Antero Putkiranta
POSSIBILITIES AND CHALLENGES OF LONGITUDINAL STUDIES IN OPERATIONS MANAGEMENT
Thesis for the degree of Doctor of Science (Technology) to be presented with due permission for public examination and criticism in the Auditorium 2310 at Lappeenranta University of Technology, Lappeenranta, Finland on the 19th of December, 2011, at 10.00.
Supervisor Professor Markku Tuominen Faculty of Technology Management Department of Industrial Management Lappeenranta University of Technology Finland
Reviewers Professor Christopher O’Brien, OBE Nottingham University Business School University of Nottingham
United Kingdom
Emeritus Professor Chris Voss Management Science and Operations London Business School
United Kingdom
Opponent Emeritus Professor Chris Voss Management Science and Operations London Business School
United Kingdom
ISBN 978-952-265-192-1 ISBN 978-952-265-193-8 (PDF) ISSN 1456-4491
Lappeenrannan teknillinen yliopisto Digipaino 2011
ABSTRACT
Antero Putkiranta
POSSIBILITIES AND CHALLENGES OF LONGITUDINAL STUDIES IN OPERATIONS MANAGEMENT
Lappeenranta: 2011 88 p.
Acta Universitatis Lappeenrantaesis 463 Diss. Lappeenranta University of Technology ISBN 978-952-265-192-1
ISBN 978-952-265-193-8 (PDF) ISSN 1456-4491
Longitudinal studies are quite rare in the area of Operations Management. One reason might be the time needed to conduct such studies, and then the lack of experience and real-life examples and results.
The aim of the thesis is to examine longitudinal studies in the area of OM and the possible advantages, challenges and pitfalls of such studies. A longitudinal benchmarking study, Made in Finland, was analyzed in terms of the study methodology and its outcomes.
The timeline of this longitudinal study is interesting. The first study was made in 1993, the second in 2004 and the third in 2010. Between these studies some major changes occurred in the Finnish business environment. Between the first and s econd studies, Finland joined the ETA and the EU, and globalization started with the rise of the Internet era, while between the second and third studies financial turmoil started in 2007.
The sample and cases used in this study w ere originally 23 manuf acturing sites in Finland. These sites were interviewed in 1993, in 2004 and 2010. One important and interesting aspect is that all the original sites participated in 2004, and 19 sites were still able to participate in 2010. Four sites had been closed and/or moved abroad.
All of this gave a good opportunity to study the chan ges that occurred in the F innish manufacturing sites and their environment, and how they reacted to these changes, and the effects on their performance.
It is very seldom, if ever, that the same manufacturing sites have been st udied in a longitudinal setting by using three data points. The results of this study are thus unique, and the experience gained is valuable for practitioners.
Keywords: Longitudinal studies, benchmarking, best practices, operations management,
For my parents,
Maija and Martti
Acknowledgements
The road I have taken to complete this thesis has been long, but fortunately not too rocky. Many people have been involved along the way, directly or indirectly, and I owe them a debt of gratitude. I have been lucky in this respect. I hope I can repay my debt sometime and help others, if needed.
First, I would like to express special thanks to my supervisor, Professor Markku Tuominen, who gave me valuable advice and support during this project. He followed my work with patience and, from the very beginning, showed faith in me. For some reason he ‘bought’ my idea for the thesis during the first phone call I made to him.
I also want to thank Professor Markku Kuula, with whom I have co-authored several papers. He opened the doors to the academic world by introducing me to the right people at the Helsinki School of Economics. And it was he who urged me to contact Professor Tuominen when I first started to speak about my plans. It seems hopeful now that this thesis with four papers will not be the end, but rather the beginning of a series of papers, and a process of getting more used to the world of academic publications.
Only time will tell where all this will lead.
My research colleagues – Jarmo Toivanen, M.Sc (Tech) and Ville Hallavo, M.Sc (Econ) – have been an invaluable help and have inspired and challenged me during the whole project. I am just waiting for them to follow the same path, and hopefully I can encourage them in return.
It is my pleasure to thank the pre-examiners, Professor Chris Voss (London Business School) and Professor Christopher O’Brien (Nottingham University Business School) for their constructive criticism and quick responses. Their theoretical and practical guidance and comments have given me new ideas, which have been extremely valuable in finalizing the thesis. The discussions I have had with Professor Voss at various conferences have helped me to understand the world of publications, and he has always shown interest in my work.
My family, my lovely wife Pirjo and our beautiful daughters Laura and Lotta, have always supported me and been there for me. Their presence shows the real priorities.
After all, a doctoral thesis is just a doctoral thesis, and it would be nothing if it could not be shared with my loved ones. I just hope that I did not ignore them (too much) during the writing process.
Last but not least, I dedicate this thesis to my parents, who always believed in me and were ready to help when needed, and I also thank both of my brothers for their support.
