gutierrez et al. (2015, p.3).

Figure 8. The RelReg lifecycle. Adopted from Maestrini et al. (2018, p. 2031).

following three phases in their framework: strategy and vision development, execution and monitoring and output analysis. These two frameworks will be included according to the design, implementation and use phases presented further on.

3.3.1 Design

3.3.1.1 Strategy alignment

According to Bourne et al. (2000) the creation of a framework starts off with the design phase. Salam (2017) mentioned that previous research found that the actors in SC collaboration need to work together to set common strategies and share information regarding performance measurements. Sena Ferreira et al. (2012) discuss that this step is the basis of a SCPMS framework. They mean that actors in the SC collaboration should work together to decide on strategic development and planning, strategic alignment and organization and business models and finally organizational performance and accountability. Furthermore, Maestrini et al. (2018) state the importance of deciding on a common strategy after the decision is made on what other actors to include in the SCPMS. It should start with agreeing on the relationship goals and then include both actors in the dyadic relation through all steps. Also, Forslund and Jonsson (2007) state that the first step is to select the metrics and ensure that the measurements is aligned with the strategy of the dyad. How well the SCPMS suits the collaboration is depending on the characteristics of both actors included. The strength of the collaboration is defined depending on to which degree the strategy is aligned with the measurement and to what extent this is jointly agreed upon.

3.3.1.2 Jointly define metrics and targets

importance of mutually deciding which metrics that should be specified for each target in the strategy.

The steps in the design phase, according to previous literature, are summarized in Table 6. These will be further used for operationalization in Table 9.

Design: Reference:

Strategy alignment Bourne et al. (2000); Forslund and Jonsson (2007); Sena Ferreira et al. (2012); Maestrini et al. (2018)

Define targets Maestrini et al. (2018); Forslund and Jonsson (2007); Simatupang and Sridharan (2005)

Define metrics with

SC partner Forslund and Jonsson (2007); Maestrini et al. (2018); Sena Ferreira et al. (2012); Lohmann et al. (2004); Simatupang and Sridharan (2005) Decide metrics Simatupang and Sridharan (2005); Sena Ferreira et al. (2012)

Table 6. Summary of the design phase. Own illustration. 3.3.2 Implementation

Following previous research, the following phase to design is the implementation phase. Forslund and Jonsson (2007) means that decisions on where in the dyad the measuring should be done must be made. The measurement part consists of multiple questions like how the reports should be generated, in which frequency of time the measuring should be done, how the performance outcome should be presented, and how and by whom the performance feedback should be done. Within this phase Bourne et al. (2000) discuss that information are collected, processed, analyzed and distributed in order to permit the continuous use of the metrics. They mean that meaningful information needs to be extracted from the current systems or that new relevant information must be gathered. Maestrini et al. (2018) further highlight the importance of full visibility in the metrics and that each actor must provide reliable data. Also, the decision on which actor that is responsible for each metric should be made.

relevant information. Either, companies are using data systems with relevant information that need to be reprogrammed to extract meaningful information, or they must start using new procedures.

Min et al. (2005) highlight the importance of information sharing in combination with mutual SCPMS. Brun et al. (2008) found the linkage between collaboration, information sharing and PMS. Their study shows that settling collaborative practices among SMEs can be difficult due to their narrow view. However, they further revealed that when collaborative programs are started, results can be convincing. Also, in some cases the implementation of information sharing through the collaborative programs lead to establishment of a SCPMS. Prajogo and Olhager (2012) discuss that SC collaboration should include connection with information systems which allows the companies to access information in real-time from the other actors in the SC. In SC collaboration, Maestrini et al. (2017) highlight information systems as vital to collect data. It is the information systems that enable information sharing, but Prajogo and Olhager (2012) highlight that the frequency, quantity, and quality of the information is most important.

The steps in the implementation phase, according to previous literature, are summarized in Table 7. These will be further used for operationalization in Table 9.

