• No results found

THE ROLE OF CONTINUOUS IMPROVEMENT ON LEAN MANUFACTURING IMPLEMENTATIONS: A CASE STUDY

N/A
N/A
Protected

Academic year: 2021

Share "THE ROLE OF CONTINUOUS IMPROVEMENT ON LEAN MANUFACTURING IMPLEMENTATIONS: A CASE STUDY"

Copied!
49
0
0

Bezig met laden.... (Bekijk nu de volledige tekst)

Hele tekst

(1)

1

THE ROLE OF CONTINUOUS IMPROVEMENT ON

LEAN MANUFACTURING IMPLEMENTATIONS:

A CASE STUDY

Author: S.H. de Boer

S2536250

MSc Thesis Technology & Operations Management

University of Groningen Faculty of Economics & Business Nettelbosje 2, 9747AE Groningen

The Netherlands

In cooperation with

(2)

2

ABSTRACT

(3)

3

TABLE OF CONTENTS

PREFACE ... 4

INTRODUCTION ... 5

THEORETICAL BACKGROUND ... 6

Continuous Improvement (CI) ... 6

Lean Manufacturing (LM) ... 7

Supply Chain Management (SCM) ... 8

Just-in-time (JIT) ... 9

Total Quality Management (TQM) ... 9

Human Resource Management (HRM) ... 9

Hypothesis development and conceptual model ... 11

METHODOLOGY ... 13

Case selection and description ... 13

Data collection and measures ... 13

Data analysis... 15

Case comparison methods ... 16

Validity and reliability ... 18

RESULTS ... 20

CI maturity and LM performance... 20

LM performance and operational performance ... 22

CI maturity and operational performance ... 23

CI maturity and LM importance consensus ... 24

LM importance consensus and operational performance ... 25

DISCUSSION ... 26

CI maturity and LM performance... 26

LM performance and operational performance ... 26

CI maturity and operational performance ... 27

CI maturity and LM importance consensus ... 27

LM importance consensus and operational performance ... 27

Limitations ... 28

CONCLUSION ... 29

Managerial implications ... 29

Suggestions for further research ... 30

REFERENCES ... 31

APPENDICES ... 34

(4)

4

PREFACE

The final part before graduation consisted of writing my Master thesis, my own contribution to the scientific world. I was frequently warned for the iterative characteristics of conducting research. By repeating, reconsidering, rewriting and rethinking every step along the way, I was caught up in the circle. Sometimes fearful of never breaking out, I realized it’s part of the process. Every iteration brings one closer to the end goal and towards the desired result.

During this process I’ve had the privilege to receive feedback and discuss preliminary results with a few people, which I would like to thank by this way. Jannes Slomp, my first supervisor and the initiator of the project who provided me with the handles to safely start my thesis. His knowledge within the research field proved to be valuable for increasing the usefulness of this thesis. Nick Ziengs, my second supervisor and a valuable source for suggestions on research methodology. By jointly brainstorming possibilities he provided me with new insights and motivation. Wilfred Knol, a colleague of Mr. Slomp who provided me crucial data and helped me with the analysis necessary to establish results. Frank Vlaming, always available for a quick second opinion on every aspect of my thesis while writing his own. Many coffee breaks together pulled us through the sometimes long days at campus. Without a doubt, these four people contributed fairly to the quality of the report.

I’m looking back on an interesting and educational period writing my thesis. With my confidence varying from time to time along the way, I can safely say I’m satisfied with the result.

(5)

5

INTRODUCTION

Continuous improvement is a phenomenon that can be considered very important in achieving business excellence for many years (Sanchez & Blanco 2014; Singh & Singh 2015). It can be defined as “a culture of sustained improvement aimed at eliminating waste in all organizational activities systems and processes, and involving all organizational participants” (Bhuiyan et al. 2006:671). Continuous improvement can offer serious competitive advantage as the patterns and behavior are deeply embedded in the organization, making it hard to imitate (Bessant,1999) because changing culture requires commitment and time (Kwak & Anbari 2006).

Though CI is linked with different operational techniques as lean and TQM, there is no agreement on the exact approaches used to achieve operational excellence with continuous improvement (Singh & Singh 2015). This indicates an interesting research area. A substantial amount of literature does however confirm a strong effect of international and organizational culture aspects on the implementation of lean manufacturing (LM) and its performance (Kull & Wacker 2010; Power et al. 2010; Marodin & Saurin 2013). There is however no linkage made between continuous improvement and the performance of LM, which is surprising as it can be considered a basic pillar of LM (Sanchez & Blanco 2014) and the two are considered to go “hand in hand” (Singh & Sing 2015:111). As Pakdil and Leonard (2015) confirmed, linking and measuring culture to LM performance would be a useful step forward in the operationalization of LM.

This study will contribute to the literature by researching the interaction of CI maturity and LM performance, filling the literature gap concerning culture and production principles and its performance. As Lam et al. (2015) point out, there is little known about effects of the involvement of employees contributing to elements of LM. Given the vitality of continuous improvement in recent businesses (de Leede & Kees Looise 1999; Sanchez & Blanco 2014), the desire for such understanding can be considered high. By performing a comparative case study regarding CI and LM performance and the operational performance in 20 SME’s in the Netherlands, conclusions can be drawn whether it is expected that CI does contribute to higher operational performance. Understanding the relationship between the concepts will allow companies to develop a culture that will enhance the LM implementation (Pakdil & Leonard 2015), and eventually contribute to a more complete understanding of lean manufacturing.

(6)

6

THEORETICAL BACKGROUND

The theoretical background will be used to provide the empirical framework and setting of the study. Existing knowledge will be described and explained to identify a possible literature gap. Firstly the concept of CI will be described, followed by LM. LM will thereafter be split into bundles to be explained in more detail.

Continuous Improvement (CI)

(7)

7

lies on increasing autonomy and empowerment of individuals. The final level contains full CI capability and can be considered a ‘learning organization’ (Bessant 1999; Bessant et al. 2001).

It is important to distinguish CI from lean manufacturing, which in some contexts can also be considered continuous improvement. Therefore in this study, CI will be viewed as the earlier discussed dynamic capability rather than a set of practices. To further emphasize the distinction between CI and LM in this study to avoid confusion regarding the CI topic, a line is drawn between practical implementations and a more ‘behavioral aspect’ that involves learning and employee attitude. The former is being located to the LM implementations and the latter being part of CI.

However, as Singh & Singh (2015) pointed out, continuous improvement and operational practices are used together, indicating that the cooperation between LM and continuous improvement could jointly enhance performance. By combining the organizational learning aspects culture of continuous improvement with the more tangible operational practices of LM, a performance efficient hybrid might arise. The understanding of these two concepts is therefore beneficial in order to improve both LM and operational performance.

Lean Manufacturing (LM)

With some of the first publications dating back to over 25 years ago (Sugimori et al. 1977; Schonberger 1982; Womack et al. 1990), Lean Manufacturing (LM) has been a topic of interest for scholars and production companies for many decades (Marodin & Saurin 2013; Jasti & Kodali 2014) and has been used on a large scale by companies to gain competitive advantage (Belekoukias et al. 2014). As Hines et al. (2004) confirmed, LM has evolved beyond its origin and continues to develop and can be considered a management philosophy aimed at eliminating waste and increase value added activities (Pont et al. 2008).

Given the rich literature history of LM, there are several recent literature reviews available that summarize LM (Marodin & Saurin 2013; Jasti & Kodali 2014; Moyano-Fuentes & Sacristán-Díaz 2012; Papadopoulou & Özbayrak 2005; Pettersen 2009). One could expect LM to be unambiguous and concise, yet the opposite is true. Papadopoulou & Özbayrak (2005) pointed out that lean cannot be considered a set of tools and practices, but required a more holistic approach. Defining and measuring lean therefore could lead to conceptual problems (Pettersen 2009; Shah & Ward 2007). It is important to clearly describe LM and state what is in- and excluded from this research.

(8)

8

the emphasis will lie on the hard practices of LM rather than employee involvement. Emphasizing on the hard practices of lean is of great importance to avoid overlapping constructs and confusion between soft LM practices and CI.

Because of the extensive attention on LM within the literature, it can be hard to find a relevant research subject to contribute to the knowledge on LM. Future research suggestions on LM are considered in a summarizing paper by Marodin & Saurin (2013), who conducted a literature review on 102 studies on lean manufacturing. Their findings concluded that deepening the knowledge on the interaction between principles and trying to integrate the mostly fragmented LM literature is of importance (Marodin & Saurin 2013). Additionally, knowing what makes companies successful in implementing lean manufacturing is considered desired knowledge (Marodin & Saurin 2013). This study tries to deepen the knowledge on LM by looking into LM on a detailed level and cluster data into one research. This knowledge contributes to a more complete understanding of the success factors of LM.