Espoo, December 2011 Antero Putkiranta
Table of Contents
ABSTRACT
ACKNOWLEDGEMENTS TABLE OF CONTENTS LIST OF FIGURES LIST OF TABLES ABBREVIATIONS
1 Introduction ... 16
1.1 Background ... 16
1.2 Results of previous MiE-related studies ... 17
1.3 Objectives of the study ... 21
1.4 Research gap ... 22
1.5 Structure of the thesis ... 23
2 Outline of theories and models used in this study ... 26
2.1 Longitudinal studies ... 27
2.2 Management theories and models used in this study ... 28
2.3 Viewpoints on Operations Management paradigms ... 32
2.4 Benchmarking ... 34
2.5 Viewpoints on performance ... 35
2.6 Best practices... 36
3 Research methodologies of the study... 38
3.1 Research methodologies used in the OM field ... 38
3.2 Key methodologies used in this study ... 43
3.2.1 Longitudinal studies ... 44
3.2.2 Questionnaire ... 45
3.2.3 Data analyses ... 46
3.2.4 Data collection ... 47
3.2.5 The sample... 47
3.2.6 Qualitative and quantitative approach ... 48
3.2.7 Summary of the methodologies used in the study ... 49
4 Summaries of the publications and key results ... 50
4.1 Paper 1 ... 50
4.2 Paper 2 ... 52
4.3 Paper 3 ... 53
5 Discussion ... 57
6 Research and managerial implications ... 60
6.1 Research and managerial implications ... 60
6.1.1 Research implications ... 60
6.1.2 Managerial implications ... 61
6.2 Limitations of the study ... 62
6.3 Validity and reliability of the study ... 63
6.4 Validity and reliability ... 63
7 Conclusions ... 66
7.1 Methodological contribution ... 66
7.2 Empirical contribution of the studies ... 68
7.3 Importance and possibilities of longitudinal studies... 69
7.4 Further studies ... 69
8 Appendix ... 80
APPENDICES
PART II:PUBLICATIONS
List of Figures
Figure 1.1 Timeframe of MiE studies ... 17
Figure 1.2 Main MiE study objectives ... 18
Figure 1.3 Structure of the study ... 24
Figure 2.1 Theories, models and methodologies used in this study ... 26
Figure 2.2 Timeline of some key Management Theories ... 29
Figure 2.3 Growth of a new product (Bass 1969) ... 30
Figure 2.4 Seven S framework (Waterman et al. 1980) ... 31
Figure 2.5 Organization star model (Galbraith 2002) ... 32
Figure 2.6 Operations management paradigms (Voss 1995) ... 32
Figure 2.7 Benchmarking types ... 34
Figure 2.8 Performance levels and views (Slack et al. 2010) ... 36
Figure 3.1 Methodologies used ... 44
Figure 3.2 Question types used ... 45
Figure 3.3 The U test model ... 47
Figure 3.4 The sample ... 48
Figure 3.5 Changes in ownership in the sample ... 48
Figure 4.1 Benchmarking users ... 54
List of Tables
Table 1.1 The main MiE-related studies and their results ... 20
Table 1.2 Objectives of the studies ... 21
Table 1.3 Input/output chart of Part 1 ... 25
Table 3.1 The positivist and phenomenological paradigms (Mangan 2004) ... 39
Table 3.2 Different methodologies used in the positivist and phenomenological paradigms (Mangan 2004) ... 40
Table 3.3 Advantages and disadvantages of rational and case research (Meredith 1998) ... 41
Table 3.4 Case research: Purpose, research question and research structure (modified from Voss 2002) ... 42
Table 3.5 Advantages and disadvantages of choices of case research ... 43
Table 3.6 The question categories used ... 46
Table 3.7 Summary of methodologies used ... 49
Table 4.1 Results from ‘Lessons learned* ... 51
Table 4.2 Key results of the papers ... 56
Table 6.1 Analyses of data quality in MiF longitudinal studies (adapted from Kuula and Putkiranta 2011). ... 61
PART II: PUBLICATIONS:
The thesis is divided into two main parts: The introduction is in the first part, and the publications are in the second part. The publications comprising the second part and summarizing the contribution of the author of this thesis are listed below. The information regarding the status of acceptance for each paper is included also.
The names in the publications listed below are in alphabetical order as also stated in the journals.
Paper 1
Kuula M. and Putkiranta A., (2011) "Longitudinal benchmarking studies in operations management: lessons learned," Benchmarking: an International Journal. Accepted 31 May 2011, forthcoming.
Author’s contribution:
The empirical part – lessons learned – and the idea for the paper was based on the author’s own experience in the overall Made in Finland study project. The literature review, the methodology used in the analyses, and the discussion and conclusion part were done in collaboration with the other author.
Paper 2
Kuula, M., Putkiranta A. and Toivanen, J. (2011), “Coping with change – a longitudinal study of changing manufacturing practices,” International Journal of Operations &
Production Management. Accepted 12 May 2011, forthcoming.
Author’s contribution:
The data was collected solely by the author in 1993 and 2004 , and in collaboration with the other authors in 2010. The model for the data analyses and the SPSS analyses was created by the author. The rest of the publication, introduction, literature review, discussion and conclusions was written in collaboration with the other authors.
Paper 3
Putkiranta, A. (2011), “Benchmarking – a longitudinal study,” Baltic Journal of Management.
Author’s contribution:
The paper was written by the author. The data was based on the same sources as in the previous publications. The research design and the data analyses were done by the author.
Accepted with minor revivions, 04th December 2011.
Paper 4
Hallavo, V., Kuula, M., Putkiranta, A. and Toivanen, J. (2011), “Effects of Plant Ownership Change on Performance and Practices: a longitudinal study.”
The paper was written in collaboration with the other authors and the same database and data collection was used as in the previous publications. The model for the data analyses was created by the author and the data was analyzed by the author using SPSS.
The paper was presented at the 6th International Performance and Management Conference, Nice, France, 6-9 September 2011.
Abbreviations
ABC Activity Based Costing ABM Activity Based Management ATO Assemble-to-Order
BPR Business Process Re-engineering BSC Balanced Score Cards
CE Concurrent Engineering CI Continuous Improvement
CIM Computer Integrated Manufacturing CT Cycle Time
HR Human Resources
IBM International Business Machines JIT Just in Time
MiB Made in Britain studies MiE Made in Europe studies MiF Made in Finland studies
MiF3 Made in Finland, longitudinal study, three data points MRP II Manufacturing Resource Planning
MTO Make-to-Order MTS Make-to-Stock
OM Operations Management Per Performance
Pra Practice
QCA Qualitative Comparative Analyses R&D Research and Development ROA Return on Assets
ROE Return on Equity
SCM Supply Chain Management
SW Software
TQC Total Quality Control TQM Total Quality Management WCM World Class Manufacturing WIP Work in Process
PART 1: OVERVIEW OF THE THESIS
1 Introduction
“The only constant is change”
Heraclitus of Ephesus (c.535 BC - 475 BC)
Longitudinal studies are relatively rare in the area of Operations Management (OM), especially when compared to other sciences such as social sciences or medical sciences, where they are more or less mandatory. Studies in the field of OM are mainly based on large cross-sectional samples and do not follow the same companies or manufacturing sites over time. There are many reasons for this. Maybe the biggest reason is lack of funding for this kind of research. Secondly, this kind of research has no tradition in Operations Management, unlike in other fields. Another reason may be that, as academics are under pressure to publish results increasingly quickly, there is no time for longitudinal studies. However, if such a study is planned with foresight, there may be opportunities to repeat it relatively easily in the future, and thus allow more research and the needed publications.
