The role of contextual factors in process harmonization · Javier, Francisco Javier, Manuel Jos e,...

The role of contextual factors in process harmonization

Citation for published version (APA):Romero, H. L. (2014). The role of contextual factors in process harmonization. Technische UniversiteitEindhoven. https://doi.org/10.6100/IR773251

DOI:10.6100/IR773251

Document status and date:Published: 01/01/2014

Document Version:Publisher’s PDF, also known as Version of Record (includes final page, issue and volume numbers)

Please check the document version of this publication:

• A submitted manuscript is the version of the article upon submission and before peer-review. There can beimportant differences between the submitted version and the official published version of record. Peopleinterested in the research are advised to contact the author for the final version of the publication, or visit theDOI to the publisher's website.• The final author version and the galley proof are versions of the publication after peer review.• The final published version features the final layout of the paper including the volume, issue and pagenumbers.Link to publication

General rightsCopyright and moral rights for the publications made accessible in the public portal are retained by the authors and/or other copyright ownersand it is a condition of accessing publications that users recognise and abide by the legal requirements associated with these rights.

• Users may download and print one copy of any publication from the public portal for the purpose of private study or research. • You may not further distribute the material or use it for any profit-making activity or commercial gain • You may freely distribute the URL identifying the publication in the public portal.

If the publication is distributed under the terms of Article 25fa of the Dutch Copyright Act, indicated by the “Taverne” license above, pleasefollow below link for the End User Agreement:www.tue.nl/taverne

Take down policyIf you believe that this document breaches copyright please contact us at:[email protected] details and we will investigate your claim.

Download date: 22. Mar. 2021

https://doi.org/10.6100/IR773251

https://doi.org/10.6100/IR773251

https://research.tue.nl/en/publications/the-role-of-contextual-factors-in-process-harmonization(822b050b-ba77-4185-b015-de8fcce7046d).html

The Role of Contextual Factors in Process Harmonization

A catalogue record is available from the Eindhoven University of TechnologyLibrary

ISBN: 978-90-8891-854-4

This thesis is number D179 of the thesis series of Beta Research School forOperations Management and Logistics.

Printed by Proefschriftmaken.nl ‖ Uitgeverij BOXPressCover design: Dayana Castro Vargas

Copyright c© 2014 by Heidi Lızabeth Romero. All Rights Reserved.


PROEFSCHRIFT

ter verkrijging van de graad van doctor aan deTechnische Universiteit Eindhoven, op gezag van de

rector magnificus, prof.dr.ir. C.J. van Duijn, voor eencommissie aangewezen door het College voor

Promoties in het openbaar te verdedigenop maandag 28 april 2014 om 16:00 uur

door

Heidi Lızabeth Romero

geboren te Santo Domingo, Dominicaanse Republiek

Dit proefschrift is goedgekeurd door de promotoren en de samenstelling van de pro-motiecommissie is als volgt:

voorzitter: prof.dr. A.G.L. Romme1e promotor: prof.dr.ir. P.W.P.J. Grefen2e promotor: prof.dr. A.J. van Weelecopromotor: dr.ir. R.M. Dijkmanleden: prof.dr. R.J. Wieringa (Universiteit Twente)

prof.dr. J. Becker (Westfalische Wilhems-Universitat Munster)dr. A. de Jongprof.dr.habil H. Schiele (Universiteit Twente)

Acknowledgements

I believe in the richness of diversity. Therefore, I feel privileged to have shared theseyears of my life with so many people from different countries, with different back-grounds, who have helped me to make this project a reality. I would like to express mygratitude to those who helped me get started down the field, those who got me overthe goal line, and those who cheered me on throughout the game.

I am grateful to my promotors Paul Grefen and Arjan van Weele. I want to thankboth for the opportunity to pursue my PhD degree. I want to thank Paul for hisconstructive and critical feedback during the project, and for challenging me to lookout of the box. I would also like to thank Arjan, for setting high standards and forkeeping an eye on the practical implications and the relevance of my research.

I am very privileged to have Remco Dijkman as my daily supervisor. He did nottry to impose his ideas, but respectfully encouraged me to navigate different options.I would like to thank Remco for his valuable advice about the method and structureof my work. During these years Remco has gained my respect for his high qualities asa person, which are shown in simple daily actions. Thanks again for all your patienceand support.

I want to thank the members of my promotion committee: professor Wieringa,professor Schiele and professor Becker for your careful examination of my thesis. Es-pecially, I would like to thank Ad de Jong, who thoroughly revised the methodologyfollowed in the chapter three.

I would like to present my sincere gratitude to the people, companies and profes-sional associations that were directly and indirectly involved in the realization of myresearch. The European Supply Chain Forum (ESCF), which co-funded my researchproject. For my data collection, I received support from the Business Process Man-agement (BPM) Round Table and six other companies that I must keep anonymous.Thanks to my colleagues from the Beta PhD Council and the Student Supply ChainForum.

I have received support from many people from the different groups of IndustrialEngineering and Innovation Sciences department. My sincere thanks to Jolanda, Nicoand Christel for their support and friendship. Especially, I would like to thank allmy current and former colleagues at Information Systems (IS) group for the goodatmosphere and nice conversations during lunch: Rob, Mohamad, Samaneh, Zhiqiang,Jana, Joel, Egon, Mariska, Ronny, Daniel, Ricardo, Marco, Ton, Jan, Samuil, Vassil,Rik, Irene, Pieter, Hajo, Ada, Annemarie, Jos, Rob, Claudia, Uzay, Harold, Oktay,Rui and Anna. In particular Shaya, Shan and Hui for your help in the cover design.Thanks for sharing nice memories during these years.

Now is the time to express my gratitude to the people who cheered me on throughout

vi Acknowledgements

this journey, my family and friends. My world is enriched with many inspiring andcaring people. I am indebted to my parents, Marıtza and Jose, for their unconditionallove and support, your interest in my work and well-being, you are an exemplar to me.Thanks to my dearest sister and friend Annie-Belle and my brother Jose Francisco, thebest that one could wish for. I would like to thank my love and partner, Anthony, forhis understanding, patience, encouragement and care during the last ten years. Also tomy family-in-law for their support. It would never be the same without all of you. AndI want to dedicate these final lines to my beloved son Alexander, and my niblings JoseJavier, Francisco Javier, Manuel Jose, Rafael Adrian, Jose Alejandro y Eva Alejandra.Quiero que esta tesis les sirva de inspiracion para que siempre luchen por alcanzar sussuenos.

Last but not least, to God for his daily blessings provided. Thanks for giving methe opportunity to share my life with all the people mentioned before.

Contents

Acknowledgements v

1 Introduction 11.1 Problem statement . . . . . . . . . . . . . . . . 21.2 Theoretical foundation . . . . . . . . . . . . . . . 3

1.2.1 Contextual factors . . . . . . . . . . . . . . 31.2.2 Process harmonization . . . . . . . . . . . . . 41.2.3 Business performance . . . . . . . . . . . . . 5

1.3 Research goal and research questions . . . . . . . . . . 61.4 Research approach . . . . . . . . . . . . . . . . 71.5 Outline . . . . . . . . . . . . . . . . . . . . 7

2 Conceptual model 112.1 Methodology . . . . . . . . . . . . . . . . . . 12

2.1.1 Problem formulation . . . . . . . . . . . . . 122.1.2 Data collection . . . . . . . . . . . . . . . 122.1.3 Data evaluation . . . . . . . . . . . . . . . 152.1.4 Data analysis and interpretation . . . . . . . . . . 162.1.5 Public presentation . . . . . . . . . . . . . . 16

2.2 Effect of contextual factors in process harmonization . . . . . . 172.3 Process harmonization . . . . . . . . . . . . . . . 21

2.3.1 Concepts . . . . . . . . . . . . . . . . . 212.3.2 Operationalization . . . . . . . . . . . . . . 21

2.4 Effect of process harmonization on business performance . . . . 262.5 Conceptual model . . . . . . . . . . . . . . . . 282.6 Conclusions . . . . . . . . . . . . . . . . . . 30

3 Measures of process harmonization 333.1 Conceptual background . . . . . . . . . . . . . . . 343.2 Methodology . . . . . . . . . . . . . . . . . . 35

3.2.1 Conceptual definition . . . . . . . . . . . . . 363.2.2 Model development . . . . . . . . . . . . . . 373.2.3 Data collection . . . . . . . . . . . . . . . 383.2.4 Model assessment . . . . . . . . . . . . . . 423.2.5 Model validation . . . . . . . . . . . . . . 43

3.3 Results . . . . . . . . . . . . . . . . . . . . 433.3.1 Relevant aspects . . . . . . . . . . . . . . 44

viii Contents

3.3.2 Initial measurement model . . . . . . . . . . . 443.3.3 Empirical data . . . . . . . . . . . . . . . 463.3.4 Model dimensions . . . . . . . . . . . . . . 503.3.5 Final measurement model . . . . . . . . . . . . 53

3.4 Conclusions . . . . . . . . . . . . . . . . . . 57

4 Relationship between contextual factors and process harmonization 594.1 Conceptual background . . . . . . . . . . . . . . . 604.2 Methodology . . . . . . . . . . . . . . . . . . 62

4.2.1 Selection of cases . . . . . . . . . . . . . . 634.2.2 Definition of information sources . . . . . . . . . . 644.2.3 Preparation of cases . . . . . . . . . . . . . 674.2.4 Execution of interviews . . . . . . . . . . . . 694.2.5 Study of additional documentation . . . . . . . . . 694.2.6 Within-case analysis . . . . . . . . . . . . . 704.2.7 Cross-case analysis . . . . . . . . . . . . . . 71

4.3 Results . . . . . . . . . . . . . . . . . . . . 734.3.1 Effect of different locations . . . . . . . . . . . 734.3.2 Effect of IT governance centralization . . . . . . . . 744.3.3 Effect of product type . . . . . . . . . . . . . 754.3.4 Effect of maturity level . . . . . . . . . . . . 764.3.5 Effect of mergers and acquisitions . . . . . . . . . 774.3.6 Effect of organizational structure centralization . . . . . 774.3.7 Effect of organizational structure formalization . . . . . 794.3.8 Effect of level of process structuredness . . . . . . . 80

4.4 Validity and reliability . . . . . . . . . . . . . . . 824.5 Conclusions . . . . . . . . . . . . . . . . . . 84

5 Conclusions 875.1 Main conclusions . . . . . . . . . . . . . . . . . 875.2 Theoretical contributions . . . . . . . . . . . . . . 925.3 Practical implications . . . . . . . . . . . . . . . 935.4 Limitations . . . . . . . . . . . . . . . . . . 94

A Literature review selection strategy 95

B Code sheet 97

C Survey 103

D Questionnaire for conducting multiple-case studies 117

E Results of pattern matching in multiple-case studies 123E.0.1 Patterns in number of locations . . . . . . . . . . 125E.0.2 Patterns in IT governance centralization . . . . . . . 126E.0.3 Patterns in product type . . . . . . . . . . . . 127E.0.4 Patterns in maturity level . . . . . . . . . . . . 129E.0.5 Patterns in number of mergers and acquisitions . . . . . 130E.0.6 Patterns in organizational structure centralization . . . . 131

Contents ix

E.0.7 Patterns in organizational structure formalization . . . . 134E.0.8 Patterns in level of structuredness . . . . . . . . . 138

Bibliography 139

Summary 149

About the Author 153

Chapter 1

Introduction

“The scientist is not a person who gives the right answers,he’s one who asks the right questions”

Claude Levi-Strauss

Business process standardization aims to make similar business processes in anorganization uniform. Intuitively, the idea of business process standardization is toensure that if an organization performs the same activity in different places, it doesso in the same way [65]. There are many claimed benefits to business process stan-dardization. It has been claimed to lower the cost of executing processes and improvecollaboration, both between departments and between an organization and its busi-ness partners [19, 37, 64]. There also exists empirical eevidence that business processstandardization decreases the throughput time of a process, reduces cost and improvesquality and control [8, 74, 98].