To define the hard practices of LM, bundles can be used to assess all practices. Shah & Ward (2007) defined 10 constructs based on literature that can be assigned to four LM bundles. For LM, the following four bundles are used to assess all constructs of Shah & Ward (2007); SCM, JIT, TQM and HRM.

Supply Chain Management (SCM)

In line with LM, the goal of supply chain management is adding value for the overall customer (Kannan & Tan 2005). By integrating and aligning processes, SCM can drive performance and quality within the supply chain (Robinson & Malhotra 2005). As customers and suppliers are involved, SCM can be initiated by the company itself but requires external commitment and cooperation. As Lamming (1996) points out, this is exactly where the challenge lies. He claims that the leanness of a supply chain is not suitable for straightforward implementation (Lamming 1996). This means that implementation of SCM improvements can be complex and might require external knowledge.

(9)

9

Just-in-time (JIT)

Just-in-time is a method to produce the correct product at the correct time (Belekoukias et al. 2014). It is useful for performance improvement (Bortolotti et al. 2013) and implemented globally as a manufacturing strategy. Aimed at increasing profit and reducing waste, it is a cost-based strategy for operations (Chen 2015). JIT is linked with lean manufacturing (Dinsdale & Bennett 2015) and considered one of four measures of lean by Shah & Ward (2007). Several studies confirm the positive influence of JIT on operational performance, confirming the value of implementing this LM concept (Belekoukias et al. 2014; Chen 2015; Bortolotti et al. 2013; García et al. 2014). Operational constructs that are dedicated to JIT are according to Shah & Ward (2007) are; ‘pull’ Kanban system and JIT production, ‘flow’ continuously flowing products and ‘set up time reduction’, aimed at reducing downtimes at changeovers.

The requirement of employee involvement with JIT is considered to be high (Marin-Garcia & Bonavia 2015; Cua et al. 2001). Therefore one could expect that the level of employee involvement has an influence on the overall performance of the implemented JIT philosophy.

Total Quality Management (TQM)

In recent years, TQM has been applied in many industries (Das et al. 2008) and the topic has received increasing attention (Zhang et al. 2000). Even though there is consensus about TQM’s ability to increase performance (Das et al. 2008; Zhang et al. 2000), there is less uniformity on the exact constructs of TQM (Zhang et al. 2000). The goal of TQM programs is “improving an organization’s ability to deliver high-quality products or services in a cost-effective manner” (Lam et al. 2015:210). Shah & Ward (2007) assigned two operational constructs to TQM; ‘Total productive maintenance (TPM)’ achieving equipment availability by optimizing maintenance and ‘Statistical process control (SPC)’ minimizing defects throughout every process step.

As Lam et al. (2015) point out, continuous improvement is the mechanism used to reach the TQM objective of performance enhancement. Therefore it could be expected that a higher level of continuous improvement leads to a higher level of performance while implementing TQM. Up to now only a few studies have investigated the link between continuous improvement initiatives and the implementation of technical TQM aspects (Lam, O’Donnell, et al. 2015).

Human Resource Management (HRM)

(10)

10

included one operational construct for HRM; ‘involved employees’. The measurements in this specific framework consists of whether the shop-floor employees of the organization use cross functional training, drive suggestion programs, lead improvement efforts and are key in problem solving (Shah & Ward 2007). Although HRM can be assigned to the soft side, the measurements have a more ‘hard’ practice orientation as they are implementation tools (programs and training) and do not consider employee attitude and cultural aspects. This reasoning led to the fact that HRM in this study can be considered applicable to be researched in combination with CI.

From the previous explanation, the difference between continuous improvement and HRM can also be derived. This distinction is important to make, as also HRM and continuous improvement could have overlapping aspects. As stated earlier, continuous improvement refers to a more long term dynamic capability culture. Whereas HRM in this scenario is more oriented towards training and job enrichment. It can be concluded that the HRM concept can be initiated whenever desirable, whereas the continuous improvement mentality takes time to evolve and form. This is the most important distinction between the two concepts.

(11)

11

Hypothesis development and conceptual model

Before constructing the hypotheses, some variables require clarification to avoid misinterpretation. The following section contains four variables; CI maturity, LM importance, LM performance, and operational performance. CI maturity refers to the maturity level discussed earlier by Bessant (1999). LM importance is the degree to which the organization values LM, as strategies might differ. LM performance determines how good the organization scores on LM, a specific LM bundle or LM construct and can be considered a combination of LM importance and LM implementation. The last variable operational performance refers to the degree the company meets customer requirements and scores compared to competitors.

As lean manufacturing has received extensive attention, the main focus of this research lies within CI. As discussed earlier, employees plays a big role in the implementation of LM principles (Marin-Garcia & Bonavia 2015). The continuous improvement capability amongst these employees might therefore be of influence on LM performance. Given this linkage, the first hypothesis therefore is constructed to test the actual influence of the CI maturity level on the degree of LM performance levels. Additionally, LM performance will be set out in bundles and constructs to see a deeper level of the influence of CI maturity on LM performance in total and as building blocks. This extra level will give valuable insights in to what detailed extent CI has an influence on LM performance. Therefore the first hypothesis is constructed as following.

H1: If the level of CI maturity is higher, it is likely that the level of LM performance is higher

LM has been linked with operational performance many times, and testing this relation will probably result in no new conclusions. Testing this relation however is necessary for this study, as the aim is to amplify the LM performance in LM constructs and LM bundles to see their relation with operational performance. This detailed level will be achieved by structurally decomposing LM into the earlier discussed bundles and constructs and compare them to operational performance piece by piece. This decomposition of LM provides the added value for researching this linkage as it gives detailed information for the mechanisms behind LM. The second hypothesis is constructed as following:

H2: If the level of LM performance is higher, it is likely that the level of operational performance is higher

As a higher CI maturity level is expected to increase the dynamic capability and thus competitive advantage of a firm, it can be expected that it also has a beneficial influence on operational performance. Not many studies have focused on the relation between CI maturity and operational performance, but the outcome could be valuable for strategy decisions trying to optimize operational performance. The third hypothesis is therefore stated as following:

(12)

12

LM is a philosophy that is implemented on an organization wide level to gain competitive advantage (Belekoukias et al. 2014). Given this organization wide character, there seems to be a need for agreement on the importance of LM. Without this consensus in the LM importance, it can be very hard to incorporate the whole organization to comply with LM, as employees simply do not agree on where to improve. As CI structurally involves employees in this continuous process, one might expect that a higher level of CI maturity also improves this consensus level of LM importance. Moreover, this might lead to the fact that CI enhances the agreement of the employees on the importance of LM. the fourth hypothesis is therefore constructed as following

H4: If the level of CI maturity is higher, it is likely that the level of LM importance consensus is higher The final hypothesis is to check whether the level of LM importance consensus has an influence on operational performance. As stated earlier, it is expected that CI maturity positively influences the LM importance consensus level, thus more agreement. But the next logical question would arise, does this consensus lead to a higher operational performance? It could be expected that when there is more consensus, improvements are more likely to succeed and lead to an increase in operational performance. Therefore this last hypothesis will entail whether the consensus level on the LM importance is positively linked to operational performance

H5: If the level of LM importance consensus is higher, it is likely that the level of operational performance is higher

Combining these five hypothesis into a conceptual framework leads to the figure shown in figure 1.

(13)

13

METHODOLOGY

The method used in this paper is case study research, as the intent is to develop a probabilistic relation between the concepts in the conceptual framework. Case study has been widely used for theory testing in complicated situations in the OM field (Karlsson 2009). A probabilistic relation presumes that when the independent value changes, it is expected that the dependent variable behaves according to expectations (Dul & Hak 2008). As Dul & Hak (2008) state, the probabilistic relations described in the hypothesis can be tested with a comparative case study. In this case study, three methods are used to determine the presence of a probabilistic relation. All three methods are further elaborated on later in this chapter.

Case selection and description

Recommended in this kind of study is to identify a small population wherein the variance is much less than a large population (Dul & Hak 2008). This avoids replication studies to yield different outcomes. Additionally, within a small population it is expected that these cases have many similarities. In this so called ‘quasi survey comparative case study’, there is no definite answer of the amount of cases needed to be included in the study (Dul & Hak 2008). The maximum amount of cases available within time constraints is therefore used for analysis.