Longitudinal studies make it possible to study change, and cause and effect. During the last two decades Finland has changed a lot. Finland joined the ETA in 1994 and the EU in 1995, which opened up possibilities to more easily export and import goods and services to and from other European countries. This also allowed the benchmarking of companies and their operations. With open borders, company takeovers became more common and easier, which could be clearly seen in the sample. New practices and new management models took root in Finland, but not only in those companies which were taken over.
1.1 Background
The background to this study is the so-called Made in Europe (MiE) study conducted by Professor Voss of the London Business School. This study was first commissioned by IBM in 1993. In the same year it was exported to Finland and conducted under the name, Made in Finland (MiF), and later, in the years 1994-1995, to many other European countries (Figure 1.1). The results were published in many articles (Table 1.1).
Later, in 2004, the MiF was repeated and thus became a longitudinal study. The manufacturing sites were the same as in 1993 and even some interviewees were the same, although this time in higher positions in their respective companies. In 2010 it was decided to conduct the MiF study once more, and in the same units as previously. It is rare for the same manufacturing sites to be followed twice in a period of ten years. But in this case they were followed on three occasions in 1993, 2004 and 2010. This is exceptional in the field of OM and thus this data is valuable and unique.
1993 1994-1995 2004 2010
Made in England Made in Finland
Made in Europe Made in Finland
- Second round Made in Finland - Third round Figure 1.1 Timeframe of MiE studies
1.2 Results of previous MiE-related studies
The results of previous MiE-related studies can be divided into three main categories. The first category is an analysis of the relationship between practice and operational performance, which was the starting point for the MiE studies. The second category is an analysis and comparison of the state of use of practices and performance based on some European countries, industries, size and ownership. The third category is the use of methodology (Figure 1.2).
Methodological issues
Practice Performance Relationship
Comparisons by
• Country
• Industry
• Size
• Ownership
Main MiE Study Objectives
Figure 1.2 Main MiE study objectives
Based on the MiE studies a wide range of articles has been written in Europe. The main finding has been that there is a clear positive correlation between the adoption of practices and operational performance. This was first shown in the works of Hanson et al. (1995) and later of Voss et al (1995). Putkiranta (2006) indicated the same in his longitudinal study, and also showed that when best practices are less used, operational performance declines. Voss et al.
(1997) noticed that business performance and benchmarking are positively related, and that operational performance and benchmarking are also positively related. In design and manufacturing, there is a relationship between enabling factors (e.g. leadership) and results (Voss et al. 1996).
The second category of results relates to country, industry, size, and ownership comparisons.
The size of the plant has an impact on the practices adopted (Voss et al. 1995). Thus, although small companies have adopted fewer best practices than bigger ones, they perform better.
The parent company origin is also significant. Sites with foreign parents have adopted more best practices and perform better than their domestic counterparts (Voss et al. 1996). World- class companies generally have good practices and good ratings in all areas (Collins et al.
1996).
The methodology has been studied, too. Davies and Kocchar (2002) noted that no detailed research methodology or justification of the model had been published. However, they noted that the advantage of the MiE type of study is that it allows comparisons between a heterogeneous group of companies. The disadvantage is that it does not take into account the unique practices of an industry. Collins et al. (1997) noted, regarding the interviews, that personal interviews give several advantages: higher quality and more consistent data, and elimination of bias. The key publications and their findings are shown in Table 1.1.
Table 1.1 The main MiE-related studies and their results
Authors Year Article, publication Key findings
Hanson P.
Voss C.
1995 “Benchmarking best practice in European manufacturing sites“, Business Process Re-engineering & Management Journal, Vol. 1, No. 1, pp. 60-77.
Adoption of best practices will lead to strong operational performance
ISO9000 is necessary but not sufficient for quality performance
Voss C.
Blackmon K.
Hanson P.
Oak B.
1995 “The Competitiveness of European Manufacturing – A Four Country Study“, Business Strategy Review , Vol. 6, No. 1, pp. 1-25
Adoption of best practices results in strong operational performance
Small companies have adopted fewer best practices than bigger ones, but they perform better
Collins R.
Cordon C.
Julien D.
1996 ”Lessons from the ‘Made in Switzerland’
Study: What Makes a World-Class Manufacturer? “, European Management Journal, Vol. 14, No. 6, 1996, pp. 576- 589
World-class companies have good practices and performance ratings in all areas
Leading companies are market-oriented and concentrate on doing things better
Leading companies have a greater international focus
The inability to implement change fast enough is a significant business vision inhibitor
The origin of the parent company can influence results: domestic companies have inferior results concerning practice and performance
Manufacturers tend to be over-optimistic about their competitive position
Voss C.
Blackmon K.
1996 “The impact of national and parent company origin on world-class manufacturing“, International Journal of Operations & Production Management , Vol. 16, No. 11, 1996, pp. 98-115
Sites with foreign parents have adopted more best practices and have better performance than domestically owned sites
Voss C.
Blackmon K.
Hanson P.
Claxton T.
1996 “Managing New Product Design and Development: an Anglo-German Study“, London Business School, Business Strategy Review, 1996 , Vol. 7, Number 3, pp. 1-15
Strong design practice leads to strong business performance
Four key drivers of manufacturing competitiveness are manufacturing capability, cost, design, investment
In design and manufacturing there is a relationship between enabling factors (e.g.
leadership, people management) and results Collins R. S.
Cordon C.
1997 “Survey methodology issues in manufacturing strategy and practice research’, International Journal of Operations & Production Management, Vol. 17, No. 7, 1997, pp. 697-706.
Personal interviews give several advantages:
higher quality and more consistent data, elimination of blatant bias, seemingly more rigorous preparations
Voss C.
Åhlström P.
Blackmon K.
1997 “Benchmarking and operational performance: some empirical results“, International Journal of Operations &
Production Management, Vol. 17, 1997, No. 1, pp. 1046-1058.
Business performance is positively related to benchmarking
Operational performance is positively related to benchmarking
Over-optimism is negatively related to benchmarking
Over-optimism is negatively related to operational performance
Learning orientation is positively related to benchmarking
Collins R. S.
Cordon C.
Denyse J.