While business process standardization has clear benefits, there may also be goodreasons to maintain some variability between business processes. Major reasons forhaving variations of a process include: requiring the ability to deal differently withdifferent types of customers and different cultures; and leaving different business unitswith reasonable autonomy to avoid micro-management [64, 88, 134]. Also, it has beenempirically shown that some level of variability cannot be avoided [51]. Exactly forthese reasons, business process modeling techniques have been developed that enablethe major variations of a business process to be mapped out, in addition to the standardflow of the business process [61, 80].

These observations lead to the conclusion that, when standardizing processes, abalance must be struck between the benefits of making those processes uniform on theone hand and the benefits of allowing variability on the other. Tregear [134] describesthis trade-off in detail. To stress this trade-off, the term ‘harmonization’ is also used,both by practitioners [47] and researchers [70]. The term is used to emphasize a strivefor balance between uniformity and variability, instead of a strive for uniformity, whichis associated with standardization. While the benefits of process harmonization arestudied in detail, little is known about the factors that drive the trade-off betweenuniformity and variability.

This introductory chapter presents the context and the topic of this thesis. Firstly,we set the scene for a thesis that examines contextual factors that determine the extentto which business processes can be harmonized and the type of influence exerted by

2 Chapter 1. Introduction

these factors. The scene is set by sketching the contingency theory guiding this research,and by defining the characteristics of contextual factors, process harmonization andbusiness performance. Next, we define the overall research goal, followed by specificresearch questions and their respective research approach. The chapter concludes withan outline of the structure of the reminder of this thesis.

1.1 Problem statement

In the attempt to enhance competitiveness, outsourcing, mergers and acquisitionshave emerged as very popular and widely adopted business strategies of this glob-alized era [20, 143]. This trend has been stimulated by changes in business envi-ronment drivers and the introduction of management concepts like business processre-engineering, benchmarking and alliance management [137].

As a result, challenges arise when organizations engage in joint activities, consid-ering that different organizations implement their business processes differently andinteract differently with the partners and supplier in their value chains. For instance,companies need to align their information systems and processes with customers anda wide range of providers including suppliers of materials and goods, service providersand partners to which activities have been outsourced. This situation becomes evenmore complex in cases of mergers and acquisitions where various processes and infor-mation systems exist, and interfaces become much more necessary to fully connect thecomplicated system landscape [117]. Significant challenges arise for the coordinationof activities and information exchange that need to take place within their value chain,considering the process variants existent among companies and business units [85].

To overcome these challenges, business process harmonization has become a neces-sity for global operating companies. Managers have been interested in the definition ofstandard processes, considering that it suggests operational economies of scale and thedevelopment of uniform best practices within a wide range of advantages. Over the lastyears business process harmonization enjoyed increasing attention from both practition-ers and researches. The advantages of process harmonization are well acknowledgedin the literature [8, 85, 146]. For instance, Munstermann et al. [99] summarized theadvantages of process harmonization in five categories: improved process performance,enhanced readiness, enhances ability to react to regulatory changes, enhance technicalinterchangeability and improve customer confidence.

However, several attempts of business processes harmonization have failed in prac-tice [18, 127], which have been attributable to factors such as their lack of flexibility toattend to the requirements of the different process variants. For instance, the world’sleading steel and mining company Arcelor Mittal, which operates in 60 countries andemploys about 245,000 people worldwide, gave up plans to standardize its businessprocesses due to the complexity of its worldwide operations. Instead, the companyimplement standardization in ten of its business units [25]. This situation shows thetrade-off faced by companies for the definition of a standard process. The trade-off con-sists of making a choice between the extent to which a process should be harmonizedand the necessary investments to achieve the expected harmonization level [134].

Despite indications of high levels of failure in harmonization initiatives and thehigh economic investments required to conduct these efforts, study of the factors thatinfluence this trade-off is still in its infancy. Knowledge of the influencing factorsthat determine the extent to which a process should be harmonized is a prerequisite

1.2. Theoretical foundation 3

to formulate strategies for ensuring the expected benefits of process harmonization.Without a complete understanding of the big picture, there are very significant risksthat the benefits of any harmonization initiative will not be exploited.

1.2 Theoretical foundation

This section reviews and outlines prior research that forms the theoretical foundationof this study. To work towards the goal of this research, we introduce a frameworkshown in Figure 1.1. In developing the framework, this study draws on research in thearea of strategy-environment fit, using contingency theory as a guiding theory.

Contextual factors

Process harmonization

Business performance

Figure 1.1: Theoretical framework based on contingency theory

Contingency theory suggests that successful organizations choose structures andprocess characteristics that fit to the degree of uncertainty in their environment [40, 92].It is a class of behavioral theory that emphasizes that there is no single best wayto manage an organization, and the optimal course of action is contingent upon theinternal and external situation. Originally proposed by Chandler [21], a contingencyapproach has been extended by many other researchers [50, 82, 116].

Using this theory, we propose that the extent to which business processes can beharmonized is dependent on contextual factors in the organization and its environment.This study fits with the assumptions of contingency theory stating that each unit ofanalysis is unique and has to be analyzed based on contextual factors [148]. For exam-ple, the extent to which order to pay processes can be harmonized in a multinationalcompany with thirty locations around the world is expected to be lower than the extentto which these processes can be harmonized in a small local business. We propose thatthe reason for those differences is systematic [34] due to differences with respect tofactors such as national culture, regulations, laws or even organizational structure.

Consequently, in addition to studying the concept of process harmonization itself,we study the contextual factors that determine the extent to which business processescan be harmonized. Also, in accordance with the goal of this study, we investigatethe effects that process harmonization has on business performance. This leads to theframework shown in Figure 1.1. In the remainder of this section, we discuss each ofthe framework’s elements in more detail.

1.2.1 Contextual factors

In contingency theory, contextual factors (also called contingency factors) are describedas environmental, organizational and individual characteristics of a firm’s external andinternal environment. Contingency theory proposes a fit between the organization andthose factors [21]. Specific contextual factors have been studied in wide variety ofsituations and topics.

In marketing literature, Homburg et al. [69] used differentiation and cost-leadershipstrategy, and distribution and customer base as internal contextual factors, and market


growth, market related uncertainty, and technological turbulence as external contextualfactors. In operations management, organizational size was used as an external contin-gency factor by [58] and [125]. In healthcare quality management, Wagner et al. [141]used contingency factors such as centralization of decision making, formalization ofregulations, and organizational size.

In organizational strategy, Ginsberg and Venkatraman [54] reviewed empirical stud-ies employing a contingency theory perspective. The contingency factors identified inthe studies reviewed, included: market structure [23], environmental uncertainty [68],product life cycle [131], environmental context [73], managerial characteristics [93], or-ganizational structure [57], managerial style [58] and internal technology [68]. In theirstudy, Ginsberg and Venkatraman [54] classified all the contingency factors identifiedin three broad groups called: environmental, performance and organizational factors.

Sila [125] evaluated the effect of contextual factors through the lens of two organiza-tional theories: institutional theory and contingency theory. The author evaluated theeffect of five contextual factors which included three institutional factors (conformanceto norms, standard accreditation and country of origin) and two contextual factors(company size and locations: local or multinational). The propositions derived relatedto institutional theory were called institutional factors and those related to contin-gency theory were called contingency factors. Damanpour [36] conducted a study inorganizational innovations, in which the impact of contextual factors and several typesof innovation were evaluated. The factors analyzed were called organizational factorsand included: functional differentiation, specialization, professionalism, administrativeintensity, organizational size and organizational slack. In this case we can see how thesame factors are just named differently by different authors based on the guiding theoryused to base their studies. Additional to the organizational and environmental factors,Rousseau and Fried [113] included worker-job factors such as prior work experienceand type of job (routine/ non-routine), and time.

Drawing on the studies outlined above, we can conclude that contextual factors arewell acknowledged in the literature and they have been studied in a variety of contextand research areas. A broad definition of contextual factors, in which they include allthe characteristics internal and external to the organization and its environment, areconsidered in this study.

1.2.2 Process harmonization

Process harmonization is the activity of aligning different variants of a family of pro-cesses, by capturing their commonality and variability in a consolidating and efficientmanner, to make them uniform or mutually compatible, in line with the businessstrategies and organizational goals. Based on this definition, process harmonization isconsider as a selective standardization, which accepts that different stakeholders in anorganization have different, possibly conflicting, requirements for a process, dependingon their context.

Different conceptualizations of process harmonization have been adopted in theliterature [47, 120, 146]. A precise specification of what we mean by process harmo-nization is dependent on the research stream that we decide to adopt. In this sectiontwo opposite research streams are discussed to derive a systematic and theoretical basisfor process harmonization.

The first stream consists of literature in which process harmonization is treated as

1.2. Theoretical foundation 5

similar to process standardization at a local level, across different locations, regions ororganizational units. For instance, Fernandez and Bhat [47] defined process harmoniza-tion as “the activity of designing and implementing business process standards acrossdifferent regions or units, so as to facilitate achieving the targeted business benefitsarising out of standardization, whilst ensuring a harmonious acceptance of the newprocesses by the different stakeholders”. Girod [55] also stated that harmonization isnot a question of force-fitting standards, but of “adopting a medium-term view whereinternal stakeholders from all around the world commonly decided which processeswould have to be local, regional, or global”. Even though Girod [55] highlighted thedistinction between harmonization and global standardization, he still treated harmo-nization as local standardization when stated that “the goal with process harmoniza-tion is not to impose one standard process, but to make selective decisions in whichmanagers can decide how much to standardize their processes” [55]. In this stream,standardization is defined in a broader sense in which local standards can also be theresult of standardization efforts.