Implementation: Reference: Data collection (from each

company) Maestrini et al. (2018) Measurement (both companies is

responsible for some metric) Maestrini et al. (2018); Forslund and Jonsson (2007) Reporting the metrics Maestrini et al. (2018); Forslund and Jonsson (2007)

Information sharing Simatupang and Sridharan (2002); Bourne et al. (2000); Min et al. (2005); Brun et al. (2008); Prajogo and Olhager (2012); Maestrini et al. (2017)

Table 7. Summary of the implementation phase. Own illustration. 3.3.3 Use

Furthermore, they state that the incentives must be managed. Simatupang and Sridharan (2002) also highlight the managing of incentives. They discuss three different kinds of incentives. Incentives can be managed in a give-and-take manner, so that the other partner is rewarded during the process of measuring. For example, if the customer shares point of sales data, then the manufacturer can share their delivery schedules in return. The second one mentioned is pay-for-performance, where good performance metric is rewarded in some way. Third is the equitable compensation, which demands open books. In equitable compensation all partners have a fair return on investment according to their risk.

When using the SCPMS, Sena Ferreira et al. (2012) suggest that analysis should be made to determine the outcome of the SCPMS. Forslund and Jonsson (2007) state that it is important to analyze for the continuous improvement of the dyadic relationship and enabling a proactive decision making. They state that the collaboration efforts in analysis can be done through meetings and discussions. The analysis should provide results that leads to continuous improvement. They discuss that the measurement and analysis steps rely on the information that is collected, communicated, and processed about the performance. Also, both Maestrini et al. (2018) and Simatupang and Sridharan (2004) address the need of continuously jointly improving the process with the SC partner.

Sena Ferreira et al. (2012) discuss that the metrics must be reviewed and modified to improve performance. When reviewing, the expectations of the collaboration should according to Simatupang and Sridharan (2004) be compared with the actual outcomes of it.

The steps in the use phase, according to previous literature, are summarized in Table 8. These will be further used for operationalization in Table 9.

Use: Reference: Communicate the

performance Maestrini et al. (2018)

Analyze the outcome Forslund and Jonsson (2007); Sena Ferreira et al. (2012); Simatupang and Sridharan (2004)

Manage incentives Maestrini et al. (2018); Simatupang and Sridharan (2002) Continuously improve the

process Maestrini et al. (2018); Forslund and Jonsson (2007); Simatupang and Sridharan (2004) Align the metrics with the

strategy Guttierez et al. (2015); Maestrini et al. (2018); Lohmann et al. (2004); Simatupang and Sridharan (2002); Sena Ferreira et al. (2012); Min et al. (2005)

Compare the outcomes

with the expectations Simatupang and Sridharan (2004)

Table 8. Summary of the use phase. Own illustration. 3.3.4 Operationalization

Steps in SCPMS Reference Operationalization Criterion for low level of collaboration

Criterion for high level of

collaboration Design:

Strategy

alignment Bourne et al. (2000); Forslund and Jonsson (2007); Sena Ferreira et al. (2012); Maestrini et al. (2018)

-To what level would you say that the business relation has a mutual strategy that considers the needs and wants of both you and your partner? When partners are unable to support common goals because they lack a collaborative strategy (Salam, 2017). Strategy developed with input from both parts in the dyad (Maestrini et al. 2018).

Define targets Maestrini et al. (2018); Forslund and Jonsson (2007);

Simatupang and Sridharan (2005)

-How do you define your targets? -How is the partner included in defining targets?

Targets are not defined mutually with SC partner (Forslund and Jonsson, 2007).

When targets are defined mutually with SC partner (Maestrini et al. 2018; Forslund and Jonsson, 2007). Define metrics with SC partner Forslund and Jonsson (2007); Maestrini et al. (2018); Sena Ferreira et al. (2012); Lohman et al. (2004); Simatupang and Sridharan (2005)

-How do you define your metrics? -How is the partner included in defining metrics? Partner is not included in defining the metrics (Maestrini et al. 2018; Forslund and Jonsson, 2007).

When metrics are defined mutually with SC partner (Maestrini et al. 2018; Forslund and Jonsson, 2007). Decide metrics Simatupang and

Sridharan (2005); Sena Ferreira et al. (2012)

-What actual metrics are present in the OTD-process with your SC partner? -Partners are using the metrics in different ways (Forslund, 2007).

If same metrics are used in the same way and summarized (Lohmann, 2004). Implementation:

Data collection Maestrini et al.

(2018) -How do you collect data? -From what sources do you collect data?

Data is only provided from one of the partners (Maestrini et al. 2018). When data is collected and provided from both companies in the dyad (Maestrini et al. 2018)

Measurement Maestrini et al. (2018); Forslund and Jonsson (2007)

Reporting the

metrics Maestrini et al. (2018); Forslund and Jonsson (2007)

-How do you report the metrics? -To whom are the metrics available? Bad visibility of metrics at partner company (Maestrini et al. (2018).