An important aspect in case selection is setting a boundary (Karlsson 2009). In this research there is chosen for small and medium (SME) production enterprises in the Netherlands. To specify this, companies had to be classified according to the North American Industry Classification System (NAICS) as sector 33 ‘manufacturing’. As the topic consists of LM, choosing production companies is most obvious and this categorization forms clear boundaries for industries. The second prerequisite for a company was its size. Expected is that within SME’s the relations between topics are much more apparent than in larger, more complex organizations. SME’s employ up to 250 employees, which was the maximum amount of employees for an organization to be included in the study. Additionally, All companies have to be located in the Netherlands to avoid large cultural influences and further reduce the total sample size available and increase similarities. Finally, the data was obtained during a period between 2008-2013. As discussed earlier, LM is evolving over time and comparing historical cases with more recent ones results in inequalities amongst cases.

Data collection and measures

To collect the data, appropriate measures are needed for each concept of the conceptual framework. In this section further explanation will be given regarding the collection and used measures of the concepts.

(14)

14

from their study in 1999 (Bessant 1999), they identified a set of constituent behaviors corresponding to maturity levels of CI. Out of these behavioral aspects, a 1-5 scale CI maturity level can be determined along with its characteristics. For each organization included in the case study, all 35 behavioral aspects are rated by respondents on a Likert scale from 1-5 (equal to the maturity level scale). From this data can be deducted what the average CI maturity level of the organization is.

Lean manufacturing is much in line with continuous improvement on terms of unambiguousness, meaning there is no perfect way to measure. Shah & Ward (2007) conducted extensive literature review to trace the main components of lean manufacturing published in the Journal of Operations Management. By doing so, they identified ten main operational constructs and 48 underlying operational measures of these components within lean manufacturing, visible in table 1.1. These operational constructs can be bundled into widely known LM bundles; SCM, JIT, TQM and HRM. By measuring all 48 identified operational measures on a 1-5 Likert scale and assessing these measurements to their corresponding constructs and bundles the implementation levels of components, bundles and overall lean can be measured. Additionally, the degree of importance of that operational measurement for each case is obtained using the same 1-5 Likert scale. For some companies, constructs and or measurements might receive less emphasis and are deliberately lower than other components due to strategy or possible trade-offs. By assessing importance within each topic, a more case specific lean performance measurement is provided.

Table 1.1: LM constructs and bundles, adapted from Shah & Ward (2007)

Lean level Lean

Bundle level SCM JIT TQM HRM

Construct level Supplier feedback JIT delivery by suppliers Supplier development Customer involvement pull Continuous flow Set up time reduction Total productive maintenance Statistical process control Employee involvement

abbreviation SUPPFEED SUPPJIT SUPPDEVT CUSTINV PULL FLOW SETUP TPM SPC EMPINV Operational

measurements 3 3 6 5 4 4 3 5 4 6

(15)

15

respondents are asked the corresponding customer importance of each aspect due to differences in operational strategy

Data analysis

Prior to analyzing the data based on the proposed methodology, a visual inspection is done regarding all cases. Some questionnaires contained empty values or missing data. In every instance, it was possible to still average the data from remaining respondents in order to extract an average score for that measurement. This resulted in no excluding cases based on missing data and it is expected this has no consequences for data validity. The total amount of cases involved in this study was concluded to be 20. Analyzing the data contained many repetitions of testing relations and relationship strengths between all constructs. Prior to this analysis, each concept’s implementation and performance had to be calculated from the obtained data. This will be done in the following order: CI, LM performance, operational performance and LM importance consensus.

Continuous improvement levels consisted of the most simple analysis approach. The goal was to obtain a CI maturity level for each case. This is achieved firstly by averaging the respondents’ results per behavioral aspect and secondly averaging all these figures to obtain the overall maturity average. This maturity level consist of a value between 1-5, equal to the Likert scale provided in the questionnaire.

Within the LM concept, simply averaging the values did not reflect LM performance reality. Some lean bundles might receive less attention (and thus a lower implementation level) as that bundle is of lesser importance to the organization. Due to this reason, LM importance has to be taken into account to calculate the weighted average of LM performance. Using a weighted average, the LM importance level is used as a denominator in the equation to weigh implementation levels. For each construct, bundle and overall LM, the same equation is used to obtain the average LM performance level. This equation is visible in figure 2.1 wherein W is weight (importance) and I equals implementation. The outcome of the equation represents the average LM performance level.

Figure 2.1: Equation lean calculation

𝑋� = ∑𝑛𝑖=1 𝑊𝑖𝑊∗ 𝐼𝑖

𝑖 𝑛 𝑖=1

(16)

16

has 6 involved (see table 1.1). Applying this weighted average formula for each case on LM, the four LM bundles and the 10 constructs results in 15 average data points for each case. Each data point contains a different scope level of LM performance and can give valuable insights in looking deeper into the working methods of LM.

To calculate overall operational performance for each organization, weighted average is used again. In this calculation, the customer importance provided by the respondent of each performance indicator is used as its weight. This results in the equation for operational performance visible in figure 2.2. In this equation the P equals the company’s performance and the I corresponds to the importance of the customer. The lower case numbers represent the 7 different operational measures by slack & Lewis (2011) discussed earlier. The outcome of the equation represents the average operational performance level.

Figure 2.2: Equation performance calculation

𝑋� = 𝑃1𝐼1+ 𝑃2𝐼𝐼2+ 𝑃3𝐼3+𝑃4𝐼4+ 𝑃5𝐼5+ 𝑃6𝐼6+ 𝑃7𝐼7

1+𝐼2+𝐼3+ 𝐼4+𝐼5+ 𝐼6+𝐼7

Where the weighted averages are widely represented and used methods, the calculation of the LM importance consensus is a more intuitive approach. This could have a few consequences for the results. There is deliberately not chosen for the more mathematical approach of standard deviations. This appeared to be very influenced by the cases containing just two measurements. Therefore in this analysis, firstly the average over all respondents in each measurement is calculated to ensure the overall level. Secondly, the deviation of each respondent in terms of that average is calculated in absolute values. For example, two respondents answered respectively 6 and 8, the average is 7 and both deviations from this average are 1. Finally, all absolute deviations are divided by the number of respondents to incorporate the different amount of respondents, in previous example resulting in 2 divided by 2, which is 1. This results in an average level of consensus for each measurement in which a lower number represents more agreement (less deviation in comparison to average). When more constructs are incorporated, for example in LM total level, all construct consensus levels are averaged. This is also the case for the LM bundles, which contain only a selected amount of constructs. Optimally this leads to a LM importance consensus score of 0 which means total agreement on the importance of LM. This calculation will only be performed on top level and bundle level for LM. Case comparison methods

(17)

17

2008). This method makes use of ranking values of independent and dependent variables to see whether the rank orders are alike. By grouping the values, based on the independent variable, the grouped averages of the corresponding dependent variable are inspected. If the grouped rank order of the dependent variable is similar to the independent variable, the hypothesis is accepted to have evidence of a probabilistic relation. Being able to answer whether there is evidence of a probabilistic relation is one of the main strengths of this approach. An important side note to this approach is that the lack of evidence does not necessarily mean there is no probabilistic relation. This is one of the downsides of this approach. Therefore two other methods are performed in order to support (or question) the claim of the ranking values approach.

As this approach uses no statistics, there is no agreement upon using a standard group size for ranking. Dul & Hak (2008) suggest quartiles. However, simply looking and choosing one group size is too short sighted and therefore three different group sizes are chosen; three groups, four groups (quartiles) and five groups. These groups are based upon Dul & Hak's (2008) suggestion but also incorporate a level below and above this suggestion. The expectation is that by viewing these three groups, the most accurate prediction can be made upon whether a relation is probabilistic or not. This is mainly to avoid selecting a single group which could bias the results. In Table 1.2 it is shown how the cases are divided in each group.

Table 1.2: Distribution of the cases

Three groups Four groups Five groups

Distribution of the 20 cases 6-7-7 5-5-5-5 4-4-4-4-4

After testing whether a relation is expected to be probabilistic, two other methods are applied to see whether the first results hold. Additionally, those outcomes will help determine the strength of each relationship. This could give valuable insights in which concepts of the conceptual framework have more influence than other concepts. In this case study methodology it is impossible to determine whether the strength is significance and therefore two approaches are used that give an indication of the relationship strength. The first approach is a more visual inspection of between case comparison and the second contains a correlation coefficient.