1998 “An empirical test of the rigid flexibility model”, Journal of Operations Management, Vol. 16 (1998), pp. 133- 146.
Inventory management, warehousing, product cycle time and total cycle time depend very strongly on the simplicity dimensions (lean management and lean attitude)
Davies, A.
Kochhar, A.K.
(2002)
2002 “Manufacturing best practice and performance studies: a critique “ International Journal of Operations &
Production Management, Vol. 22, No. 3, 2002, pp. 289-305
The advantage of a Made in Europe type study is that it allows comparisons between a heterogeneous group of companies
The disadvantage is that it does not take into account any unique practices that a company may have and that affect performance significantly
No detailed research methodology or justification of the model has been published
Putkiranta A. 2006 Putkiranta, A. (2006), Industrial Benchmarks: From World Class to Best in Class - Experiences from Finnish Manufacturing at Plant Level, Helsinki School of Economics A-268, Helsinki.
Adoption of practices and improved performance also correlate in the long run
Performance deteriorates if companies do not continue to adopt more practices
1.3 Objectives of the study
The longitudinal data set of manufacturing practices and their outcomes over three periods from 1993 to 2010 presents a unique opportunity to study changes at the factory level over a period of 17 years. The objective of the research in this thesis is first to explore the longitudinal changes in three areas: benchmarking, management practices and the impact of ownership changes. The second objective is to learn from the research and to examine the possibilities of and obstacles to longitudinal research in the area of Operations Management.
The results of this research have been delivered as four separate papers.
The overall methodology used in the longitudinal MiF study is analyzed in Paper 1, and the changes in management practices are studied in Paper 2. Paper 3 concentrates on benchmarking, and Paper 4 focuses on ownership changes. A summary of the objectives of the publications is presented in Table 1.2.
Table 1.2 Objectives of the studies
Paper number Objectives
Paper 1 To explore the possibilities and pitfalls of longitudinal studies, especially in the field of Operations Management and how to design such studies.
Paper 2 To study how manufacturing companies have reacted to the turbulent business environment by changing their improvement and/or management practices.
Paper 3 To study how benchmarking has changed in manufacturing organizations and what effect it has had on operational performance.
Paper 4 To study the changes in ownership and especially the impact these changes have had.
The key driver in all of these papers is change. Lots of changes occurred in the business
changes in their operations and in the ways they approached different markets. Consequently, the companies also changed some the practices they used. By studying these operations and companies longitudinally the effect of the performed changes could be explored.
1.4 Research gap
Longitudinal studies are relatively rare in the area of OM and thus the longitudinal data collected in this research presents a unique opportunity to study OM practices and the process of longitudinal research itself.
Voss et al. (2002) recommended longitudinal studies as one way of doing research in OM. In their book entitled “Researching Operations Management” (2009) Karlsson (ed.) dedicated one chapter to longitudinal methodology. However, no practical experience was reported and thus this first paper is valuable in that it fills the gap by offering conclusions drawn from real life experience, which hopefully will encourage others to conduct longitudinal studies.
Methodological issues are usually discussed at quite a general level or in several different publications, but the first paper reports the main ‘lessons learned’ regarding research using a standard questionnaire and face-to-face interviews.
Normally, researchers try to understand and explain phenomena on the basis of quantitative or qualitative cross-sectional studies. But, as discussed in these papers, changes occur over time, and very few researchers have tried to explain phenomena by using longitudinal data in the area of OM. Especially in situations where company environments change dramatically, – as they did when Finland joined the ETA and later the EU, or when new business models arose through the influence of the internet, or when a global financial crisis affected all companies, – longitudinal research affords an excellent opportunity to study how these companies reacted to the changes and what models they adopted to survive or make the best of the new situation.
Schönberger (2001) postulated that there is a life cycle for management practices, which is ever-shortening. This life cycle has not been tested before with new practices and longitudinal data. The longitudinal setting in this research has given a good opportunity to study the behavior of various practices used in the companies and to test the behavior of these practices by using models such as the Bass diffusion model, which require longitudinal data.
While the effect of benchmarking has been tested in many studies (e.g. Voss et al. 1997), the use of benchmarking has not been so widely studied, and especially the effect of changes in use. Nor has the effect of benchmarking on different performance classes (absolute or relative performance) been studied. This research gives an indication of how the use of benchmarking has developed over time, what the current focus of benchmarking is, and what the effect of benchmarking is in the various performance categories.
Ownership and parent company origin have been widely studied, even in the MiE cases (Voss et al. 1996). However, it has been more difficult to study ownership change and its effect on the practices adopted and on performance. A longitudinal approach following the same companies over time provides a unique opportunity to study the effect of ownership change.
In a single case study, it is obviously impossible to compare different companies, but in this study, with an original sample of 23 companies, it is possible to compare the acquired companies with those that remained under the old ownership.
1.5 Structure of the thesis
This thesis is divided into two parts. In the first part (Part 1) an overview is given of the publications and in the second part (Part 2) the publications are presented in their published form. The structure of the first part can be seen in Figure 1.3. In the first chapter the subject is introduced and background information is given. The linkages to Made-in-Europe-based studies and other studies are shown and the results of these studies are given. The objectives of the study are shown with the research gap and research questions, and finally the structure of the thesis is presented.
The second chapter contains a general discussion on the models and theories used in the study. It first deals with longitudinal studies, which are the key theme of the thesis, and goes on to discuss general management theories and phases of operations management, and finally benchmarking. The third chapter focuses on the research methodologies used in the study, starting with the general research paradigms used in operations management. This is followed by a discussion of the motivation behind the research strategy used in this study – the longitudinal setting, why case research methodology, the questionnaire used, the data
The fourth chapter summarizes the publications in terms of their purpose and objectives, the methodologies used, and the key results. The fifth chapter concludes Part 1 by discussing the results in a wider context, offering some conclusions and pointing out the managerial and research implications. Finally the limitations of the study are discussed and ideas are suggested for future research.
Pa
POSSIBILITIES AND CHALLENGES OF LONGITUDINAL STUDIES IN OPERATIONS MANAGEMENT
Part 1: Introductory part Part 2: Articles
Longitudinal benchmarking studies in operations management: lessons learned
Coping with the change – a longitudinal study into
the changing manufacturing
practices
Benchmarking – a longitudinal study
Effects of Ownership Change on Performance
and Practices:
a longitudinal study
Figure 1.3 Structure of the study
An input/output chart of the chapters is shown in (Table 3.1).