In contrast, the second stream distinguishes differences in goals between harmoniza-tion and standardization. The goal of process standardization is to achieve uniformityof process activities across the value chain and across firm boundaries [146]. Empha-sizing this position, Schafermeyer et al. [120] defined process standardization as “theunification of business processes and the underlying actions within a company in orderto facilitate communication about how the business operates, to enable handoff acrossprocess boundaries in terms of information, and to improve collaboration and developcomparative measures of process performance”. Unlike process standardization, whichmainly pursues the unification of processes, the goal of harmonization is to align similarprocesses based on a single, focused business objective [70].

However, when analyzing the differences between harmonization and standardiza-tion as they are discussed in these two streams, we notice that differences only existwith respect to strict standardization. In this strict view, standardization leads to asingle unified process that does not allow variability [108]. However, in a more broadview, local variations on the standard process are also allowed. Therefore, we claim thatharmonization and standardization are similar concepts that differ only with respect totheir focus: standardization stresses the unification of processes, while harmonizationstresses a trade-off between global unification and local variation. In this paper focuseson that trade-off, therefore we prefer to use the term process harmonization.

1.2.3 Business performance

Business performance is an overall concept of organizational effectiveness, which in-cludes both indicators of operational performance (i.e. nonfinancial indicators) andindicators of financial performance [138]. The level of performance a business attainsis a function of the efciency and effectiveness of the actions it undertakes [101].

Business performance can be measured using different metrics and evaluated at dif-ferent levels in the organization, such as the process level and the overall organizationallevel. At a process level, the operational efficiency of specific business processes is eval-uated, using different measures, such as customer satisfaction and operational costs.Some examples of measures at the organizational overall firm level include productivityand level of responsiveness [91].


1.3 Research goal and research questions

Based upon the problem statement described in the previous section, the goal of thisthesis is to contribute to the development of theory regarding process harmonization byidentifying contextual factors that determine the level of harmonization, operationaliz-ing process harmonization, and specifying the type of relationship between the contextualfactors identified and the level of harmonization. The central research question is asfollows:

What contextual factors should be considered to decide on the right level ofharmonization of a business process in an organization?

This central research question is divided in four sub-questions organized aroundthree studies. Each study has a different research focus, objective, sub-questions andmethodology. The first study presents the contextual factors that determine the level ofprocess harmonization. The aim of this study is to identify the factors that determinethe extent to which business processes can be harmonized, relevant aspects for theoperationalization of process harmonization and the performance benefits that theorganization achieves through process harmonization. The sub-questions of study oneare:

RQ1. What contextual factors determine the level of harmonization of a businessprocess?

RQ2. Which aspects of a business process are relevant to measure the level of processharmonization?

The second study examines the operationalization of process harmonization. Theaim of this study is to develop and validate a set of operational measures to evaluatethe level of business process harmonization in an organization. As part of the con-ceptualization, the aspects identified in the first study are used as an input to providenew insights over the concept. Based on this consideration, two sub-questions areformulated for the second study:

RQ3. What measures can be used to quantify the level of harmonization of a businessprocess in an organization?

The third study examines how contextual factors determine the level of harmoniza-tion of a business process. The goal of this study is to identify the type of effect ofcontextual factors on different aspects of process harmonization. The sub-question ofstudy three is:

RQ4. How do contextual factors influence the level of harmonization of eachindividual aspect of process harmonization?

The contribution of these studies is cumulative. The first study contributes inte-grating the findings of previous research in a general conceptual model. This conceptualmodel includes the most relevant contextual factors which have been associated withprocess harmonization in previous research and the list of aspects that have been as-sessed through different measures used in the literature. The second study contributesto research by providing an entirely new measurement model to assess the level ofharmonization which distinguish different aspects of the process. This is the first mea-surement model providing a distinction between aspects in the literature. Finally, the

1.4. Research approach 7

third study contributes to theory in process harmonization by identifying the relation-ship between contextual factors and each aspect of the harmonization level. It providesnew insights about which aspects are related with each factor and the type of relation(positive or negative).

1.4 Research approach

The research methodology followed to answer the research questions stated in theprevious section, is defined around three studies.

In the first study we used a literature review to answer RQ1 and RQ2. The lit-erature review explored the contextual factors identified by different authors in theliterature and the type of associations defined between these factors, the operational-ization of process harmonization and business performance. The operationalizationof the concept in previous research was used to identify the most relevant aspects ofprocess harmonization. A conceptual model was developed at the end of this studyto integrate the contribution of different authors, and to summarize all the contextualfactors and aspects considered relevant in literature.

The second study answers RQ3 by the means of a combination of survey-basedresearch and interviews. The research approach consisted of two phases: an exploratoryphase and a confirmatory phase. In the exploratory phase, interviews were conducted toderive a the first set of measures and a measurement model for the operationalizationof process harmonization. For the confirmatory phase, a survey-based research wasused to empirically validate the measurement model defined in the exploratory phase.

With the third study, RQ4 is answered through case study research. Six compar-ative case studies are used to explore the relationship between the contextual factorsidentified in the first study and the level of harmonization measured using the modelderived in the second study. This evaluation is limited to large organizations, andparticularly explored for two procurement processes: supplier’s selection and ordering.The case study approach provides a deeper understanding of the relationships understudy, which makes it the most suitable approach to answer RQ4.

1.5 Outline

Figure 1.2 provides an overview of the structure of this thesis. The thesis is organizedas follows:

Chapter 1, the current chapter, firstly introduces the context of the thesis, describ-ing contingency theory, contextual factors, process harmonization and businessperformance. Additionally, we present the problem statement, research goal,research questions and the research approach followed to answer the researchquestions. Finally, we provide an outline of the general structure of the thesis.

Chapter 2 presents the first study of this thesis. Research questions 1 and 2 areaddressed in this chapter by presenting a conceptual model in which all the con-textual factors associated with the level of process harmonization are integrated,and their relationship and effect on business performance are also summarized.

Chapter 3 describes the second study. It shows the different steps followed to developa new measurement model for assessing the level of harmonization. It answers


research question 3.

Chapter 4 discussed the third study. It concludes with propositions about the rela-tionship between the contextual factors identified in chapter 2 and the level ofharmonization using the measures developed in chapter 3. Research question 4is addressed in this chapter.

Chapter 5 concludes this thesis by providing a final discussion as well as theoreticalcontributions, practical implications, limitations, and future work.

1.5. Outline 9

Research Outline

Chapter 1

Chapter 2

Chapter 4

Chapter 3

Chapter 5

Introduction

(Study #1)Conceptual Model

(Study #3)Relationship between factors and specific aspects of process harmonization

(Study #2)Measures of process harmonization

Conclusions

RQ1. What contextual factors determine the level of harmonization of a business process?

RQ2. Which aspects of a business process are relevant to measure the level of process harmonization?

RQ4. How do contextual factors influence the level of harmonization of each individual aspect of process harmonization?

RQ3. What measures can be used to quantify the level of harmonization of a business process in an organization?

Literature review

Case study

Survey and inteviews

Research questions

Research method

Contextual factors


Process harmonization

M‐1

M‐2

M‐3

M‐1

M‐2

M‐3

External Factors

Internal Factors

Immediate Factors

Strategic

Tactical

Operational

External Factors

Internal Factors

Immediate Factors

External Factors

Internal Factors

Immediate Factors

External Factors

Internal Factors

Immediate Factors

Figure 1.2: Overview of this thesis

Chapter 2

Conceptual model

“If I have seen further it is by standing on the shoulders of Giants”Isaac Newton

Positive effects of process harmonization are well acknowledged by practitioners andacademics. However, little is known about the factors that determine the success ofprocess harmonization initiatives. This chapter fills that gap, by presenting a concep-tual model that defines the operationalization of process harmonization, factors thatdrive the success of process harmonization, performance properties that are influencedby that success, and relations between these concepts.

Contextual factors are identified through a structured analysis of existing literature.The results draw from 35 relevant sources identified, are then used to answer RQ1.What contextual factors determine the appropriate level of harmonization of a businessprocess? and RQ2. Which aspects of a business process are relevant to measure thelevel of process harmonization? In doing so, we contribute to literature by showinga structured overview of the contextual factors, performance benefits and relationsand an operationalization of process harmonization. This overview can be used from ascientific perspective to derive and prove hypotheses about drivers and effects of processharmonization. It can be used from a practical perspective to get an overview of thefactors that must be taken into account when engaging in a process harmonizationproject and the effect that those factors have on the success of that project and thesubsequent potential to gain performance benefits.

The remainder of this chapter is structured as follows. The first section presents theresearch method, including a literature survey protocol. An analysis of the literaturefound is shown in the following sections, classifying it in three main groups: literaturethat identifies the effect of contextual factors in process harmonization (Section 2.2),literature that operationalizes the concept of process harmonization (Section 2.3), andliterature that describes the effect of process harmonization on business performance(Section 2.4). These findings are summarized in a conceptual model that is describedin Section 2.5. The final section concludes the chapter, by discussing the main findingsand limitations of our work.

12 Chapter 2. Conceptual model

2.1 Methodology

The research method selected to answer RQ2 and RQ3 is a literature review. A liter-ature review is the best approach to broaden the knowledge base in a research area,such as process harmonization, that is fragmented and for which the contribution ofdifferent authors has not been contextualized [132]. Cooper’s [30] Taxonomy of Liter-ature Reviews was used to classify our study according to six characteristics: focus,goal, perspective, coverage, organization and audience.

The focus of this review is on research outcomes and research methods. We areinterested in the identification of variables, measures and research methods used inprevious research on process harmonization. The goal is to critically analyze previousresearch, integrate different results and identify critical issues to propose further re-search opportunities in the field. The perspective from which this review is conductedis based on the predefined idea of what process harmonization is and the expected linkwith contextual factors and business performance. This idea is explained in the theo-retical foundations that are presented in Section 1.2. The coverage of our review is anexhaustive review with selective citations. All the articles identified that satisfied a setof criteria defined in Section 2.1.1 are included. The organization of this review is builtaround the concepts that are explained in the theoretical foundation and shown in Fig-ure 1.1. The audiences for this review are scholars within the field of business processmanagement and practitioners involved or interested in harmonization projects.

The research approach that we used, followed scientific guidelines for conductingintegrative research reviews defined by Cooper [29] and depicted in Figure 2.1. Theyinclude five stages: problem formulation, data collection, data evaluation, analysis andinterpretation, and public presentation. Each stage is described in detail in furthersubsections.

2.1.1 Problem formulation

A problem formulation includes the definition of the research questions that will guidethe literature review [30] and the set of inclusion criteria used for the selection process.

The research questions are based on the goal and focus of the review [107]. Threeresearch questions are formulated: From the previous literature, what is the effect ofcontextual factors on the level of harmonization of a business process? What is theeffect of the level of harmonization on business performance? What methods have beenused to investigate this effect?

A set of inclusion criteria was defined and summarized in Table 2.1. The set ofcriteria includes general, relevance and quality criteria. Articles selected should satisfyall the general and quality criteria, and at least one of the relevance criteria. Therelevance criteria include all the possible topics of interest for our collection. Therefore,an article that covers at least one of these topics should be selected.