Between case comparison

(18)

18

amount of values, a percentage is shown. This percentage determines the percentage of times the hypothesis hold and thus the likelihood of the hypothesis to hold in other cases. Obviously a higher percentage indicates a more solid linkage. A strong aspect of this method is that it is a good indicator of the likelihood a hypothesis holds, giving an indication of whether there is a probabilistic relation and the strength of this relation.

Correlation coefficient

In addition to the between case comparison, a correlation coefficient is also calculated. There is awareness that coefficients are very subject to data outliers and that the sample size for this method is relatively small. The correlation coefficient is mainly used to assist the relationship strength of the case comparison and widen the view of determining the strength. The coefficient will be looked upon with caution due to previous mentioned weaknesses in this research, but a secondary method will be valued above using a single method as it can help in assisting the relationship strength. By comparing all values of all cases on terms of their independent an dependent value, a coefficient is determined. The coefficient that is used is the Pearson’s product-moment coefficient, as this is the most familiar coefficient. This coefficient is expected to be in line with the between case comparison in terms of relationship strength. As the data is bounded 1-5 and 1-10, expected is that this effect is of lesser influence. It should be kept in mind that this is a statistic approach used in a case study, hence the assisting role. The coefficient is rounded down in two decimals.

Validity and reliability

(19)

19

expected is that this will enhance the generalizability. Additionally, there is barely any researcher bias apparent in the questionnaires.

(20)

20

RESULTS

In this section all results of the three approaches will be summarized in tables. They will be structured according to the hypotheses provided. The first and second columns respectively describe the independent and dependent variable of the relation. The third column provides a positive or negative answer on the ranking method described in the methodology. A yes provides evidence of a probabilistic relation. A yes will be generated if there’s evidence in one or more groups mentioned in the methodology. In the fourth column the between case percentage is shown. This percentage is the number of times the hypothesis holds in the case comparison. The last column is the correlation coefficient. The outcome of the correlation coefficient can lie between -1 and 1. 0 indicates total lack of correlation whereas both ultimatums determine the strongest correlation, -1 being a negative reversed correlation and 1 a positive correlation. Logically, the further the value lies towards the ends of this scale, the stronger the correlation is considered. All calculations with the ranking order, the between case comparison explanation and the coefficient calculation of all relationships measured are shown in the appendix. Therefore the in-depth analysis of the ranking and level determining is shown in appendix A. This section solely displays and describes the results from the analysis.

CI maturity and LM performance

Testing the first hypothesis entailed whether there is evidence of a probabilistic relation between CI maturity levels and LM performance. Table 2.1 shows the analysis outcome for this hypothesis. It can be seen that there is evidence of a probabilistic relation, and both the between case percentage and correlation coefficient are considered strong.

Table 2.1: Results CI and LM

(21)

21

Table 2.2: Results CI and LM bundles Independent variable Dependent variable Evidence of probabilistic relation Between case percentage Correlation coefficient CI SCM yes 72,63% 0,74 JIT yes 65,26% 0,46 TQM yes 71,05% 0,64 HRM yes 68,95% 0,56

The next step is to look at the individual constructs making up the bundles. Similar to LM, the LM bundles still are somewhat broad. Therefore the impact of CI on all 10 constructs identified by Shah & Ward (2007) is shown in table 2.3. Here it is visible that for most cases evidence of probabilistic relation goes paired with a between case percentage of 65% or higher and a coefficient of 0,50 or higher. In this table however there are two exceptions. For ‘supplier feedback’ both the between case percentage and the coefficient are high, but the ranking order is low. For ‘flow’ the opposite is visible; low percentage and low coefficient but evidence of a probabilistic relation. As stated earlier, the ranking order method of Dul & Hak (2008) purely gives evidence of a probabilistic relation, it does not confirm nor disprove the presence of a probabilistic relation. Therefore a yes or no is not binding and the other indicators should be taken into account. In this case expected is that for ‘supplier feedback’ the chances are likely there is a probabilistic relation, given the high between case percentage and correlation coefficient. On the other hand it can be assumed that this is not the case for ‘flow’. The coefficient turned out to be 0 and the between case percentage is very low. All other construct outcomes are in line with each other, resulting in evidence of a probabilistic relation between CI maturity and the constructs ‘supplier just-in-time’, ‘supplier development’, ‘setup time reduction’, ‘statistical process control’, ‘total productive maintenance’ and ‘employee involvement’. On the other hand there is no evidence for CI maturity and the construct ‘customer involvement’ and ‘pull’.

(22)

22

Based on the results, there can be concluded that there is a lot of evidence of probabilistic relations between CI maturity and LM performance, CI maturity and LM bundle performance and CI maturity and LM construct performance. The top level has strong evidence according to the ranking, a between case percentage of 70,53% and a high coefficient of 0,73. The LM bundle levels all proved to have evidence as well, followed by corresponding high between case percentages and correlation coefficients. Connecting the outcomes of the in total 15 relations rested to the hypothesis, there is no disagreement on the fact that there is strong evidence of the fact that it is likely that the level of LM performance is higher once the level of CI maturity is higher, therefore the first hypothesis is accepted. LM performance and operational performance

LM performance on operational performance proved both to have evidence of a probabilistic relation as well as a strong between case percentage and coefficient. All three factors indicate a probabilistic relation for both concepts, visible in table 2.4.

Table 2.4: Results LM and operational performance Independent variable Dependent variable Evidence of probabilistic relation Between case percentage Correlation coefficient LM PERF yes 71,05% 0,60

The bundle level of LM however showed that in 75% of the cases the bundles have no evidence of a probabilistic relation, visible in table 2.5. Only SCM proved to have strong evidence of a probabilistic relation in all three measurements. One reason for this outcome could be the fact that SCM has most constructs involved and thus a higher influence overall. The lack of JIT bundle evidence can be discussed, given the high between case comparison, but still there is no convincing result of the LM bundles’ probabilistic relation to operational performance.

(23)

23

The division of evidence of a probabilistic relation within the LM constructs is 50/50. From table 2.6 it is visible that ‘supplier JIT’ and ‘setup time reduction’ have the strongest link to operational performance. Also ‘customer involvement’ and ‘statistical process control’ are considered positive on a probabilistic relation, given their strong between case comparison and coefficient. These four construct have evidence of a probabilistic relation with operational performance. On the other hand, ‘supplier feedback’, ‘supplier development’, ‘flow’ and ‘employee involvement’ proved to have no evidence of a relation with operational performance, supported by the between case comparison and correlation coefficient. ‘Pull’ and ‘total productive maintenance’ have contradicting results. The former has no evidence but rather high percentage and coefficient, the latter has evidence but a low percentage and a very low coefficient. For these abnormalities there is no explanation.

Table 2.6: Results LM constructs and operational performance Independent variable Dependent variable Evidence of probabilistic relation Between case percentage Correlation coefficient SUPPFEED PERF no 58,95% 0,35 SUPPJIT yes 75,26% 0,68 SUPPDEVT no 56,84% 0,39 CUSTINV yes 67,37% 0,53 PULL no 67,37% 0,49 FLOW no 50,53% -0,13* SETUP yes 70,53% 0,56 SPC yes 67,89% 0,52 TPM yes 55,78% 0,18 EMPINV no 58,95% 0,39

It is likely that the level of operational performance is higher when the level of LM performance is higher thus the hypothesis can be approved. This outcome is specifically based on the top level outcome, which proved to be very high. With 71,05% chance of the hypothesis holding, evidence according to the ranking and a strong coefficient, all indicators point towards a probabilistic relation. There is no reason to reject the hypothesis based on the lack of evidence in 3 out of 4 LM bundles, as all the between case percentages and correlation results are average and therefore do not disprove the evidence of the top level.

CI maturity and operational performance

(24)

24

Table 2.7: Results CI and operational performance

Independent variable Dependent variable

Evidence of probabilistic relation

Between case

percentage Correlation coefficient

CI PERF no 58,42% 0,31

It is too nearsighted to claim that CI maturity does not lead to operational performance improvements at all, but from previous results it is very clear that CI works best on operational performance throughout the tangible implementations of LM. When considering both philosophies this lies within expectations, as LM requires human involvement, and continuous improvement requires tangible practices to improve. Therefore the hypothesis of a probabilistic relation between CI maturity and operational performance will be rejected

CI maturity and LM importance consensus

Important to note for this part is the reversed correlation coefficient. As discussed earlier, a lower consensus grade is associated with higher level of CI maturity and operational performance. This situation causes the coefficient to behave negatively due to the reversed relation. There is evidence found of CI maturity on the LM importance consensus level. Followed by a somewhat average between case percentage and medium correlation coefficient. This is visible in table 2.8.