Table 1.3 Input/output chart of Part 1
Input Chapter Output
- Background of the study - Motives of the study
1.Introduction - Objectives of the study - Research questions - Structure of the study - Longitudinal studies
- Management theories - OM paradigms - Benchmarking - Performance - Best practices
2. Outline of models and theories - Overview of the studies in the field
- Viewpoint to performance and practices
- Various research methodologies - The case study approach - Questionnaire used - Analysis tools of small
samples
3. Research methodology - The research process - Arguments for the chosen
methods
- Motivation for analysis tools
- Publications presented in Part 2
4.Summary of the cases and main results
- Description of the cases - Review of the results - Results of the papers 5. Discussion - Interpretation of the results in
wider context - Results of the papers 6. Research and managerial
implications, limitations of the study, and validity and reliability of the results
- Implications for researchers and managers in the field - Reliability aspects of the
results - Results of the papers 7. Conclusions, contribution
from methodological and empirical point of view, and further research
- Methodological contribution - Empirical contribution - Avenues and ideas for future
research
2 Outline of theories and models used in this study
Various theoretical models have been used in the publications to analyze the changes occurring between the studies of 1993 and 2010. These models are basic OM research methodologies, benchmarking models, management and organizational theories, and operations management models. A summary of the methodologies used in the publications is given in Figure 2.1.
Benchmarking Longitudinal studies
Bass diffusion model
Benchmarking Longitudinal studies
Organizational Star model (Galbraith 2002)
Seven S (Waterman et al (1980)
Paper 1 Paper 2 Paper 3
Paper 4
Figure 2.1 Theories, models and methodologies used in this study
In the first paper (Paper 1) the input was the author’s experience and the lessons learned in the Made in Finland research project. The output of the paper was a model for analyzing the research methodology used in the MiF studies, as well as the lessons learned and guidance on how to conduct such studies and what to take into consideration when planning longitudinal studies, or indeed, when planning a cross-sectional study that might become a longitudinal study.
In the second paper the objective was to find out whether the practices follow a pattern in their usage. This was analyzed by comparing the results to the curve shown in the Bass
diffusion model which is used in other research fields. The input was the selected practices and their usage over the years, and the output was the behavior curve of these practices.
In the third paper the input was the data gathered in the MiF studies. The objective was to study the use of benchmarking and how it has changed over time and what effect it has had on absolute and relational performance. The data used was the benchmarking question used in the study, and the output was whether benchmarking has an effect on these performance categories. The test used to analyze the results was the Mann-Whitney test (U-test), which is a non-parametric test for small samples.
In the fourth paper the input was also the data gathered in the MiF studies. The objective was to analyze the effect of changes in ownership. Thus, the framework or models chosen for analyzing the data were picked from the management area – the Organizational Star model by Galbraith (2002) and the Seven S model by Waterman et al. (1980). They motivated the creation of a suitable framework for categorizing and choosing the used practices from the MiF questionnaire. The output was whether there was any difference in absolute or relational performance between the acquired companies and those that remained in the same ownership.
2.1 Longitudinal studies
Longitudinal studies are quite rare in the area of Operations Management. In other fields they are more common and even almost mandatory, as in the social sciences where objects are followed over time. However, the benefits of longitudinal studies are clear. The main benefits of such studies are that they can reveal causal relations, give alternative explanations for events and results, and uncover areas for research (Bynner and Joshi 2007, Rindfleich et al.
2008, Handfield and Melnyk, 1998). However, longitudinal studies are not suitable for theory testing, according to Åhlström and Karlsson (2009).
What makes a longitudinal study? According to Åhlström and Karlsson (2009), phenomena are studied over time in longitudinal studies. But how many data points are needed? Or how long should one study the phenomena? Obviously at least two data points are needed, but is that enough? According to Ruspini (1999) and Menard (2002), the data is collected for two or more periods. However, Åhlström and Karlsson (2009) questioned whether two data sets are
There are various challenges in longitudinal studies. First they are time-consuming (Voss 2002) and they are very complex to perform (Ruspini 2002). Data quality is also an issue.
Data sampling, and its reliability and validity, are always key challenges in research, and are especially problematic with longitudinal data (Magnusson and Bergaman 1991).
Documentation is always crucial, but with longitudinal studies this factor is even more crucial (Ruspini 2002, Åhlström and Karlsson 2009).
2.2 Management theories and models used in this study
The history of management paradigms is fairly long with many different aspects and viewpoints. In the Harvard Business Review’s (HBR) 75th anniversary issue, Sibbet (1997) summarized the key time-line of management ideas and practices, using the key publications and articles of HBR, starting from 1922. The categorization Sibbet used was: Scientific Management, Government Regulation, Marketing and Diversification, Strategy and Social Change, Competitive Challenge and Restructuring, and Globalization and Knowledge.
Kuokkanen et al. (2010) also used management paradigms to conceptualize the history of management, but from a somewhat different angle than Sibbet in HBR. They used five names for their modeling: scientific management, human relationship theories, structural analyses, organizational culture, and the innovation paradigm (Figure 2.2).
Scientific management (also called Taylorism) started as long ago as the early 1900s and was launched by Frederick Taylor. The focus was on work flows and applying science in factories to get the best possible measurable results (e.g. Taylor 1911).
Human relationship theories by Mayo (1933) focused on the relationship between work conditions and employees and performance. An employee should belong to a group, and this need or incentive is more important than monetary incentives or working conditions.
Structural analyses studied the business environment and the changes that occurred after the 1950s and their relationship to the organizational structures of companies and their strategies.
Hamel and Prahaland (1990) and Porter (1980) continued this and developed the structural analyses to strategic dimensions.
The idea of an Organizational culture paradigm at the beginning of the 1970s was that the opinion of workers should be better taken into account when making management decisions.
The Japanese work culture was seen as a model example of this (e.g. Morgan 1997).
The research focus of Innovation Paradigm was the ever-increasing pace at which products and services are brought into the marketplace to satisfy the differing and fast-changing needs of customers, and how these needs were created in many cases.