2.1.2 Data collection

The data collection step includes a definition of the search strategy and the selectionof relevant articles that are included in the review. A search was conducted, whichconsists of a pre-search, a systematic search and a cross-reference search.

A pre-search was conducted using Google Scholar as a search engine, to identifyhow extensive the literature on this topic is and to determine the keywords to be used

2.1. Methodology 13

DescriptionStage of Research

• Define research questions

• Inclusion/exclusion criteria for selection

• Pre‐search: Google scholar

• Systematic search: ABI‐INFORMS, EMERALD, SPRINGER

• Cross‐reference search: backward tracing technique

• Extract outcomes and methods

• Qualitative synthesis: free‐coding, descriptive and analytical themes

• Present three topics: (1) contextual factors, (2) process harmonization: concepts and measures; and (3) business performance

Problem Formulation

Step 1

Data Collection

Step 2

Data evaluation

Step 3

Data analysis ‐interpretation

Step 4

Public presentation

Step 5

Figure 2.1: Five stage approach for conducting integrative research reviews proposed by Cooper [29]

Table 2.1: Outline of the inclusion criteria used for the selection process.

General criteria

1. The report is written in English2. Data from one study did not overlap data from another study

Relevance criteria

1. The article provides a definition of harmonization in the context of businessprocesses.2. The article provides a description of a harmonization process, includingdifferent stepsthat provide enough information to derive a definition of the concept.3. The study describes a methodology to conduct harmonization projects.4. The study presents an empirical application of process harmonization.5. The study describes a relationship (direct effect, moderator, and mediator)between contextual factors and process harmonization.6. The study describes a relationship (direct effect, moderator, and mediator)between process harmonization and business performance.

Quality criteria

1. The article is a published journal or conference paper.2. The method used to derive the results is explained in the article


for the search. This search engine was chosen considering that it has a broad scope andis therefore suitable to efficiently find a large collection of papers and keywords thatare related to harmonization. Using the keyword ‘harmonization’ only a limited num-ber of articles are found, and very few specifically focused on process harmonization.The keyword ‘process standardization’ was also used for the pre-search considering theoverlap with this concept as explained in detail in Section 1.2.2. When adding thiskeyword a high number of papers are retrieved. A high occurrence of keywords identi-fied during the pre-search was interpreted as an indicator for keyword relevance. Thekeywords selected are shown in Table 2.2.

After conducting the pre-search step, a systematic search step was conducted, us-ing the keywords that were identified in the pre-search step and three search engines:ABI/INFORMS, EMERALD and SPRINGER. These search engines were used, be-cause they cover most of the high-quality journals and conferences in the area of inter-est. A pre-selection of the articles was performed looking for keywords in the title orabstracts of the papers retrieved.

Table 2.2: List of keywords used for the systematic search.

Keywords

“harmonization of business processes”“process harmonization”“business processes harmonization”“process harmonization” AND “business processes”“harmonization of processes”“harmonize” AND “business processes”“harmonization” AND “business process”“standardization of business processes”“process standardization”“business processes standardization”“process standardization” AND “business processes”“standard process” AND “business processes”“standardize” AND “business processes”“process variants”“franchising” AND “standardization”(mergers or acquisitions) AND “standardization”

The cross-reference search is performed using a backward tracing technique, aftercompleting the first search cycle and selecting a set of articles using the selection criteriaspecified in Section 2.1.1. The cross-reference search provides an effective method thatcan lead to locate a significant number of the articles that comprise an exhaustivereview [107]. The backward technique is promising in this case, considering that itallows us to find articles that are relevant to our study, but that cannot be found withthe systematic search because of differences in the terminology used.

At the end of the data collection process, a selection of the articles that are rele-vant for this review is conducted. The selection process was conducted by inspectingthe title and abstract to separate the potentially relevant articles from the obviouslyirrelevant studies, using the relevance and general criteria described in Table 2.1. Af-ter the irrelevant articles were discarded, the articles were evaluated using the qualitycriteria also described in Table 2.1. This evaluation was conducted by reading themethod, analysis and conclusion sections. This literature review was conducted by one

2.1. Methodology 15

researcher. The list of articles selected and the results of the selection strategy areprovided in Appendix A.

2.1.3 Data evaluation

In this step, we describe the type of data that is extracted from each article selectedduring the data collection. Because the focus of this literature review is on outcome andmethods, we extracted information about contextual factors, process harmonizationand business performance, and the methods used by authors to derive their results.

All selected articles were classified using a theoretical model that was derived by ex-tending the framework of Figure 1.1 with the framework described by Kumar [77]. Ourframework presents the different relations that we investigate, while Kumar’s presentsthe different types of relations that may exist among variables. This results in thetheoretical model depicted in Figure 2.2, which shows a chain of causes and effects,where in the first segment contextual factors are the cause that determine the extentto which the effect of processes harmonization can be achieved, while in the secondsegment process harmonization is the cause that determines to which extent the ef-fect of business performance benefits can be achieved. Consequently, the theoreticalmodel contains three concepts that are interrelated: (A) process harmonization, (B)contextual factors, and (C) business performance.

Contextual factors

Process harmonization:

- concept- measures


Mediators Mediators

Moderators Moderators

Figure 2.2: A framework for classifying the literature in process harmonization

For the first concept, contextual factors, we extracted the different factors, the typeof relationship exerted on process harmonization (direct effect, moderator or mediator)and the methodology used by authors to derive these relations such as case studies orliterature review. A moderator is a factor that increase or decrease the type of influenceexerted (positive or negative) by contextual factors, while a mediator explains how orwhy these relations exists. For instance, a legal requirement that specifies rules forpurchase orders above a certain amount of money, can add specific activities to thesupplier selection process only for this specific type of orders and not for orders bellowthe specified amount. If we want to harmonize the supplier selection process for allthe orders, this legal requirement imposes a restriction in terms of how far can it beharmonized. This restriction is not necessarily present in all companies around theworld producing different product types. Therefore, the contextual factor called legalrequirement negatively influences the level of harmonization of the supplier selectionprocess in this organization. The type of order is a mediator, because if all ordershandled are above a certain amount of money, there is no need to adapt the process fordifferent types of orders and the level of harmonization increases. However, if there aredifferent types of orders that must be handled, the level of harmonization decreases.


For the second concept, process harmonization, we identify how different authorsdescribe process harmonization as well as its operationalization. The operationalizationincludes the identification of variables or indicators that define the level of harmoniza-tion and the identification of the methods used to evaluate the validity and reliabilityof the indicators derived. For instance, Munstermann et al. [99] defines process harmo-nization as the process of aligning business process variants to an archetype process.Based on this definition, Munstermann [99] presents a measure to quantify the extentto which processes are harmonized, on a scale from one to seven. In order to measurethe extent to which processes are harmonized, Munstermann et al. [98] uses severalindicators, including for example we have documented our actions to a great extentthat organizations can measure, using a Likert scale, based on how much they agree(7) or disagree (1) with the indicator statement.

The third concept, business performance, includes different performance indicatorsthat are used to evaluate the effect of process harmonization on business performance.

The model depicted in Figure 2.2 provides an appropriate theoretical frameworkfor this study because it emphasizes the interactions between process harmonization,contextual factors and business performance, aligned with the guiding theory usedto conduct this study (See Figure 1.1) as suggested by Webster and Watson [142].A coding sheet is used to evaluate the information in the articles in a systematicmanner [128]. The coding sheet is included in Appendix B.

2.1.4 Data analysis and interpretation

A qualitative synthesis of the data extracted is performed. This type of analysis isappropriate in our case, because we have to combine a mixture of quantitative andqualitative results [107]. The method used is a thematic synthesis proposed in [130].This synthesis has three stages which overlap to some degree: the free line-by-linecoding of the findings of primary studies; the organization of these free codes intorelated areas to construct descriptive themes; and the development of analytical themes.

In particular, one of the authors went through the papers that were identified inthe data collection step, to identify sentences that relate to elements of the frameworkshown in Figure 2.2. In particular, sentences were identified that contain statementsabout: factors that influence the extent to which processes can be standardized, meansto measure the extent of to which processes are standardized, or effects of process stan-dardization on organizational performance. Subsequently, the author unified factors,measures or effects that were similar in nature according to the author. The results ofthat effort are the results that are presented. In order to identify the different aspectsof business process standardization operationalization, we studied the operationaliza-tions that were found in the literature to determine what was measured. For example,Munstermann et al. [99] measure the extent to which processes, and actions are stan-dardized and Schafermeyer et al. [119] also measure the extent to which documents arestandardized.

2.1.5 Public presentation

The information that is relevant is organized to be discussed in the next section. Itis organized conceptually, using the structure proposed in Figure 2.2. The three mainconcepts are: (1) contextual factors and their effect on process harmonization, (2)

2.2. Effect of contextual factors in process harmonization 17

concepts and measures of process harmonization, and (3) the effect of process harmo-nization on business performance. All data extracted in subsection 2.1.3 are analyzedand presented in the next section.

2.2 Effect of contextual factors in processharmonization

From the set of 35 articles selected, 10 presented only contextual factors that impactprocess harmonization, 15 were focused on the effect of process harmonization on busi-ness performance, 6 include both sides (contextual factors and effect), and the last 4described only the concept of process harmonization at a theoretical level.

Table 2.3 shows the effect of contextual factors on the level of harmonization ofbusiness processes in an organization. The contextual factors are depicted in the firstcolumn, whereas the second column presents a different possible settings of each factorwhich is related to an expected ‘low’ or ‘high’ level of harmonization. For instance,one factor is the level of process structuredness. If the process is nonroutine, such as asupplier’s selection process, it is expected that when trying to harmonize this process alow level of harmonization can be achieved. While for routine or standard processes theexpected level of harmonization is high. The next column shows the validation methodused by authors to support their statements about the potential of each contextualfactor to explain the level of harmonization of a set of business processes. Finally,references are included in the last column.

The contextual factors identified in the literature include: cultural differences, dif-ferent regulations, power distance, number of different locations, IT governance cen-tralization, product type, maturity level, organizational structure, number of mergersand acquisitions, process type, level of process structuredness, and personal differences(level of experience and knowledge). For these contextual factors, a number of relationsto process standardization were identified and proven in the literature collected.

Ang and Massingham [2] analyzes how the decision of harmonization versus adapta-tion should be made taken into consideration the level of differences in national culture.Standardization refers to a common approach to business throughout the world, whileadaptation requires a different approach in each market. It suggests that the greaterthe cultural differences, the greater the difficulty in knowledge transfer across cultures.The cultural difference is a criteria to base the decisions of harmonization versus adap-tation.

There are mandatory and unavoidable variations that come from differences in reg-ulations such as financial regulations, taxation regimes and employment practices [134].In order to comply with these legal requirements companies need to adapt their pro-cesses and end up with multiple variations of the same process. Therefore, this situationdefinitely influences the level of harmonization that can be achieved for global com-panies and must be considered when strategic decisions are made such as acquiringsubsidiaries in other country or region with different legislations.