Table 2.8: Results CI and LM importance consensus Independent

variable Dependent variable

Evidence of probabilistic relation Between case percentage Correlation coefficient

CI LM importance consensus yes 60,53% -0,44

For this analysis table 2.9 shows that 2 out of 4 bundles proved to have evidence of a probabilistic relation. All between case percentages can be considered mediocre and much alike. Only the coefficient of JIT and TQM are on the higher side. Remarkable is the positive coefficient of HRM, which is not in line with expectations.

Table 2.9: Results CI and LM bundle importance consensus Independent

variable Dependent variable

Evidence of probabilistic relation Between case percentage Correlation coefficient CI

SCM importance consensus yes 57,89% -0,33

JIT importance consensus yes 60,53% -0,50

TQM importance consensus no 58,95% -0,40

HRM importance consensus no 57,37% 0,15

(25)

25

probabilistic relation and 50% of the bundles as well. Although numbers are not convincingly, the majority points towards a probabilistic nature and therefore the hypothesis is accepted.

LM importance consensus and operational performance

The LM importance consensus is shown to have no probabilistic relation with operational performance. This outcome is enhanced by the low between case percentage and coefficient, visible in table 2.10.

Table 2.10: Results LM importance consensus and operational performance Independent variable Dependent

variable

Evidence of probabilistic relation

Between case

percentage Correlation coefficient

LM importance consensus PERF no 46,32% -0,07

Table 2.10 shows that the LM importance consensus of none of the bundles proved to have evidence of a probabilistic relation with operational performance as well. All between case percentages are very low, and their coefficients are in line with this number. In this situation again, two correlations appeared to be positive. This is rather surprising given the earlier discussed reversed relation. Although the reversed correlations are low and do not necessarily imply a reversed relation, their presence is remarkable.

Table 2.11: Results LM bundle importance consensus and operational performance Independent variable Dependent

variable Evidence of probabilistic relation Between case percentage Correlation coefficient SCM importance consensus PERF no 47,89% -0,09

JIT importance consensus no 45,79% 0,042

TQM importance consensus no 47,89% -0,26

HRM importance consensus no 44,74% 0,21

(26)

26

DISCUSSION

The purpose of this section is to discuss the results proposed in this study. In figure 3 the conceptual framework is shown with the updated hypotheses results. The section will be structured according to the order of the hypothesis. The discussion ends with a few limitations encountered in this study.

Figure 3: Results conceptual model

CI maturity and LM performance

This study found results that there is a probabilistic relation between CI maturity and LM performance. This is in line with the findings of Singh & Singh (2015), who claim that both go hand in hand and operational practices benefit from CI implementation programs. In addition, as Marodin & Saurin (2013) stated, LM relies heavily on employee involvement. This relation can be considered strong, given the amount of evidence found in all levels of comparison in all three methods. The division of the probabilistic relation amongst the constructs is less uniform. This might indicate that some bundles indeed are more influenced by CI maturity levels.

LM performance and operational performance

(27)

27

CI maturity and operational performance

There is no evidence of a probabilistic relation with CI maturity and operational performance, followed by a low between case comparison percentage (58,42) and a low coefficient (0,31). This possible forms the most interesting conclusion. The strong relationship between CI maturity and LM performance and the subsequent relation of LM performance on operational performance has been confirmed. But we now have results that CI maturity does not have a probabilistic relation with operational performance directly. This indicates that the usefulness of CI is only exploited through the implementation of tangible LM principles. This confirms again that employees and operational practices are best used together (Marodin & Saurin 2013; Singh & Singh 2015).

CI maturity and LM importance consensus

The level of CI maturity proved to have evidence of a probabilistic relation compared to the consensus on LM importance with a between case percentage of 60,53% and correlation of -0,44. This is in line with expectations, as a higher level of CI maturity encompasses more employee involvement and thus more agreement upon importance of LM. Also on bundle level, CI maturity was in 50% of the cases linked with a probabilistic relation to more agreement on the importance of the bundles. The between case percentages and correlation coefficients were largely in line with the evidence result. This result should be viewed with the knowledge that the consensus calculation was an intuitive one, and might therefore be under viewed or exaggerated. Further study on this topic is necessary to give a more trustworthy outcome.

LM importance consensus and operational performance

(28)

28

Limitations

In addition to the conclusions found, some downsides are apparent in this research. Some of these downsides are the reason of performing a case study on data that has more survey characteristics. Measuring the concepts in the conceptual model lacked unambiguousness. Even though widely accepted scientific measurements are used to determine the levels of the concepts, there is no perfect way of measuring. Measuring operational performance for example can also be performed on a KPI level including actual data from the organization. Although expected is the most applicable measures are used, one could discuss to use other measures to evaluate CI maturity, LM performance and operational performance.

Furthermore, the calculation of the data can be subject to researcher’s bias. Many operationalizations and calculations of the constructs were considered, but eventually not incorporated. There is tried to examine a wide range of calculations and select the one that displayed the data best. Expected is that the calculations realistically represent the desired measurements, but it remains rather subjective. This is especially the case with the consensus calculations. As mentioned earlier, based upon calculations that were not apparent in literature, this could have implications for the conclusions of the last two hypothesis.

(29)

29

CONCLUSION

Both continuous improvement and lean manufacturing have been used extensively by manufacturing firms worldwide. This draws attention to the focus on the mechanisms of how these philosophies work separately and together. This study tried to fill in the literature gap by examining the relation between continuous improvement, lean and operational performance. Moreover it was aimed towards creating understanding in how to improve operational performance by using continuous improvement and lean manufacturing. Additionally there is tried to examine the effects of CI maturity on LM performance, and LM performance on operational performance on a very detailed level.

There is concluded that in fact there is evidence of a probabilistic relation between continuous improvement and lean manufacturing performance, lean manufacturing performance and operational performance and continuous improvement maturity and LM importance consensus. In addition to this, the influence of continuous improvement on the underlying LM bundles and constructs and their influence on operational performance is explored. This turned out to have varying results in how the constructs behave. Subsequently, no evidence is found of a probabilistic relation between continuous improvement and operational performance directly, as well as lean importance consensus on operational performance. The knowledge of the tested hypothesis can give valuable benefits in creating a business advantage throughout combining both philosophies or specific bundles or constructs.

Managerial implications

(30)

30

Suggestions for further research

As the results solely give evidence of probabilistic relations, this does not necessarily imply that there indeed is a relation nor disproves the fact that there is a relation. Logically the suggestion for further research would be to perform this study with a larger sample size to be able to give significant answers. These statistics will strengthen the relationships explored by us and it would be no surprise if results on construct or bundle level would differ on some aspects. Additionally, it will give more detailed information on the sometimes scattered results within the LM constructs.

(31)

31

REFERENCES

1. Ambrosini, V., Bowman, C. & Collier, N., 2009. Dynamic Capabilities: An Exploration of How Firms Renew their Resource Base. British Journal of Management, 20, pp.S9–S24. 2. Bamber, G.J. et al., 2014. Human resource management, Lean processes and outcomes for

employees: towards a research agenda. The International Journal of Human Resource Management, 25(21), pp.2881–2891.

3. Belekoukias, I., Garza-Reyes, J.A. & Kumar, V., 2014. The impact of lean methods and tools on the operational performance of manufacturing organisations. International Journal of Production Research, 7543(July 2014), pp.1–21.

4. Bessant, J., 1999. Developing Continuous Improvement Capability. International Journal of Innovation Management, 02(04), pp.409–429.

5. Bessant, J., Caffyn, S. & Gallagher, M., 2001. An evolutionary model of continuous improvement behaviour. Technovation, 21(2), pp.67–77.

6. Bhuiyan, N., Baghel, A. & Wilson, J., 2006. A sustainable continuous improvement methodology at an aerospace company. International Journal of Productivity and Performance Management, 55(8), pp.671–687.

7. Bortolotti, T., Danese, P. & Romano, P., 2013. Assessing the impact of just-in-time on operational performance at varying degrees of repetitiveness. International Journal of Production Research, 51(4), pp.1–14.

8. Chen, Z., 2015. The relationships among JIT, TQM and production operations performance. Business Process Management Journal, 21(5), pp.1015–1039.

9. Cua, K.O., McKone, K.E. & Schroeder, R.G., 2001. Relationships between implementation of TQM, JIT, and TPM and manufacturing performance. Journal of Operations Management, 19(6), pp.675–694.

10. Das, A., Paul, H. & Swierczek, F.W., 2008. Developing and validating total quality management (TQM) constructs in the context of Thailand’s manufacturing industry. Benchmarking: An International Journal, 15(1), pp.52–72.