Organizational Culture Paradigm Scientific Management
Human Relation Theories Structural Analyses
Innovation Paradigm
1900 1930 1950 1970 1990
Figure 2.2 Timeline of some key Management Theories
Not only have the management paradigms changed, but the overall operating environments of companies have also been in turmoil. With globalization and financial imperatives, the need for change has been evident. Hayes et al. (2005) made a comparison between the ‘old economy’ and the ‘new economy’ as they called it, and the key parameters, and how they have changed. The main parameters were the unit of analysis, the overall goal, the primary measures of performance, the dominant OM activity, the OM tools, the domain of OM, and the competitive imperative. The main difference is that the unit in the new economy is the overall supply chain – the network of independent players. The goal is to create a relationship with all the players in the supply chain, unlike earlier when there was only the vendor and the customer. The difficulty lies in managing the overall ever-changing network, and at the same time getting the product to the market faster and faster. Tools and methodologies change with the network, and the life cycle of used practices is ever shortening, as Schoneberger (2001) noted.
timing of a consumer’s initial purchase is related to the number of previous buyers”. Bass named five types of consumer groups based on the buying behavior, or rather when the buying occurs: innovators, early adopters, early majority, late majority, and laggards. At the beginning the growth of a new product is quite slow as only the innovators buy the product, followed later by the early adopters. The real purchase growth starts with the early majority and later with the late majority. The interest of the model lies in the forecasting of the sales peak and the magnitude of the sales. The model is widely used in marketing, but the application goes beyond marketing.
Number adopting the innovation
Elapsed time from launch to complete diffusion Innovators
Early adopters
Early
majority Late majority
Laggards
Figure 2.3 Growth of a new product (Bass 1969)
Waterman et al. (1980) introduced the seven S framework with the idea that certain parts of a company should work together well in order to be effective and to help create an effective and working entity (Figure 2.4). There are always several factors that affect the organization’s ability to change and these factors should not be forgotten. A change made in one part of the organization or in one process inevitably affects some other parts of the organization and processes.
The seven parts they named are:
Structure: The structure of the company, e.g. centralized or decentralized, organization form e.g. matrix, etc.
Strategy: All the actions the company takes in order to win customers.
Systems: All the formal and informal procedures to make the organization go.
Style: The way the top management comes across to the organization, and how and where they spend their time.
Staff: The attitude, morale and motivation of the staff, as well as pay scales and appraisal systems.
Skills: The various skills needed in the market place and how flexible the company is to change the needed skills.
Superordinate
goals: A set of values and basic beliefs the business is originally built around.
The Seven S framework has become a powerful tool to analyze the organization’s ability to deliver and implement strategy.
Superordinate goals Structure
Staff Skills
Strategy Systems
Style
Figure 2.4 Seven S framework (Waterman et al. 1980)
Galbraith (2002) argued that there has to be a balance between the strategy, structure, processes, people, and reward systems (Figure 2.5). If vision and strategy changes, as it does quite often in the new economy, it is bound to have an effect on the organizational structure, skills and type of people needed, the delivery process and other processes in the company and outside the company, and the reward and measurement systems.
Strategy
Structure
Processes People
Rewards
Figure 2.5 Organization star model (Galbraith 2002)
2.3 Viewpoints on Operations Management paradigms
In his papers “Alternative paradigms for manufacturing strategy” (1995) and “Paradigms of manufacturing revisited” (2006), Voss proposed three key paradigms in the area of OM.
These paradigms are: Competing through manufacturing, Strategic choices in manufacturing strategy, and Best practices (Figure 2.6).
Strategic choices in manufacturing Competing through
manufactturing
Best practice
Figure 2.6 Operations management paradigms (Voss 1995)
Competing through manufacturing refers to the capabilities of the company and these capabilities should be aligned with key success factors (Hill 2000) in order to obtain the best performance in the marketplace.
Strategic choices in manufacturing strategy refer to the decisions made regarding e.g. choice of process, equipment, labor force and locations. Hayes and Wheelwright (1979a, 1979b) argued that the company should closely follow the volumes, and align the production process with these volumes. There always seems to be an optimum production performance.
Best practices are usually used closely with World Class Manufacturing. Manufacturing practices such as JIT, lean, TQM, ABC costing, and Concurrent Engineering are usually referred to as best practices.
2.4 Benchmarking
The objective of benchmarking is to stimulate activity and creativeness (Slack et al. 2010).
According to Stevenson (2009) the objective is to find standards against which performance can then be measured. Fitzsimmons and Fitzsimmons (2010) defined benchmarking in the following way: the “practice of comparing one’s performance with that of other firms that are known as best in class”. Slack wrote “in order to learn from them and/or assess performance”.
Benchmarking can be categorized and used in many different ways depending on the purpose and possibilities for benchmarking. Prasnikar et al. (2005) proposed a four-category division for benchmarking: benchmarking of competitive advantages, benchmarking of strategies, process benchmarking, and performance benchmarking. Slack et al. (2010) divided benchmarking into external and internal benchmarking, competitive and non-competitive benchmarking, performance and practice benchmarking. Rigby took up the question of product benchmarking (2006). The various benchmarking types and the use of the types are illustrated in Figure 2.7.
Product benchmarking External
benchmarking
Competitive benchmarking
Process and performance Benchmarking Competitive advantages
Benchmarking of strategies
Non-competitive benchmarking
Internal benchmarking
Figure 2.7 Benchmarking types
Internal benchmarking is used for benchmarking performance between various units and especially in cases where there are similar processes in different locations. Internal benchmarking reveals the differences in processes and it is quite easy to find the reasons for performance differences. External benchmarking is more for the differences between products e.g. prices and features, and strategies. Competitive advantages can be found also with external benchmarking. With non-competitive benchmarking the best-in-class processes and performance figures can be found. It is hard to get this information from the competition. By using competitive benchmarking, it is possible to get ideas for competitive advantages as well as new innovative strategies. Competitive benchmarking can reveal the differences between products e.g. pricing and channels, and strategies.