18C

hap

ter2.

Con

ceptu

alm

od

elTable 2.3: Relationship between contextual factors and the level of harmonization.

FactorsHarmonization level

Validation ReferencesLow High

Differences in national andorganizational culture

high low Literature review [2, 10, 95]

Different regulationsmany differentregulations

a few or nonedifferent regulations

Case studies (1project at nationallevel; 1 company; 3 companies: 1company in 30 countries)

[67, 110, 111,134]

Power distance high low Case study (3 companies)[95]

Number of locations several fewCase study (1 company in 30countries)

[134]

IT governance centralizationdecentralized localdecisions

centralized globaldecisions

Literature review, case studies (3companies; 8 multinationalcompanies) and surveys

[10, 55, 112]

Differences in product type(product innovation, volume ortechnological uncertainty)

high low Literature review

[106]

Maturity level low moderate/high Case study (3 companies)[111]

Organizational structure — hierarchical networkCase study (3 companies), data setanalysis (152 companies)

[55, 126]

2.2.

Eff

ectof

contex

tual

factorsin

pro

cessh

armon

ization19

Table 2.3 – Continued from previous page

FactorsHarmonization level

Validation ReferencesLow High

Number of mergers andacquisitions

applying practicesseveral commonproc. with differentversions of eachprocess

without practicesfew common proc.with few versions ofeach process

Case study (1 company in 30countries)

[134]

Process type not transactional highly transactionalCase study (3 companies; 1company in 30 countries)

[35, 55, 134]

Level of process structuredness nonroutine processesroutine and standardprocesses

Literature review/ Case studies (3companies; 2 companies)

[84, 111, 120]

Personal preferences based onauthority level

several differences few differencesCase study (1 company in 30countries)

[134]

Work experience combined withskills and knowledge

medium/highexpertise high skilledwho needs lowproceduralknowledge moretacit knowledge

low experience withexplicit and highproceduralknowledge

Case study (2 companies)

[120]


In inter-firm collaborations, the characteristic of the relationship among firms is anaspect of organizational differentiation that has been widely recognized for affectingthe scope, structure and performance these collaborations. It includes various factorssuch as power distance, partners’ financial and legal independence, and operationaland cultural diversity. Table 2.3 only shows a relationship between power distance andprocess harmonization. It was observed that organizations with low power distancehad a higher level of integration of their business practices, while those with mediumand high power distance had a low level of integration [95].

For globally operating companies, having different locations do not only affect thelevel of harmonization because of differences in their legal requirements, but due to thefrequency of interaction between individuals performing different tasks. Each individ-ual has their own way of working (personal differences) and it is shaped by its culturalbackground. What works in one location may not be feasible in another location con-sidering that the necessary resources are not available or affordable. Also differences inlocal market imperatives increase process variations. It is difficult to isolate the effectof location because the effect is mixed with differences in national or regional culture,customer expectations, market maturity and local market conditions [134].

The centralization of IT governance may lead to a higher level of harmonization. ITplays a significant role in reaching business objectives. In heterogeneous IT landscapescharacterized by different systems in similar functional areas, decentralized IT depart-ments, or insufficient IT service levels, the decision to standardize definitely providessignificant improvements [17]. However, in some cases the initial investment neededto centralize this infrastructure is too high, and savings that can be achieved throughprocess harmonization do not balance this investment.

Mergers between companies with a different range of products or services demandalso adaptations of their supporting activities such as purchasing and marketing. Dif-ferences in products and services may require variation in the processes that create,deliver and maintain them [134].

Organizational structure was also identified as a contextual factor that exerts aninfluence in the level of harmonization. Girod and Bellin [55] describes how a hier-archical network based on both vertical and horizontal relationships, facilitates thecentralization of decisions by headquarters. Harmonization leverages the principle ofdistributed leadership contained in the hierarchical network to allow top, functional,regional, business and geographic unit managers to: (1) share responsibility of decid-ing which managerial processes should be standardized or remain local, and (2) assesswhere the best standards should come from in the organization.

The managerial practices identified in our review are mergers and acquisitions. Theimplementation of one or several managerial practices, such as mergers and acquisitions,definitely influences the level of harmonization of business processes. This increasesthe number of process variants that coexists. These variants might differ with respectto their performance such as efficiency, quality and cycle time. The harmonization ofthese variants consolidates processing volumes and allows the organization to exploiteconomies of scale. If the firm merges several variants of the same process, it canidentify the variant showing the highest performance and apply it as the new processstandard [8].

The level of process structuredness has an impact on the harmonization potential,and therefore on the success of a process harmonization initiative [84, 120]. “Nonrou-tine processes are less applicable to standardization than routine processes; it became

2.3. Process harmonization 21

apparent that even those processes that are nonroutine, or even creative, may comprisesections that may in fact become subject to process standardization” [111]. The rea-son underlying is that different parts of a process need to be open for creative decisionmaking, while others have to meet legal requirements of different countries. There arealso unstructured, unmeasured and unrepeatable processes that can lead to a low levelof harmonization. For instance, “Knowledge work is often said to be impossible todocument and model as a process” [134].

Authors observed that organizations which performed better in their harmonizationinitiatives, have at least a moderate level of process maturity. They conclude thatmaturity level has a positive correlation with harmonization potential [111].

The potential of a process to be successfully standardized also depends on personaldifferences such as level of experience and knowledge of the people involved in theprocess. Table 2.3 shows that for processes demanding employees with medium orhigh work experience or tacit knowledge to be performed, have less potential to besuccessfully standardized. The lack of interpretative assessment via employees duringa process suggests that standardizing this process is possible and leads to a successfulstandardization process [120]. Additionally, people are used to work in a specific wayand may not be willing to change, while some others see the sharing of knowledgeabout how a process works as a loss of control and power [134]. If these differencesare significant between people with the same role, they increase business processesvariation.

2.3 Process harmonization

In this section two aspects of process harmonization are analyzed: (1) the conceptof process harmonization, and (2) the operationalization used to measure the level ofharmonization of a set of business processes in an organization.

2.3.1 Concepts

The construct of process harmonization has been defined in multiple ways by differentauthors in the literature, by relating it to other concepts as shown in Table 2.4. Therelated concepts are: standardization, adaptation, integration, alignment, collabora-tion, centralization, modularization, coordination, homogenization and customization.Standardization was the concept most frequently related with process harmonizationin the literature in 29 out of 35 documents selected, followed by integration in only 10appearances. This result shows that the majority of the articles define harmonizationas a variation of standardization, as we discussed in Section 1.2.2.

2.3.2 Operationalization

Table 2.5 shows how different authors operationalize the construct of harmonization.It shows which types of variables are used to measure the level of harmonization, onwhich scale the level of harmonization is expressed, and how measures are evaluated.The literature found can be classified in three groups based on their type of opera-tionalization of process harmonization: (1) no operationalization provided, (2) usingattributes, and (3) using indicators with a numerical scale.


Table 2.4: Concepts describing process harmonization in the literature.

Article Sta

nd

ard

iza

tio

n

Ad

ap

tati

on

Inte

gra

tio

n

Ali

gnm

ent

Co

lla

bora

tio

n

Cen

tra

liza

tio

n

Mod

ula

riza

tio

n

Coo

rdin

ati

on

Ho

mog

eniz

ati

on

Cu

sto

miz

ati

on

1. Ang and Massingham [2] X X2. Beimborn et al. [8] X3. Boersma and Kingma [10] X X4. Buchta et al. [17] X X X5. Dai et al. [35] X6. Girod and Bellin [55] X X X7. Helfert [67] X8. Hufgard and Gerhardt [70] X9. Kobayashi et al. [76] X X10. Kumar and Harm [78] X X11. Lillrank [84] X12. Manrod and Vitasek [88] X13. Martins and da Silva [89] X14. McLaren et al. [90] X X X15. Moffat and Archer [95] X X X16. Mortensen and Lemoine [96] X X17. Munstermann et al. [97] X X18. Munstermann et al. [98] X19. Munstermann et al. [99] X X20. Norta and Eshuis [102] X X21. Perego and Salgaro [104] X22. Quintens et al. [106] X X X23. Ricken and Steinhorst [108] X X24. Rosemann et al. [110] X25. Rosenkranz et al. [111] X26. Ross et al. [112] X X X27. Schafermeyer et al. [120] X28. Schafermeyer et al. [119] X29. Shang and Seddon [122] X X30. Sorenson and Sorensen [126] X X31. Tregear [134] X X32. van Leijen [135] X X X X33. van Liere et al. [136] X X34. Wullenweber et al. [146] X35. Zhao [149] X

Total 29 1 10 3 2 3 2 6 3 4


Table 2.5: Classification of the literature based on the operationalization of process harmonization.

Type Scale Evaluation Amount References

None — — 22 ∗

Attributeslow/high None 1 [112]current/target/global None 1 [134]

Indicators 1-7 scale

Content validity,

5 [8, 97, 99, 119, 146]convergent validity,discriminant validity,reliability

∗ All the references used in Table 2.4 except for those classified in attributes and indicators

The majority of the articles selected (22) belongs to the first type of articles, whichonly described the concept at a theoretical level without providing any way to measurethe level of harmonization. For instance, Lillrank [84] discussed the inner workings ofroutines, defined as attributes of processes. It suggests that a process can be standard,routine or nonroutine, depending on how it is structured in relation to its environmentand resources. This classification was defined by the author as a conceptual constructand differences between each process type were discussed in detail through the paper.

The second type of articles defines attributes, which are characteristics of an orga-nization that has achieved harmonization to a certain extent. For instance, Ross etal. [112] present a classification of harmonization that consists of two levels: ‘low’ and‘high’. One of the attributes of the low level is that IT application decisions are madein business units. An organization that has this attribute can therefore be consideredto have a low level of harmonization and the aspect evaluated with this attribute is In-formation Technology (IT). In addition to the two levels presented by Ross et al. [112],Tregear [134] presents three levels. These levels describe a trajectory for reengineer-ing processes towards a global standard. Tregear [134] states that local variation mayneed to be accepted as necessary, therefore the harmonization effort may end at the‘current’ or ‘target’ level. Tregear [134] neither presents a detailed distinction betweenthe three different levels of standardization, nor presents specific criteria to indicatewhen a shift from one level to another occurs. In these studies there is no commonal-ity among the scales defined, however they evaluate a wider spectrum of aspects thanthe articles using indicators, including: activities, control-flow, IT, resources, data andmanagement.

The third type of articles uses a numerical scale to assess the level of harmonization.Articles in this category defined process harmonization as a construct that was mea-sured using reflective indicators [8, 97, 98, 119, 146]. Reflective indicators are measuresthat change with changes in the construct. For instance, Schafermeyer et al. [119] usedthe construct business process standardization and one of the three indicators used tooperationalize the construct was we have documented all actions of the business processto a great extent. This indicator is reflective because it is expected that if the extentof process standardization is high, the extent of documentation of all the actions of aprocess should also be high. But the fact that the extent of documentation is high donot necessarily implies that the extent of process standardization is also high.