11. Dinsdale, E.J. & Bennett, D., 2015. Benefits; drawbacks and boundaries to deliver JIT. Benchmarking: An International Journal, 22(6), pp.1081–1095.

12. Dul, J. & Hak, T., 2008. Case Study Methodology in Business Research, Routledge.

13. García, J.L. et al., 2014. Structural equations modelling for relational analysis of JIT performance in maquiladora sector. International Journal of Production Research, 7543(February), pp.1–19.

14. Hines, P., Holweg, M. & Rich, N., 2004. Learning to evolve. International Journal of Operations & Production Management, 24(10), pp.994–1011.

15. Jasti, N.V.K. & Kodali, R., 2014. Lean production: literature review and trends. International Journal of Production Research, 7543(August), pp.1–19.

16. Kannan, V. & Tan, K.C., 2005. Just in time, total quality management, and supply chain management: understanding their linkages and impact on business performance. Omega, 33(2), pp.153–162.

17. Karlsson, C. Ed. 2009. Researching operations management. Routledge

18. Kull, T.J. & Wacker, J.G., 2010. Quality management effectiveness in Asia: The influence of culture. Journal of Operations Management, 28(3), pp.223–239.

(32)

32

20. Lam, M., Donnell, M.O. & Robertson, D., 2015. Achieving employee commitment for continuous improvement initiatives Article information :

21. Lam, M., O’Donnell, M. & Robertson, D., 2015. Achieving employee commitment for continuous improvement initiatives. International Journal of Operations & Production Management, 35(2), pp.201–215.

22. Lamming, R., 1996. Squaring lean supply with supply chain management. International Journal of Operations & Production Management, 16(2), pp.183–196.

23. de Leede, J. & Kees Looise, J., 1999. Continuous improvement and the mini‐company concept. International Journal of Operations & Production Management, 19(11), pp.1188– 1202.

24. Marin-Garcia, J. a. & Bonavia, T., 2015. Relationship between employee involvement and lean manufacturing and its effect on performance in a rigid continuous process industry. International Journal of Production Research, 53(11), pp.3260–3275.

25. Marodin, G.A. & Saurin, T.A., 2013. Implementing lean production systems : research areas and opportunities for future studies. , 51(August), pp.37–41.

26. Moyano-Fuentes, J. & Sacristán-Díaz, M., 2012. Learning on lean: a review of thinking and research. International Journal of Operations & Production Management, 32(5), pp.551–582. 27. Pakdil, F. & Leonard, K.M., 2015. The effect of organizational culture on implementing and

sustaining lean processes. Journal of Manufacturing Technology Management, 26(5), pp.725– 743.

28. Papadopoulou, T.C. & Özbayrak, M., 2005. Leanness: experiences from the journey to date. Journal of Manufacturing Technology Management, 16(7), pp.784–807.

29. Pettersen, J., 2009. Defining lean production: some conceptual and practical issues. The TQM Journal, 21(2), pp.127–142.

30. Pont, G.D., Furlan, A. & Vinelli, A., 2008. Interrelationships among lean bundles and their effects on operational performance. Operations Management Research, 1(2), pp.150–158. 31. Power, D., Schoenherr, T. & Samson, D., 2010. The cultural characteristic of

individualism/collectivism: A comparative study of implications for investment in operations between emerging Asian and industrialized Western countries. Journal of Operations Management, 28(3), pp.206–222.

32. Robinson, C.J. & Malhotra, M.K., 2005. Defining the concept of supply chain quality management and its relevance to academic and industrial practice. International Journal of Production Economics, 96(3), pp.315–337.

33. Sanchez, L. & Blanco, B., 2014. Three decades of continuous improvement. Total Quality Management & Business Excellence, 25(9-10), pp.986–1001.

34. Schonberger, R.J., 1982. The Transfer of Japanese Manufacturing Management Approaches to U.S. Industry. Academy of Management Review, 7(3), pp.479–487.

35. Shah, R. & Ward, P.T., 2007. Defining and developing measures of lean production. Journal of Operations Management, 25(4), pp.785–805.

36. Shah, R. & Ward, P.T., 2003. Lean manufacturing: Context, practice bundles, and performance. Journal of Operations Management, 21(2), pp.129–149.

37. Singh, J. & Singh, H., 2015. Continuous improvement philosophy – literature review and directions. Benchmarking: An International Journal, 22(1), pp.75–119.

38. Slack, N. & Lewis, M. 2011. Operations strategy. Pearson education, Ltd.

(33)

33

15(6), pp.553–564.

40. Womack, J.P., Jones, D.T. & Roos, D., 1990. The Machine that Changed the World: The Story of Lean Production,

41. Yin, R.K., 1994. Case Study Research: Design and Methods. Applied Social Research Methods Series, 5, p.219.

(34)

34

APPENDICES

APPENDIX A: RANKING METHOD, BETWEEN CASE TABLE AND CORRELATION

1. CI on LM

1.1 CI on LM; bundle level

Company CI score CI rank LM score LM rank 3 groups 4 groups 5 groups CI LM 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 7 3,850 1 3,094 4 2,133 2,305 1 X 13 3,743 2 3,519 1 3,114 3,387 2 1 X 19 3,214 3 3,082 5 2,329 2,828 3 1 1 X 2 3,114 4 3,387 2 3 2,499 2,607 4 1 1 0 X 21 2,752 5 3,188 3 3 2,671 2,632 5 1 1 0 1 X 23 2,743 6 2,507 16 5,16667 2,461 2,656 6 1 1 0 0 0 X 9 2,737 7 3,031 6 3,850 3,094 7 1 0 1 1 1 1 X 14 2,695 8 2,335 19 11 2,737 3,031 9 1 1 1 1 1 1 1 X 5 2,671 9 2,632 12 2,171 2,625 10 1 1 1 0 1 1 1 1 X 4 2,499 10 2,607 14 13,4 2,293 2,388 11 1 1 1 1 1 1 1 1 0 X 6 2,461 11 2,656 11 2,219 2,847 12 1 1 0 0 0 0 1 1 1 0 X 18 2,425 12 2,677 10 11,75 3,743 3,519 13 1 1 1 1 1 1 0 1 1 1 1 X 22 2,379 13 2,493 17 12,7143 2,695 2,335 14 1 1 0 0 0 0 1 1 0 0 0 1 X 15 2,343 14 2,714 9 2,343 2,714 15 1 1 0 0 0 0 1 1 1 1 0 1 0 X 3 2,329 15 2,828 8 11 1,844 2,543 16 0 1 1 1 1 1 1 1 1 0 1 1 0 1 X 11 2,293 16 2,388 18 13 2,425 2,677 18 1 1 0 0 0 0 1 1 1 1 0 1 0 0 1 X 12 2,219 17 2,847 7 3,214 3,082 19 1 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 X 10 2,171 18 2,625 13 2,752 3,188 21 1 1 1 1 1 1 0 1 1 1 1 1 1 1 1 1 0 X 1 2,133 19 2,305 20 2,379 2,493 22 1 1 0 1 1 1 1 1 0 1 0 1 0 0 0 1 1 1 X 16 1,844 20 2,543 15 12,8571 14,6 13,75 2,743 2,507 23 1 1 0 0 0 0 1 0 0 1 0 1 1 0 0 0 1 1 1 X CI LM total 190 positive 134 part 70,53% Correlation 0,7236489

Company CI score CI rank SCM score

SCM

rank 3 groups 4 groups 5 groups CI SCM 1 2 3 4 5 6 7 9 10 11 12 13 14 15 16 18 19 21 22 23

(35)

35

Company CI score CI rank JIT score JIT rank 3 groups 4 groups 5 groups CI JIT 1 2 3 4 5 6 7 9 10 11 12 13 14 15 16 18 19 21 22 23