Metrics are visible for both companies (Maestrini et al. 2018).

Information

sharing Simatupang and Sridharan (2002); Bourne et al. (2000); Min et al. (2005); Brun et al. (2008); Prajogo and Olhager (2012); Maestrini et al. (2017) -How is information about performance in the OTD-process shared? Only parts of information are shared randomly (Forslund, 2015). Information sharing is taking place via Internet is shared at a regular basis (Forslund, 2015).

Use:

Communicate the

performance Maestrini et al. (2018) -How is the performance output communicated? -To whom is the performance output communicated? -How often is the performance output communicated? The performance of the OTD-process is not shared (Bagchi and Skjoett-Larsen, 2003). Performance information is communicated continuously (Bagchi and Skjoett-Larsen, 2003), preferably via Internet tools (Forslund, 2015) Analyze the

outcome Forslund and Jonsson (2007); Sena Ferreira et al. (2012);

Simatupang and Sridharan (2004)

-How is the

Adapt the strategy and update metrics and targets Guttierez et al. (2015); Maestrini et al. (2018); Lohmann et al. (2004); Simatupang and Sridharan (2002); Sena Ferreira et al. (2012)

-What factors is considered when reviewing the metrics?

-How are targets updated? Only company specific factors are considered (Maestrini et al. 2018). The adaptation considers contextual SC factors (Maestrini et al. 2018) and is aligned with the dyadic strategy (Lohmann et al. 2004). Compare the outcomes with the expectations Simatupang and

Sridharan (2004) -Do you have stated expectations of the collaborative measurement? Partner expectations are not compared with the outcomes (Simatupang and Sridharan, 2004). The expectations of both partners are compared with the outcomes

(Simatupang and Sridharan, 2004).

Table 9. Operationalization of the steps in the SCPMS framework. Own Illustration.

3.4 Challenges for collaboration in the SCPMS

When it comes to challenges with collaboration and SCPMS different authors have focused on different aspects. The layout of this chapter is according to the phases in SCPMS, which is design, implementation and use. Some of the challenges can be connected to a certain step in the framework presented in Table 9, whereas others are connected to the phase but might not specifically be connected to a certain step. Also, some of the challenges affect more than one step.

3.4.1 Design challenges

When designing the PMS Papakiriakopoulos and Pramatari (2010) found challenges

most companies focus only on the internal. Barratt and Oliveira (2001) also found that companies mostly focused on internal metrics, and therefore miss out on the external focus. Another challenge concerns the capacity/ability of the company to implement collaborative practices. Panahifar et al. (2015) found that some companies thought that their partners lacked abilities, and lack of knowledge as big challenges. Furthermore, Archer et al. (2008) looked on SMEs and found that lack of resources is a big challenge, which Kumar and Singh (2017) also stated as a finding of their study that SMEs, due to lower resources and innovative capabilities, were facing many constraints. Hudson et al. (2001) studied the PMS in SMEs, finding that the development process demands too much resources and is too strategically oriented, which is in line with the limited resources and more dynamic strategies that are often associated with SMEs. Further, they mean that this leads to specific challenges for SMEs, since the development of PMS need strategic metrics in the long-term. Brun et al. (2008) found that collaboration and information

sharing practices are more challenging to SMEs. Michalski et al. (2018) and Barratt and

Oliveira (2001) also found lack of common objectives as a challenge and Panahifar et al. (2015) mentions lack of shared targets. Forslund (2011) also mentions that there is a risk that companies have different objectives.

The major challenges for the design phase, according to previous literature, are summarized in Table 10. These will be further used for operationalization in Table 13.

Design challenges: Reference:

Affecting processes not included Papakiriakopoulos and Pramatari (2010) Lack of common culture and metrics

selection Busi and Bititci (2006); Forslund (2011) Uneven focus on External and

Internal metrics Busi and Bititci (2006); Barratt and Oliveira (2001) Lack in ability of the company to

implement collaboration practices Brun et al. (2008); Archer et al. (2008); Kumar and Singh (2017); Hudson et al. (2001); Panahifar et al. (2015) Difficulties to set shared objectives

and targets Michalski et al. (2018); Barratt and Oliveira (2001); Forslund (2011); Forslund and Jonsson (2007); Panahifar et al. (2015)

Table 10. Summary of challenges for the design phase. Own illustration.

3.4.2 Implementation challenges

In the implementation phase there can also be some challenges. Papakiriakopoulos and