Five of the articles reviewed defined indicators for process harmonization and all ofthem used multiple reflective indicators. The list of indicators is depicted in Table 2.6.It can be observed that the level of harmonization of a process is associated with


the level of harmonization of their: data content, activities, control-flow, informationtechnology, resources and management. The number and type of indicators vary perstudy. However, each study uses multiple indicators, to take into account severalaspects of the concept of harmonization. There is little commonality among authorswith respect to the indicators used, except for the fact that they all use a 7-point Likertscale. Each author defines his own set of indicators.

Table 2.6 also shows how the measures were assessed to test the quality of theirindicators. These studies describe the methods followed to analyze the validity andreliability of the indicators, to justify that they actually reflect the intended meaningof the construct. To evaluate the quality of the reflective indicators they evaluate:content validity, reliability and construct validity, which is composed of convergentand discriminant validity. The content validity shows the degree in which a constructreflects the expected meaning in its measurement. Authors evaluated the contentvalidity consulting several experts in the process under study and taking the indicatorsfrom preceding research.

Another evaluation of the quality of the indicators used in these studies is thereliability. Reliability defines the linkage between an indicator and its construct. Thet-test to measure the significance of each indicator was conducted using bootstrappingwith a large sample size (greater than 200 samples). Finally, authors also measure theconstruct validity of their indicators. Construct validity consists of convergent anddiscriminant validity. Convergent validity shows the internal consistency of indicatorswhen multiple measures are used to represent one construct, which is always the casein the studies described in Table 2.6. While discriminant validity indicates a goodfit between the measurement item and the latent variable. They evaluate it usingthe square root of the AVE which should always be greater than the inter-correlationbetween latent variables. These methods do not apply to measures derived usingattributes, as indicated in Table 2.6.

Based on the evaluation of the different indicators shown in Table 2.6, we canconclude that the set of six indicators derived by Munstermann et al. [97] best reflectsthe concept of process harmonization. First, it has a high reliability, especially forindicators ten to fifteen. Second, the convergent and discriminant validity tested usingthe square root of AVE is high, showing a good internal consistency of the indicatorsto reflect the construct. This result could be expected, considering that the sets ofindicators evaluated by the other authors consist of three items instead of six. However,even with more indicators in this construct the authors covered less aspects (activitiesand control-flow) than the ones covered with three indicators, as shown in Table 2.6.

To conclude, the level of harmonization has been operationalized in three ways: nooperationalization provided, using attributes or using an indicators with a numericalscale. The majority of the articles do not provide an operationalization of the concept.The articles that use attributes have different scales which are not comparable, butthey covered a wider variety of aspects than the ones using indicators. In the last typeof studies, all the indicators used to define the level of harmonization were reflectiveand use a 7 point Likert scale.


Table 2.6: Measures of process harmonization and their evaluation.

Type Measure AspectsValidation method

Indicator Convergent Discrim.reliability validity validity

Att

ribute

s

1. Business units have similar oroverlapping operations.

Activities

2. Data is standardized acrossbusiness units.

Data

3. IT applications decisions aremade in business units.

InformationTechnology None None

4. There are globally integratedbusiness processes often withsupport of enterprise systems.

Information [112] [112]Technology

5. Centralized management.Resources

6.There is a centralized controlover business process design.

Resources-Management

Att

ribute

s

7. Process performance is reportedglobally.

Data-Management

8. Global quality assurance onprocess work.

Management None None[134] [134]

9. Apply functional/ process/business units matrices.

Resources

Indic

ato

rs

10.The process cycle tracking iswell-regulated.

control- 0.9891flow

11.Process is efficient andorganized with transparent andcomprehensible activities.

activities 0.9799control-flow

12. The procedures are highlystandardized in the process

activities 0.9801

13. Processes and activities aredocumented to a great extent.

activities 0.9685control-flow

14. There is a fixed procedure forthe collaboration betweendepartments.

control- 0.9893

Compositerelia-bility:0.9839,AVE:0.9112 [97]

AVESquareRoot:0.9546 [97]

flow

15. The department alreadyworked with process standards inthe process under study.

not 0.8070specified

Indic

ato

rs

16. The process runs throughmandatory process steps.

activities 0.814Compositerelia-bility:0.857,AVE:0.669

AVESquareRoot:0.818 [8,99,119,146]

flow17. There are mandatoryspecifications for each step of theprocess.

data 0.894

18. The process is highlystandardized.

activities, 0.737control-flow


2.4 Effect of process harmonization on businessperformance

The effect of process harmonization or standardization is well acknowledged in the lit-erature. Process harmonization is recognized in literature as a driver of performanceimprovements in terms of cost, time, efficiency, effectiveness, quality, and responsive-ness. Table 2.7 summarizes the contributions found in the literature that indicate atleast one effect of process harmonization on business performance and the researchmethods used for the validation of the relations identified. These effects are measuredusing performance indicators such as efficiency, effectiveness and growth rate as shownin Table 2.7.

The effect of harmonization initiatives on financial performance is the one thatattracted the most attention, both at firm [76, 95, 96] and at operational level [104,106, 146]. This is followed by the effect on operational performance, including time,quality, efficiency and effectiveness [98, 149]. Finally, a few studies explore the impactof process harmonization on strategic performance, such as growth rate or the successof a managerial strategy.

Table 2.7: The effect of process harmonization on business performance.

Performancemeasure

Mediator Moderator Effect Validation method

Globalintegration

— — IncreaseCase studies (3companies) [55]

Businessvalue

Marketvolatility

— IncreaseAnalyze secondary data(98 samples from 59firms) [35, 106]

Outsourcingsuccess

Relationaland con-tractualgovernance

— IncreaseSurvey (335 samples)[146]

Profit growth — — IncreaseCase study (3companies), literaturereview [95, 106]

Risk growthrate

— — DecreaseCase studies (2companies) [120]

Costs

IT man-agement,relationaland con-tractualgovernance

— Decrease

Literature reviews (2), 2case studies (1company), case study (3companies; 7companies), survey (191samples), case study (1company) and survey(156 samples), casestudy (5 cases); casestudy (1 company in 30countries), 2 surveys(335 and 51samples) [17, 78, 88, 90,95–99, 104, 106, 134,146, 149]

2.4. Effect of process harmonization on business performance 27


Performancemeasure

Mediator Moderator Effect Validation method

Costs of ITsystemdevelopmentand maint.

— — DecreaseCase study [76]

Service levels — — IncreaseLiterature review, casestudy (1 company in 30countries) [90, 134]

Efficiency

Processcontrol,ITmanagement

ITintensity

Increase

Survey (65 samples),literature review, Casestudies (3,1,3 and 7companies), case study(5 cases), case study (1company in 30countries), case study (1company), survey (51samples) [8, 17, 55, 78,96, 104, 134, 136, 149]

EffectivenessITmanagement

— IncreaseLiterature review, casestudy (1 company in 30countries) [17, 134]

Quality

ITintensity,relationaland con-tractualgovernance

— Increase

Survey (65 samples),Case study (1company), survey (191samples), case study (1company) and survey(156 samples), casestudy (5 cases), casestudy (1 company),Survey (335samples) [8, 78, 97–99, 104, 106, 136, 146]

Responsiveness — — IncreaseCase studies (3companies), literaturereview [55, 106]

The effect of process harmonization on business performance is not always direct.For instance, Wullenweber et al. [146] describe the case in which the effect is medi-ated by relational governance (specifically communication, coordination and consensus)and contractual governance. They argue that: “Using process standards allows for abetter understanding about how the business operates and can be improved. Thisfacilitates communication and coordination between exchange partners and allows re-aligning disparate goals and actions to solve day-to-day problems. These findings showthat process standardization increases the effectiveness of relational governance” [146].This applies in a context with a high business and technological uncertainty wherecontractual provisions can hardly be designed. The second indirect effect is mediatedby contractual governance. Process harmonization provides transparency with better


documentation of processes. That leads to a higher measurability of process outputand control throughout the process. Under this situation, more specific and completecontracts can be designed and negotiated between parties involved, positively affectingthe outsourcing success.

Another case in which the effect of process harmonization is not direct, was de-scribed by Beimborn et al. [8]. In this paper the effect of process standardization onperformance is mediated by process control and moderated by IT intensity. The me-diation effect through process control is significant only on efficiency, while the effectof IT intensity is direct and also moderates the effect of process standardization.

Another interesting finding was presented by Dai et al. [35]. In this paper processharmonization is presented as the mediator in the effect of market volatility and busi-ness performance. When market volatility rises, firm performance increases also as aresult of primary business process standardization. In this case the type of businessprocess (primary or secondary) serves as a moderator. “Standardizing primary businessprocesses exerts a direct impact on how the firm delivers value activities. Standardizingsupport business processes will have a less direct impact, and should result in leveragefor tactical improvements, not strategic ones” [35].

Finally, we also observed that must of the relations between process harmoniza-tion and business performances were validated using empirical research, including casestudies and surveys. The majority of these studies were exploratory and provided de-scriptive interpretations of their results without using quantitative methods to guidethem.

Summarizing, process harmonization has an effect at different levels of businessperformance, including: strategic, tactical and operational performance. This effecthas been evaluated in many contexts, using performance indicators such as efficiency,costs and effectiveness. The effect can be direct or indirect through mediators such asthe level of process control, or moderators such as process type (primary or secondary)and IT management and IT intensity.

2.5 Conceptual model

This section presents a conceptual model that integrates the literature in process har-monization, based on the literature survey described in the previous sections. Thiscan be used as a meta-model for classifying the exiting literature and to examine thedifferent factors that influence the extent to which a process can be harmonized andthe subsequent effect on business performance.

The model consists of three parts as shown in Figure 2.3. In the first part we clas-sify the contextual factors identified in Table 2.3 in three categories (external, internaland immediate), depending on the level of the organization that they characterize.The second part indicates different aspects of process harmonization that can be af-fected by contextual factors in an organization. The third part concerns the type ofbusiness performance effect that can be expected with improvements in the level ofharmonization of an organization. The arrows linking the three parts articulate theinterdependencies among them (i.e. between a contextual factor and level of harmo-nization, or level of harmonization and business performance) and the signs indicatethe type of relationship (positive, i.e. an increase of the source implies an increase inthe target; or negative, i.e. an increase in the source implies a decrease in the target).