7 3,850 1 2,869 8 2,1333 2,3146 1 X 13 3,743 2 3,590 3 3,1143 3,7772 2 1 X 19 3,214 3 3,606 2 2,3286 2,4808 3 1 1 X 2 3,114 4 3,777 1 3,5 2,4993 2,7359 4 1 1 1 X 21 2,752 5 3,293 4 3,6 2,6714 2,7271 5 1 1 1 0 X 23 2,743 6 2,087 18 6 2,4612 2,5198 6 1 1 1 1 1 X 9 2,737 7 3,231 5 3,85 2,8689 7 1 0 1 1 1 1 X 14 2,695 8 1,796 20 11,75 2,7371 3,2311 9 1 1 1 1 1 1 0 X 5 2,671 9 2,727 13 2,171 2,7319 10 1 1 0 1 0 0 1 1 X 4 2,499 10 2,736 11 13,4 2,2929 1,9869 11 0 1 1 1 1 1 1 1 0 X 6 2,461 11 2,520 15 2,219 3,1488 12 1 1 0 0 0 0 0 1 1 0 X 18 2,425 12 2,943 7 11,5 3,7429 3,5902 13 1 0 1 1 1 1 0 1 1 1 1 X 22 2,379 13 2,686 14 12,1429 2,6952 1,7956 14 0 1 0 0 0 0 1 1 0 0 0 1 X 15 2,343 14 2,748 10 2,3429 2,7481 15 1 1 1 0 0 0 1 1 1 1 0 1 0 X 3 2,329 15 2,481 16 12,4 1,8438 2,8181 16 0 1 0 0 0 0 1 1 0 0 1 1 0 0 X 11 2,293 16 1,987 19 14,75 2,4247 2,9435 18 1 1 1 0 0 0 0 1 1 1 0 1 0 1 1 X 12 2,219 17 3,149 6 3,2143 3,6061 19 1 0 1 1 1 1 0 1 1 1 1 0 1 1 1 1 X 10 2,171 18 2,732 12 2,7524 3,2931 21 1 1 1 1 1 1 0 1 1 1 1 1 1 1 1 1 1 X 1 2,133 19 2,315 17 2,3786 2,686 22 1 1 1 1 1 0 1 1 0 1 0 1 0 0 0 1 1 1 X 16 1,844 20 2,818 9 12,7143 12,6 11 2,7429 2,0868 23 0 1 0 0 0 0 1 0 0 1 0 1 1 0 0 0 1 1 0 X CI JIT total 190 positive 124 part 65,26% Correlation 0,46306

Company CI score CI rank TQM score

TQM

rank 3 groups 4 groups 5 groups CI TQM 1 2 3 4 5 6 7 9 10 11 12 13 14 15 16 18 19 21 22 23

7 3,850 1 2,920 4 2,1333 1,757 1 X 13 3,743 2 3,549 1 3,1143 2,565 2 1 X 19 3,214 3 2,552 9 2,3286 2,419 3 1 1 X 2 3,114 4 2,565 8 5,5 2,4993 2,308 4 1 1 0 X 21 2,752 5 3,313 2 4,8 2,6714 2,584 5 1 0 1 1 X 23 2,743 6 2,028 16 6,66667 2,4612 2,737 6 1 0 1 0 0 X 9 2,737 7 2,998 3 3,85 2,920 7 1 1 1 1 1 1 X 14 2,695 8 1,747 19 10 2,7371 2,998 9 1 0 1 1 1 1 0 X 5 2,671 9 2,584 7 2,171 2,510 10 1 1 0 0 1 1 1 1 X 4 2,499 10 2,308 12 11,4 2,2929 2,225 11 1 1 1 1 1 1 1 1 0 X 6 2,461 11 2,737 5 2,219 2,227 12 1 1 1 1 1 1 1 1 0 0 X 18 2,425 12 2,186 15 9,75 3,7429 3,549 13 1 1 1 1 1 1 0 1 1 1 1 X 22 2,379 13 1,888 17 11,1429 2,6952 1,747 14 0 1 0 0 0 0 1 1 0 0 0 1 X 15 2,343 14 2,732 6 2,3429 2,732 15 1 0 1 0 0 1 1 1 1 1 1 1 0 X 3 2,329 15 2,419 11 10,8 1,8438 1,715 16 1 1 1 1 1 1 1 1 1 1 1 1 1 1 X 11 2,293 16 2,225 14 12 2,4247 2,186 18 1 1 0 1 1 1 1 1 0 0 0 1 0 0 1 X 12 2,219 17 2,227 13 3,2143 2,552 19 1 0 1 1 0 0 1 0 1 1 1 1 1 0 1 1 X 10 2,171 18 2,510 10 2,7524 3,313 21 1 0 1 1 1 1 0 1 1 1 1 1 1 1 1 1 0 X 1 2,133 19 1,757 18 2,3786 1,888 22 1 1 0 1 1 1 1 1 0 0 0 1 0 0 1 1 1 1 X 16 1,844 20 1,715 20 13,1429 15 15,25 2,7429 2,028 23 1 1 0 0 0 0 1 0 0 0 0 1 1 0 1 0 1 1 1 X CI TQM total 190 positive 135 part 71,05% Correlation 0,634585

Company CI score CI rank HRM score

HRM

rank 3 groups 4 groups 5 groups CI HRM 1 2 3 4 5 6 7 9 10 11 12 13 14 15 16 18 19 21 22 23

(36)

36

1.1.1 CI on construct level SCM bundle

Company CI score CI rank SUPPFE ED score

SUPPFE

ED rank 3 groups 4 groups 5 groups CI

SUPPFE ED 1 2 3 4 5 6 7 9 10 11 12 13 14 15 16 18 19 21 22 23 7 3,850 1 3,839 3 2,1333 3 1 X 13 3,743 2 4,160 1 3,1143 3,5 2 1 X 19 3,214 3 3,728 4 2,3286 3,3333 3 1 1 X 2 3,114 4 3,500 9 4,25 2,4993 3,1492 4 1 1 0 X 21 2,752 5 3,577 7 4,8 2,6714 3,25 5 1 1 0 1 X 23 2,743 6 3,905 2 4,33333 2,4612 3,3597 6 1 1 1 0 0 X 9 2,737 7 3,327 13 3,85 3,8393 7 1 1 1 1 1 1 X 14 2,695 8 3,536 8 7,5 2,7371 3,327 9 1 1 0 1 1 0 1 X 5 2,671 9 3,250 14 2,171 3,4619 10 1 1 0 0 0 0 1 0 X 4 2,499 10 3,149 15 10,4 2,2929 3,11 11 1 1 1 1 1 1 1 1 0 X 6 2,461 11 3,360 11 2,219 3,6579 12 1 0 0 0 0 0 1 0 1 0 X 18 2,425 12 3,072 19 14,75 3,7429 4,16 13 1 1 1 1 1 1 0 1 1 1 1 X 22 2,379 13 3,130 16 13,7143 2,6952 3,5355 14 1 0 1 1 1 1 1 0 1 1 0 1 X 15 2,343 14 3,112 17 2,3429 3,112 15 1 1 0 1 1 1 1 1 0 1 0 1 1 X 3 2,329 15 3,333 12 15 1,8438 3,5856 16 0 0 0 0 0 0 1 0 0 0 1 1 0 0 X 11 2,293 16 3,110 18 15,75 2,4247 3,0716 18 1 1 0 1 1 1 1 1 0 0 0 1 1 0 0 X 12 2,219 17 3,658 5 3,2143 3,7284 19 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 X 10 2,171 18 3,462 10 2,7524 3,5766 21 1 0 1 1 1 1 1 1 1 1 0 1 1 1 0 1 1 X 1 2,133 19 3,000 20 2,3786 3,1304 22 1 1 0 1 1 1 1 1 0 1 0 1 1 1 0 0 1 1 X 16 1,844 20 3,586 6 12,5714 11,8 10,25 2,7429 3,9048 23 1 0 1 1 1 1 0 1 1 1 1 1 1 1 1 1 0 0 1 X SUPPFEED CI total 190 positive 135 part 71,05% Correlation 0,64827961