Benchmarking must go beyond the figures in order to provide real results and useful data (Stauffer 2003). However, in many cases the information received in benchmarking is difficult to use (Tucker et al. 1987). The figures may be exact, but the difficulty lies in the processes – how to reach the figures. Competitive benchmarking is especially difficult in this area. Non-competitive benchmarking is easier to perform and people tend to be more receptive towards these ideas. This also allows the choosing of real world-class companies.
When performing benchmarking and analyzing the results, the time needed to implement some practices – when the output of these practices is available – should also be taken into account (Davies and Kochhar 2002). One of the recent studies in benchmarking (Adebanjo et al. 2010) shows that the use of benchmarking is still strong. It is used and will continue to be used as an important tool to support organizations.
2.5 Viewpoints on performance
Hayes et al. (2005) noted that “in a competitive environment a company’s competitive effectiveness (lower cost, better quality, more features, longer product life, etc...) must improve over time”. This indicates one aspect of competitive performance – performance in relation to others. Similarly, Slack et al. (2010) divided performance into external and absolute performance, strategic targets and historically based targets. There are several viewpoints on performance in operations. Slack et al. (2010) mentioned quality, speed, dependability, flexibility, and cost. Hudson et al. (2001) focused on quality, time, flexibility,
The outside view on performance is typically quality, lead times, reliability (of the product and of the delivery), while the internal view is internal quality, throughput times and utilization rates, and costs (Slack et al. 2010). At different organizational levels the same performance can be viewed differently (Figure 2.8).
Broad Overall High
strategic strategic relevance and
measures objectives aggregation
Functional strategic Market Operations Financial
measures strategic strategic strategic
objectives objectives objectives Composite
performance Customer Agility Resilience
measures satisfaction
Generic operations Quality Dependability Speed Flexibility Cost measures
Examples of Defect per Mean time Customer Time to Transaction
detailed unit between query time market costs High
performance Level of Lateness Order lead Product Labour diagnostic
measures customer complaints time range productivity power and
complaints Throughput Machine frequency of
Scrap level time efficiency measurement
Figure 2.8 Performance levels and views (Slack et al. 2010)
At the operations level or plant level, performance is viewed in a more detailed way, and the same performance is ‘aggregated’ to strategic-level targets in other functions. One very important aspect of performance is that all the different performance metrics at various levels must be aligned to the same end results, as introduced in the concept of Balanced Score Cards and Strategic Maps by Kaplan and Norton (1992, 2003).
2.6 Best practices
As mentioned earlier in section 2.3 best practices are part of the Operations Management Paradigms proposed by Voss (1995, 2006). Practices such as Just in Time (JIT), Lean, Total Quality Management (TQM), and Activity Based Costing (ABC) are normally referred to as
‘best practices’.
Slack et al. (2010) noted that best practice analyses can identify the gaps between the best benchmarked performance and the company’s performance. According to them a best practice has the following characteristics:
is current – not outdated
is structured – clear goals, scope and process is proven – success has been demonstrated
is repeatable – has been used and proved in various contexts
has an unambiguous method – it can be connected to business processes, operations strategy, technology, supply relations, and ICT systems
has a positive impact on results.
The implementation of best practices should lead to improved operational performance, as shown in the studies by Hanson et al. (1994), Hanson and Voss (1995), Collins et al (1995), Voss et al (1997). The results of Putkiranta (2006) also indicated that this linkage is valid longitudinally, and showed that, when the use of best practices or the level of usage decreases, this has a negative impact on operational performance, or, at best, operational performance does not improve. However, the best practices used in companies do have a life span, which is shortening, as noted by Schönberger (2001).
3 Research methodologies of the study
This chapter discusses the research methodologies used in OM from various angles and especially in the context of this study.
3.1 Research methodologies used in the OM field
Hanfiled and Melnyk (1998) noted concerning OM research that “its basic aim is not to create theory, but to create scientific knowledge”. This scientific knowledge is (Reynolds 1971):
a method of organizing and categorizing things prediction of future events
explanations of past events
a sense of understanding about what causes events the potential for controlling events.
In the area of OM a variety of research methodologies are used. Groom (2009) stated that these methodologies can be categorized in many ways and they depend on the phenomena in hand and what kind of research strategy will be used. Sometimes the strategy used depends on the continent (Drejer et al. 2000), such as the fact that case study methodology is more common in Europe, and quantitative studies are more common in the U.S.
In Quantitative Research the emphasis is on mathematical and statistical tools. It is associated with logical positivism (e.g. Croom2009, Golafshani 2003). In quantitative research a large amount of data is used to test hypotheses and causal relations between variables.
Qualitative Research tries to understand the phenomena in its natural setting and environment (e.g. Golafshani 2003, Croom 2009). Broadly defined qualitative research means all the research that does not use statistical or quantification methods (Strauss and Corbin, 1990).
Croom (2009) stated that “Qualitative approaches are .. at their extreme concerned with constructivism, interpretation and perception, rather than with identification of a rational, objective truth.”
One division is the Positivist paradigm and the Phenomenological paradigm. Managan (2004) summarized and explained the key characteristics of these two paradigms by the basic belief concerning what a researcher should do, and what the preferred methods are, as seen in Table 3.1.
Table 3.1 The positivist and phenomenological paradigms (Mangan 2004)
Positivist paradigm Phenomenological paradigm Basic belief The world is external
Observers are independent Science is value-free
The world is subjective and socially constructed
The researcher should Focus on facts
Look for causality and fundamental laws
Reduce phenomena to the simplest events
Formulate hypotheses and test them
Focus on meanings
Try to understand what is happening
Look at the totality of each situation
Develop ideas through induction from the data
The preferred methods include
Operationalizing concepts so they can be measured Taking large samples
Using multiple methods to establish different views of phenomena
Small samples investigated in-depth over time
There are many different methodologies available for researchers in both paradigms as summarized in Table 3.1 (Mangan 2004) and some methodologies can be used under both paradigms.
Table 3.2 Different methodologies used in the positivist and phenomenological paradigms (Mangan 2004)
Positivist Paradigm Phenomenal Paradigm
Cross-sectional studies Experimental studies Longitudinal studies Surveys
Models and simulation
Action research Case studies Ethnography Construct elicitation Grounded theory Hermeneutics Participative enquiry
Another way of dividing research is rational and case research. Meredith analyzed these two types of research in his article (1998). The basic belief of rational research is that “the phenomenon studied exists out there’” as he put it. Some typical models in this kind of research are production and inventory simulation models and similar mathematical models.