2.5. Conceptual model 29

+

Strategic• Global integration

• Business value

• Outsourcing success

• Profit growth

•Risk growth rate

Operational• Efficiency

• Effectiveness

• Quality

• Responsiveness

• Time

+

+

+

‐

+/‐

+/‐

+

+

+


Tactical• Costs

• Service level+

Extent of standardization

Data

Resources

‐

‐

Activities

Information technology

Management

Different regulations

Product type

External

Internal

IT Governance centralization

Maturity level

Cultural differences

Power distance

Number of different locations

Number of mergers and acquisitions

Organizational structure

Immediate Level of structuredness

Contextual factors

‐

‐

‐

+

+/‐

+

+

‐

+

Personal differences+

Control‐flow

Figure 2.3: A conceptual model explaining the determinants and effect of process harmonization

The first part of the model distinguishes three different levels in the organizationalcontext. The immediate level includes factors that are internal and directly relatedto the process under study, such as the level of structuredness of a process and per-sonal differences in terms of knowledge and experience. The internal level considersfactors that are part of the internal environment of an organization, i.e. organizationalstructure and number of different locations. The external level considers factors thatcharacterize the business network and the macro-economic context in which the orga-nization operates and that are beyond the control of an individual organization, suchas legal requirements for specific industries. The eleven contextual factors identifiedare divided in two immediate factors, six internal and three external, as depicted inFigure 2.3.

The second part of the conceptual model shows six aspects that can be differentiatedwhen evaluating the level of harmonization of business processes. These aspects werederived from the set of indicators described in the literature to measure the level ofharmonization, as shown in Table 2.6. They include: data, activities, control-flow,resources, information technology and management. The conceptual model does notdefine how the contextual factors affect each aspect separately, because such relationswere not identified in literature. However, it suggests that when analyzing the effectof contextual factors, we should not only think about harmonization of a process as awhole, but also about harmonization of each individual aspect of a process.

The third part of the model shows twelve elements of business performance that are


affected by changes in the level of process harmonization. The performance measuresidentified in the literature were further classified in three groups: strategic, tacticaland operational. They should be evaluated in a particular organizational unit in whichharmonization takes place.

The last element of the model are the arrows and their signs, which indicate thetype of relations between different parts of the model. For instance, an arrow with apositive sign between the level of harmonization and Efficiency, which is an operationalmeasure of business performance, shows that an increase in the level of harmonizationwill increase the level of efficiency at an operational level. In some cases we indicate twosigns; they refer to cases in which we found opposite relations suggested by differentauthors. For example, this is the case for costs. It is expected that the level ofharmonization will reduce costs at the operational level for the economies of scale thatcan be achieved, however there are also significant costs involved in the implementationof harmonization initiatives. Therefore it is not straightforward that a higher level ofharmonization will always produce a reduction in costs.

2.6 Conclusions

The main finding of this study is that contextuality shapes the way in which processharmonization takes place in an organization, and that contextual factors define theextent to which a set of business processes can be harmonized. The result of thischapter can be summarized in three main points.

Firstly, we found eleven contextual factors that determine the level of harmoniza-tion of a business process, to answer RQ1. What contextual factors determine theappropriate level of harmonization of a business process? These contextual factorswere classified in three groups (immediate, internal and external) based on the level ofthe organization that they characterize. Immediate factors are contextual factors thatdescribe a process that is being harmonized. Two immediate factors were identified:the level of process structuredness and personal differences in terms of experience andknowledge of resources. Internal factors are those that describe the internal environ-ment of an organization. Six internal contextual factors identified include: the numberof locations, level of centralization of IT governance, differences in product type, matu-rity level, organizational structure and number of mergers and acquisitions. Externalfactor are those beyond the control of an individual organization. We identified threethat are relevant to our study: cultural differences, different regulations and powerdistance with other organizations in the business network in which the organizationoperates.

Secondly, to answer RQ2. Which aspects of a business process are relevant to mea-sure the level of process harmonization? we found that measures of process harmo-nization identified in the literature are focused on the assessment of six main aspects:data, activities, control-flow, information technology, resources and management. Ac-tivities refer to the level of standardization of specific steps in the process. Control-flowmeasures the level of standardization of the sequence of activities. Data measures thelevel of standardization of input and output data used in the process. InformationTechnology refers to the level of standardization of IT systems. Management measuresthe standardization of the process assessment. And resources refers to the level ofstandardization of human resources involved in the process.

2.6. Conclusions 31

Finally, we developed a model that conceptualizes and operationalizes process har-monization, factors that drive the extent to which processes can be harmonized, perfor-mance benefits that can be gained through harmonization and relations between theseconcepts. Our conceptual model builds on existing theory about process harmonizationand is developed by means of a literature survey. It contributes to theory by show-ing relations that (hypothetically) exist between process standardization, contextualfactors that influence the extent of process standardization and the effect of processstandardization on organizational performance. Our conceptual model structures andsummarizes the body of knowledge in the area of process standardization. In additionthe chapter shows which hypotheses have been proven, and the validity and reliabilityof the different measures used to assess the extent of standardization. By doing so, itprovides clarity about the status of the research in this area and the opportunities forfurther research.

In terms of the managerial implications, the conceptual model that is presented inthis chapter helps managers to recognize different contextual factors that play a rolein successful process harmonization and to determine the type of influence that thesefactors can exert. Knowledge of these factors provides insights to build managerialguidelines to gain the expected benefits of process standardization.

The research presented in this chapter should be considered with the following lim-itations. Firstly, the conceptual model that is presented contains both proven andhypothesized relations. Future research is necessary to determine which of the hy-pothesized relations also hold. Secondly, it is difficult to compare the results obtainedfrom different studies considering that majority of them discuss their relations in adescriptive way without conducting any quantitative analysis. Therefore, in our modelwe cannot define which of the identified relations are more significant to provide bettermeans for control than others. In order to study that, a uniform study is necessarythat analyses the different relations in a uniform, comparable, manner. Thirdly, it ispossible that there is overlap between contextual factors. It is also possible that somecontextual factors have a moderating or mediating effect, rather than a direct effect.While we found some moderating and mediating effects in the literature, we did notincorporate these into the conceptual model, because a mediating or moderating effectwith respect to other contextual factors as well as overlap between factors, can onlybe shown in a study where all factors are considered in a uniform comparable manner.Finally, the search and selection of relevant literature was done by a single researcher.This can cause bias in the selection process. We reduced the risk of bias due to this,by presenting clearly verifiable inclusion and exclusion criteria.

Chapter 3

Measures of processharmonization

“Not everything that can be counted counts,and not everything that counts can be counted”

Albert Einstein

Given the importance of process harmonization attributed in the literature [49, 63],there is a need for measures to evaluate to what extent processes are harmonized.Performance measurement has critical importance for organizational management [38].Measures provide information about the current level of harmonization of a process inan organization, which compared to the expected level can be used to derive directionsfor improvements [41]. Previous research have failed to identify performance measuresto evaluate changes in the level of harmonization as a result of harmonization initiatives

Based on this motivation, the research question that we addressed in this chapteris RQ3. What measures can be used to quantify the level of process harmonization ofa business process? To answer this question, a set of operational measures to evaluatethe level of harmonization of business processes in an organization is developed andvalidated using an approach for construct measures and validation procedures thatis discussed in the next chapter. This set of operational measures is useful for twomain reasons: first, researchers can use them to develop normative theory based uponempirical investigation on process harmonization; and second, practitioners can havea mechanism for relating process performance in terms of the level of harmonization,to specific actions at an organization. For instance, InBev is a brewing company thatresulted from the merger between Belgium-based company Interbrew and Brazilianbrewer AmBev that occurred in 2004. In 2013, InBev realized 43.2 billion dollarsrevenue and employed more than 150,000 people based in 24 countries worldwide.The chief procurement officer in charge of the global operation intends to harmonizetheir ‘order to pay’ process to improve comparability among operations worldwideand to show only one face to the customers. In this situation, measures of processharmonization can help to identify which aspects of harmonization must be addressedto achieve the expected benefits of harmonization, i.e. focus on investments in a newIT systems to support the ‘order to pay’ process or on harmonization of the resourcesinvolved in this process.

34 Chapter 3. Measures of process harmonization

The remainder of this chapter is organized as follows: The next section presents theconceptual background that supports our approach for developing measures of processharmonization. Section 3.2 presents an overview of the methodology for developingmeasures of process harmonization, including a detailed description of the execution ofeach step (See Figure 3.2). Results are discussed in section 3.3. Section 3.4 concludesthe chapter presenting the main findings, limitations and directions for future research.

3.1 Conceptual background

A number of rationales have been brought forward to explain different decisions in-volved in the operationalization of concepts in strategic management research. In theprevious chapter, Section 2.3.2, we presented the operationalization of process harmo-nization found in the literature. We found that only 7 articles out of 35 identified as rel-evant in the literature of process harmonization, show some type of operationalizationof this concept. Types of operationalization of process harmonization that we foundinclude: (1) using attributes [112, 134] and (2) using indicators [8, 97, 99, 119, 146].Attributes were used to describe characteristics of an organization that has achievedharmonization to a certain extent, while indicators use a numerical scale to measurethe extent of process harmonization. We found that attributes have the problem ofa lack of commonality among scales used by different authors and also difficulties inthe quality assessment of these scales. These issues are better approached when usingindicators with a numerical scale, in which different methods can be used to evaluatethe validity and reliability of the measures developed.

Operationalization

Construct

Aspects

Indicators

Measurement model

Figure 3.1: Approach for the operationalization of process harmonization

Based on the previous observations, in this chapter we adopt an integral approach,based on the procedures proposed by Churchill’s [24] and Mackenzie et al. [87] todevelop a measurement model for assessing the level of harmonization of a businessprocess. Our approach is depicted in Figure 3.1, in which ‘process harmonization’

3.2. Methodology 35

is consider a construct. A construct is defined as a “conceptual definition of a vari-able” [121] and “an abstract entity that represents the non-observational state or natureof a phenomenon” [6]. Our operationalization of the construct consisted on identifyingdifferent aspects of process harmonization that is relevant for a further identificationof indicators to quantify each one of these individual aspects [144]. With the list ofmeasures identified, a hierarchical measurement model is built. This approach usesdifferent methods for data collection, including literature review, interviews with ex-perts, workshops and an online survey, to build and validate the measurement modelproposed. Details about the methodology are explained in the next section.

Our study extends existing literature by developing a hierarchical (second order)model, taking into account the most recent insights in academic literature, using PartialLeast Square (PLS) [7, 109]. This is the first measurement model, in which a distinctionis made between the different aspects of the concept of process harmonization in theliterature, providing new insights and directions for future research.

3.2 Methodology

The outline of the methodology followed is depicted in Figure 3.2. It is divided in twophases, an exploratory phase and a confirmatory phase.

Exploratory phase Confirmatory phase

Conceptual definition

Step 1Aspects

Model development

Step 2

Empirical dataData collection

Step 3

Measurement model - initial

Measurement model - finalModel

assessment

Step 4

Model validation

Step 5

Related measures

Figure 3.2: Methodology for developing measures of process harmonization


The goal of the exploratory phase is to build a model for measuring the level ofharmonization of a business process. In this phase two steps were defined. In thefirst step, the domain of the concept process harmonization is specified, by defining as-pects of the construct that must be considered for further measurement. The aspectswere identified using as an input 35 literature sources relevant in the domain of pro-cess harmonization, and the results of a workshop conducted with seven participants,including practitioners and academics with experience in the topic. This first step pro-vides two main outcomes: first, a list of relevant aspects of the construct under study;and second, measures of related concepts of harmonization, which are used to furthervalidate the measurement model developed with our approach. The second step of theexploratory phase consists of deriving measures for each aspect of process harmoniza-tion. Measures were derived by interviewing two experts in process harmonization andthree academics. As a result of the second step, a measurement model is built usingthe indicators derived in this step.