Company CI score CI rank SUPPJIT

score SUPPJIT

rank 3 groups 4 groups 5 groups CI

SUPPJI T 1 2 3 4 5 6 7 9 10 11 12 13 14 15 16 18 19 21 22 23 7 3,850 1 2,771 9 2,1333 2,1923 1 X 13 3,743 2 3,091 2 3,1143 3,0435 2 1 X 19 3,214 3 2,814 7 2,3286 2,881 3 1 1 X 2 3,114 4 3,043 3 5,25 2,4993 2,2427 4 1 1 0 X 21 2,752 5 3,402 1 4,4 2,6714 2,5714 5 1 1 0 1 X 23 2,743 6 2,352 16 6,33333 2,4612 2,5645 6 1 1 0 0 1 X 9 2,737 7 3,010 4 3,85 2,7713 7 1 0 0 1 1 1 X 14 2,695 8 2,608 12 8,25 2,7371 3,0097 9 1 1 1 1 1 1 0 X 5 2,671 9 2,571 14 2,171 2,7257 10 1 1 1 0 0 0 1 1 X 4 2,499 10 2,243 18 12,8 2,2929 2,575 11 1 1 1 0 0 0 1 1 0 X 6 2,461 11 2,565 15 2,219 2,7941 12 1 1 1 0 0 0 0 1 1 0 X 18 2,425 12 2,269 17 16 3,7429 3,0909 13 1 1 1 1 1 1 0 1 1 1 1 X 22 2,379 13 2,868 6 12,2857 2,6952 2,608 14 1 1 0 1 1 1 1 1 0 1 0 1 X 15 2,343 14 2,660 11 2,3429 2,6604 15 1 1 0 0 0 0 1 1 0 1 0 1 0 X 3 2,329 15 2,881 5 10,8 1,8438 2,0313 16 1 1 1 1 1 1 1 1 1 1 1 1 1 1 X 11 2,293 16 2,575 13 8,75 2,4247 2,2692 18 1 1 0 0 1 1 1 1 0 0 0 1 1 0 1 X 12 2,219 17 2,794 8 3,2143 2,8145 19 1 0 0 1 1 1 0 0 1 1 1 1 1 1 1 1 X 10 2,171 18 2,726 10 2,7524 3,4023 21 1 0 1 1 1 1 0 1 1 1 1 0 1 1 1 1 0 X 1 2,133 19 2,192 19 2,3786 2,8684 22 1 1 0 0 0 0 0 1 1 1 1 1 0 1 1 0 0 1 X 16 1,844 20 2,031 20 12,2857 14 14,25 2,7429 2,3524 23 1 1 0 1 0 0 1 0 0 0 0 1 0 0 1 1 1 1 0 X CI SUPPJIT total 190 positive 128 part 67,37% Correlation 0,50301955

Company CI score CI rank SUPPDE VT score

SUPPDE

VT rank 3 groups 4 groups 5 groups CI

(37)

37

1.1.2 CI on construct level JIT bundle

Company CI score CI rank CUSTIN

V score CUSTIN

V rank 3 groups 4 groups 5 groups CI

CUSTIN V 1 2 3 4 5 6 7 9 10 11 12 13 14 15 16 18 19 21 22 23 7 3,850 1 3,569 5 2,1333 3,0222 1 X 13 3,743 2 3,561 6 3,1143 3,9375 2 1 X 19 3,214 3 3,557 7 2,3286 3,4767 3 1 1 X 2 3,114 4 3,938 1 4,75 2,4993 2,7557 4 0 1 0 X 21 2,752 5 3,609 4 4,6 2,6714 2,3482 5 0 1 0 0 X 23 2,743 6 3,183 11 5,66667 2,4612 2,916 6 0 1 0 0 0 X 9 2,737 7 3,109 14 3,85 3,5685 7 1 0 1 1 1 1 X 14 2,695 8 3,129 13 10,5 2,7371 3,1091 9 1 1 0 1 1 1 1 X 5 2,671 9 2,348 20 2,171 3,1787 10 1 1 1 0 0 0 1 0 X 4 2,499 10 2,756 18 15,2 2,2929 2,7348 11 0 1 1 1 0 1 1 1 0 X 6 2,461 11 2,916 16 2,219 3,7604 12 1 1 0 0 0 0 0 0 1 0 X 18 2,425 12 3,880 2 14 3,7429 3,561 13 1 0 1 1 1 1 1 1 1 1 0 X 22 2,379 13 3,306 9 13,1429 2,6952 3,1295 14 1 1 0 1 1 1 1 0 0 1 0 1 X 15 2,343 14 2,795 17 2,3429 2,7955 15 0 1 0 0 0 1 1 1 0 1 0 1 1 X 3 2,329 15 3,477 8 10,4 1,8438 3,1877 16 0 1 1 0 0 0 1 0 0 0 1 1 0 0 X 11 2,293 16 2,735 19 13,25 2,4247 3,8797 18 1 1 1 0 0 0 0 0 1 1 1 0 0 1 1 X 12 2,219 17 3,760 3 3,2143 3,5574 19 1 0 1 1 1 1 1 1 1 1 0 1 1 1 1 0 X 10 2,171 18 3,179 12 2,7524 3,609 21 1 1 1 1 1 1 0 1 1 1 0 0 1 1 1 0 0 X 1 2,133 19 3,022 15 2,3786 3,3056 22 1 1 0 0 0 0 1 0 1 1 0 1 0 1 1 1 1 1 X 16 1,844 20 3,188 10 12 11,8 10 2,7429 3,1828 23 1 1 0 1 1 1 1 1 1 1 0 1 1 1 0 0 1 1 0 X CUSTINV CI total 190 positive 117 part 61,58% Correlation 0,34655508

Company CI score CI rank PULL

score PULL

rank 3 groups 4 groups 5 groups CI PULL 1 2 3 4 5 6 7 9 10 11 12 13 14 15 16 18 19 21 22 23

7 3,850 1 1,892 17 2,1333 2,6 1 X 13 3,743 2 4,514 1 3,1143 3,7647 2 1 X 19 3,214 3 3,352 3 2,3286 2,1964 3 0 1 X 2 3,114 4 3,765 2 5,75 2,4993 2,8166 4 1 1 1 X 21 2,752 5 3,227 4 5,4 2,6714 2,5313 5 0 1 1 0 X 23 2,743 6 1,889 18 7,5 2,4612 2,3577 6 0 1 1 1 1 X 9 2,737 7 2,519 12 3,85 1,8919 7 0 0 0 0 0 0 X 14 2,695 8 1,830 19 13,25 2,7371 2,5188 9 0 1 1 0 0 1 0 X 5 2,671 9 2,531 11 2,171 2,4651 10 0 1 0 1 1 0 0 1 X 4 2,499 10 2,817 9 13,8 2,2929 1,9573 11 0 1 1 1 1 1 0 1 0 X 6 2,461 11 2,358 14 2,219 3,1406 12 1 1 0 0 0 0 0 0 1 0 X 18 2,425 12 3,142 5 9,75 3,7429 4,5135 13 1 1 1 1 1 1 0 1 1 1 1 X 22 2,379 13 1,758 20 12,8571 2,6952 1,8303 14 0 1 0 0 0 0 1 1 0 0 0 1 X 15 2,343 14 3,011 8 2,3429 3,0108 15 1 1 1 0 0 0 0 0 1 1 0 1 0 X 3 2,329 15 2,196 15 12,4 1,8438 3,0635 16 0 1 0 0 0 0 0 0 0 0 1 1 0 0 X 11 2,293 16 1,957 16 14,75 2,4247 3,1423 18 1 1 1 0 0 0 0 0 1 1 1 1 0 1 1 X 12 2,219 17 3,141 6 3,2143 3,3517 19 1 0 1 1 1 1 0 1 1 1 1 1 1 1 1 1 X 10 2,171 18 2,465 13 2,7524 3,2273 21 1 1 1 1 1 1 0 1 1 1 1 1 1 1 1 1 1 X 1 2,133 19 2,600 10 2,3786 1,7581 22 0 1 0 1 1 1 1 1 0 0 0 1 1 0 0 1 1 1 X 16 1,844 20 3,064 7 10,7143 10,4 9 2,7429 1,8889 23 0 1 0 0 0 0 1 0 0 0 0 1 1 0 0 0 1 1 1 X CI PULL total 190 positive 107 part 56,32% Correlation 0,28694772

Company CI score CI rank FLOW

score FLOW

rank 3 groups 4 groups 5 groups CI FLOW 1 2 3 4 5 6 7 9 10 11 12 13 14 15 16 18 19 21 22 23

Referenties

GERELATEERDE DOCUMENTEN

genomen houdt bijna de helft van de automobilisten zich niet aan de aangegeven limiet. Met andere woorden, via een 'meer-van-hetzelfde'- aanpak lijkt het buitengewoon

To support this finding, Wottrich and Voorveld (2016) replicated Dahlén’s (2005) study, further confirming the strength of overlapping associations with the creative media

Voor wat betreft de overige aspecten van de publiekrechtelijke zorgplicht en de door de AFM opgelegde regels jegens de geldgever wordt de visie van de AFM en van Bierens dat

This study investigated which factors contribute to building and maintaining a strong relationship between MAs and OMs, with the help of Social Exchange Theory (SET)

- Reduce non-value adding activities Goals • Lower delivery time • Lower inventories at central company • Higher product quality • Higher innovative capability due to

Whereas previous research studied the procedures mainly separately from each other, this study opts to develop an integrated framework for the Continuous Improvement of

Consequently, due to mainly modularity, a flow of common products was established on the shop floor and structure was brought into the processes (e.g. Together with the desire

Aaker and Keller (1992): The effects of sequential introduction of brand extensions?. What is the likelihood that you buy the K-Swiss bags assuming a purchase was planned in