Case research uses multiple tools and methods in the natural setting of the phenomenon. Both quantitative and qualitative tools can be used. The goal is to understand the phenomena as well and as thoroughly as possible. Both of these two methods have their advantages and disadvantages. Table 3.3 illustrates the advantages and disadvantages of these two types of research. According to Meredith, rationalist methods are more suitable for testing or verifying theories, whereas case methodologies are better for modifying, extending or even creating theories.
Table 3.3 Advantages and disadvantages of rational and case research (Meredith 1998)
Advantages Disadvantages
Rationalist Precision
Reliability
Standard procedures Testability
Sample difficulties Trivial data Model-limited
Low explained variance Variable restrictions Thin results
Case Relevance
Understanding Exploratory depth
Access and time
Triangulation requirements Lack of controls
Unfamiliarity of procedures
Regarding triangulation, Hussey and Hussey (1997) stated that it can overcome the potential bias of the single methods approach. Easterby-Smith et al. (1991) categorized triangulation in four different categories:
1. Data triangulation: data is collected at different times or from different sources 2. Investigator triangulation: different investigators collect the data independently from
each other
3. Methodological triangulation: both qualitative and quantitative methods and techniques are used
4. Triangulation of theories: a theory is taken from other disciplines and applied and used to explain a phenomenon in another discipline.
According to Voss et al. (2002) case research has become “one of the most powerful research methods in operations management”. However, pure case research is used more in Europe than in the U.S. Case research can be used for many different types of purposes. These different purposes and their key characteristics are summarized in Table 3.4.
Table 3.4 Case research: Purpose, research question and research structure (modified from Voss 2002)
Purpose Research question Research structure
Exploration
Uncover areas for research and theory development
Is there something interesting to justify the research?
In-depth case studies
Unfocused, longitudinal field studies
Theory building
Identify/describe key variables
Identify linkages between variables
Identify why these linkages exist
What are the key variables?
What are the patterns or linkages of these variables?
Why should these relationships exist?
Few focused case studies In-depth field studies Multi-site case studies Best-in-class studies
Theory testing
Test the theories developed in the previous stages Predict future outcomes
Are the theories we have generated able to survive the testing of empirical data?
Did we get the behavior that was predicted by the theory, or did we observe another unanticipated behavior?
Experiment Quasi experiment Multiple case studies Large sample of population
Theory extension/refinement To better structure the theories in light of the observed results
How generalizable is the theory?
Where does the theory apply?
Experiment Quasi experiment Case studies
Large-scale sample of population
Case selection and sampling in case research are always vital questions. One way of seeing the alternatives (Voss 2002) are single cases or multiple cases, and retrospective cases or longitudinal data. The advantages and disadvantages of these choices are summarized in Table 3.5..
Table 3.5 Advantages and disadvantages of choices of case research
Choice Advantage Disadvantage
Single cases Greater depth Limits on the generalizability of conclusions drawn.
Biases such as misjudging representatives of a single event and exaggerating easily available data
Multiple cases Augment external validity Help guard against observer bias
More resources needed Less depth per case Retrospective cases Allow collection of data on
historical events
May be difficult to determine cause and effect
Participants may not recall important events
Longitudinal cases Overcome the problems of retrospective cases
Have long elapsed time and thus may be difficult to do
3.2 Key methodologies used in this study
The key methodologies used in the papers can be divided in several ways (Figure 3.1). The study is both qualitative and quantitative by nature. The first publication is purely qualitative and reports the lessons learned from the MiF longitudinal study. The other publications are partly qualitative and partly quantitative, although the sample does not allow statistical analyses that are too far-reaching and unequivocal. The data was collected by using a standard questionnaire in all three surveys. This study is thus longitudinal and has three data points.
The data was analyzed by using statistical methods suitable for small samples.
Questionnaire Data collection
Data analyses
Longitudinal study Qualitative and
Quantitative
Figure 3.1 Methodologies used
Overall, this study uses multiple methods, depending on the problem in hand. The linking factor is the longitudinal data and the questionnaire used.
3.2.1 Longitudinal studies
Regarding data collection, the methodology is longitudinal studies. Longitudinal studies are suitable (e.g. Voss et al. 2002, Bynner and Joshi 2007):
- in relations between cause and effect - when searching for alternative explanations - in new areas of research.
However, Åhlström and Karlsson argue that these methods are not suitable for theory building and theory testing (Åhlström and Karllson 2009).
What makes a longitudinal study? There is no clear answer to that, but neither is there any agreement on how long the phenomena should be studied. In a longitudinal study at least two measurement points are needed. Obviously, however, more data points reveal more – at least that was the case in these studies in Publication 2 (Kuula et al. 2011).
Many of the studies within the area of OM are snapshots of one point in time, and are cross- sectional or study a single industry or line of business. This is the case with the MiE studies, which are cross-sectional and have only one data point. In this study the data was collected at three different data points (1993, 2004, and 2010). The data was collected by using a structured standard questionnaire and interviewing people. The interviews lasted for about 1.5 hours to 3 hours. At the same time some qualitative data was collected.
3.2.2 Questionnaire
The original questionnaire used was based on the MiE study. It contained 46 questions using Likert scaling from 1 to 5. On top of that, five open questions were asked regarding business visions, business vision inhibitors, and sources of advice, ability to compete, and where they can compete. In 2004, 29 questions were added to better classify the companies in the area of production and strategic intentions (Figure 3.2). The Questionnaire used is shown in the Appendix .
Performance Data
Data of Practices used
Production Structure Strategic
Intentions
Figure 3.2 Question types used
To make it easier to answer the questions, the Likert scaling had verbal indicators showing what kind of result scores 1, 3, or 5. The new questions asked were only classification questions. Even though some of the data asked for was performance-related data, it was regarded as ordinal data, not interval data, on the Likert scale. The new questions asked were only for classification purposes, even though the scaling/numbering was the same.
The questions used in the questionnaire were divided into various subcategories based on their relevance and level, as discussed in sections 2.5 and 2.6. The categories used were in the following performance categories:
competitive performance absolute performance