In the confirmatory phase, a sample of measures of process harmonization is empir-ically assessed using as an input 119 responses gathered by means of an online survey.In Step 3, an evaluation of the participant’s demographics is conducted. In step 4, fac-tor analysis is used to evaluate the number of dimensions that represent the measurespreviously derived. A revised measurement model is built, modifying the number ofdimensions and indicators included in the original model. The validity and reliabilityof the measurement model are evaluated in step 5. A detailed description of differentsteps previously mentioned is included in this section.

3.2.1 Conceptual definition

In this step, we derive a set of relevant aspects of process harmonization by means ofa workshop, in which the experience of a group of practitioners in procurement andacademics in process harmonization was the main mode of gathering data. However, “itis imperative that researchers consult the literature when conceptualizing constructsand specifying domains” [24]. Therefore, we used an initial set of measures derivedfrom previous literature as a starting point to further assess which aspects are alsorelevant from practical perspective during the workshop.

The first set of aspects was derived by means of a literature review described inchapter 2. It is appropriate to use the output of this literature review because weuse the same definition of the concept, in which process harmonization is defined as aselective standardization, which accepts different requirements for a process, depend-ing on their context. The methodology followed to conduct this literature review isdescribed in details in section 2.1.2. From the documents found in the literature, onlyresearch papers which present measures of the relevant concepts were selected. Firstly,documents collected were analyzed looking at phrases or words in the text to capturethe meaning of the concept as suggested in Strauss and Corbin [129]. Concepts usedto describe process harmonization were highlighted and linked to each document. Ta-ble 2.4 shows a concept matrix suggested in Webster and Watson [142], which was usedto summarize the findings and to identify which concepts define process harmonization.Secondly, measures of the most relevant concepts found in the previous step were iden-tified and summarized in Table 2.6. Each measure was classified looking at words inthe text that capture aspects of the process that they measure. For instance, measure14 ‘There is a fixed procedure for the collaboration between departments’ refers to the

3.2. Methodology 37

order in which the activities in the process are performed. In this example the keywordis ‘procedure’. Therefore, the aspect associated with this measure is ‘control flow’. Thesame analysis was done with all the measures indicated in Table 2.6 and indicated inthe third column ‘Aspect’. Finally, an initial set of aspects is defined to be furtherevaluated in the workshop.

A workshop was conducted to identify the set of aspects relevant for process har-monization from a practical perspective. The workshop included seven participantsorganized in two mixed groups with in total three academics and four practitioners.The selection of practitioners from only one specific process type ‘procurement’ wasmade to focus on identifying aspects and measures of process harmonization, ignoringdifferences about different processes. In this case, the process selected was well-knownby all the participants and the scope of the discussions were reduced to harmoniza-tion aspects. Participants were asked to identify factors that exert an influence in theharmonization level of a procurement process and specific aspects of process harmo-nization related to these factors. As part of the outputs of this workshop, a set ofaspects of harmonization was collected. The output of both, the literature review andthe workshop were compared to decide which aspects are relevant in theory and alsopractice. All the aspects identified in both, literature and workshop, were included asan input for next step. The output of this step is a set of the relevant aspects of processharmonization and a list of measures of the related concepts that are further used toassess the validity and reliability of the measurement model in step 5.

3.2.2 Model development

To evaluate how the different aspects previously identified really reflect different parts ofprocess harmonization, a measurement model was built and empirically tested. Firstly,interviews were conducted using a questionnaire with open questions to identify specificmeasures that have been used in companies involved in harmonization initiatives, toevaluate their level of harmonization improvements. Interviewees were also asked aboutthe motivation and expected benefits when conducting these initiatives, to identifywhich measures are significant in practice. Secondly, a meeting with academics wasperformed for brainstorming about ideas of more measures aligned with those gatheredin practice. The original list of measures was rewritten to guarantee that its wordingwas simple and precise, as suggested in the literature by Tourangeau et al. [133]. Thisprocess was not exhaustive because it was only intended to generate a first set ofmeasures which allows a quantitative validation of the aspects proposed. Finally, ahierarchical (second order) measurement model was build using the measures derivedin the previous steps.

Our measurement model was specified as a hierarchical factor model type II (re-flective first order and formative second-order). The second order construct is pro-cess harmonization, and the first order constructs are the relevant aspects defined insection 3.2.1. Measures are assigned to one specific aspect that they assessed. Ourmeasurement model is the initial hierarchical (second-order) model that will be em-pirically tested and validated in the confirmatory phase. Interviews were conducted


with two experts engaged in harmonization projects in large scale companies and threeacademics. One of the experts is a professional responsible to conduct a large scaleharmonization project in a governmental organization in the last six years. The otherexpert conducted a study about the impact of factors such as information technology inprocess standardization versus harmonization in value chain management. This studyincluded six multinational companies, in which two of them show successful results intheir harmonization efforts. The academics were exposed to literature and practicalexperiences in the topic. This group of professionals and academic experts was selectedconsidering their experience, to identify indicators of process harmonization that theyhave seen in practice in the companies studied and in previous literature on the topic.

3.2.3 Data collection

Empirical research was conducted to identify how much the current set of measuresactually reflects different aspects of process harmonization. An online survey wasconducted to achieve the goal of this step. The survey instrument is included in Ap-pendix C and it required around 20 minutes for completion. It was built using a Googledocument and the link for access was distributed to targeted respondents.

Targeted respondents for the online survey were professionals with experience inbusiness process management (BPM) practices. The survey instrument prepared wasdistributed in two languages: English and Dutch, targeting a different group of pro-fessionals within BPM. The Dutch questionnaire was sent to members of the BPMround table at Eindhoven University of Technology, with more than 300 members reg-istered when the survey was conducted. The questionnaire translated to English wasdistributed throughout BPM experts worldwide invited through advertisements madeon LinkedIn in a group called BPMInstitute.org with 1,311 members. It was also pub-lished though an internal communication portal used in an international consultancyfirm in The Netherlands (with 1,779 members registered). Professionals in BPM as-sumed the role of key informants because they provide information on an aggregateunit of analysis (in this case a harmonization project) by reporting on organizationalcharacteristics [39].

A set of criteria described in Dillman [39] were considered to design the question-naire. This set of criteria focuses on reducing the sources of error when conductingsurvey research, to be able to generalize sample results to a defined population. Thesources of error include: coverage error, sampling error, measurement error and non-response error. A coverage error is the result of all units in a defined population nothaving a known nonzero probability of being included in the sample drawn to repre-sent the population. The sampling error is the result of surveying a sample of thepopulation rather than the entire population. A measurement error is the result ofinaccurate responses that stem from poor question wording, poor interviewing, surveymode effects and/or some aspect of the respondent’s behavior. A non-response error isthe result of no response from people in the sample, who, if they had responded, wouldhave provided different answers to the survey questions than those who did respond tothe survey. This section discusses the decisions made to reduce the sources of errorsdescribed before when designing the survey instrument.

To reduce coverage errors, the survey was promoted through a group called BP-MInstitute.org using a social network (LinkedIn). Using the link provided to accessthe survey, respondents do not need any password or type of authorization, facilitating

3.2. Methodology 39

their access to the survey.

To reduce sampling errors, three big communities of BPM professionals were se-lected: BPM round table (300 members), BPMInstitute.org (1,313 members) and pro-fessionals from an international consultancy firm in The Netherlands (with 1,779 mem-bers registered). They received an invitation online to fill the survey with a link. Twoweeks after the initial invitation, they received a reminder to fill the questionnaire.This enhances the chances to reach a representative population, because they are morelikely to accept the invitation. Results of the questionnaire were gathered four weeksafter the first invitation. Additionally, the questionnaire was translated in English andDutch, considering that the members of the BPM round table mainly speak Dutch.Even though most of them are familiar with English; they are more willing to fill aquestionnaire in Dutch because it requires less effort on translation. This can guaranteethat we do not exclude a significant portion of the population for language barriers.

To reduce measurement errors, a pre-test was performed with a selected groupof academics with expertise in BPM (6) and in survey research (2). Furthermore, ananalysis of the responses was performed to check whether they understood the questionand provided an appropriate answer.

To reduce non-response errors, there are different design decisions that we consid-ered when building this survey in order to get respondents to sign on to the survey siteand to keep them motivated to complete the survey once begun. An invitation to fillthe questionnaire was prepared including the goal of this study, a description of ourcontribution to future research, and practical implications. The invitation is criticalin eliciting responses to web surveys [32]. The same text was included in the welcomescreen, making it more motivational for the respondents [39]. The first questions weresimple, just to assess the level of experience and knowledge that the respondents haveon process harmonization. Questions were stated very direct and using a simple lan-guage, using short line-length. They were also presented using a conventional formatsuch as paper questionnaires, recommended in the literature [39].

Each question has instructions on how to proceed, instead of providing them atthe beginning of the questionnaire. We did include in any question the option to forcepeople to answer before continuing to the next question, to give people the flexibilityto scroll from question to question if necessary. In this case we could manage to dothis considering that the order effect is not a major concern [39].

A few open questions were included to get information from the respondents. Openquestions are not recommended because it requires extra effort (cognitive and psy-chomotor) from respondents and they can increase the abandonment rate [32]. How-ever, in this case it is necessary for our research goal and it has been written thatpeople are more willing to answer open questions in web surveys compared to othertypes of surveys [39].

Following the decision of using open questions, we decided not to use a progresspointer. Basically, we consider that in our questionnaire with open questions, if weinclude a progress pointer based on number of items completed, this may produce thebackfire effect of increasing the abandonment rate [31, 32]. The main reason is becausewe may underestimate the duration of the survey due to the open questions. We madethis decision even though there are contradictory findings with respect to the effectof progress pointers used in surveys. Some authors arguing that indicating progressthrough the survey increase the motivation to complete the survey while reducingabandonment [39].


Another decision that we considered in the design of the survey instrument is thenumber of questions on a single screen. We decided to include only one question perscreen, to reduce the correlation in responses among items sharing one single screen [31],except for question 7 and 9 in which they are related. The downside of this decision isthat it can increase the time taken to answer the survey. This is not very critical in thiscase, considering that the survey is relatively short with few questions. In questions 4and 8, we used check boxes, but we restricted it to only a few options to avoid thatthe respondent only focus the attention on the first options ignoring the last ones [31].

In the survey besides the set of indicators derived in the previous phase, two con-structs were evaluated: business process standardization and business process complex-ity. Measures of these two concepts can be seen as reflective measures of process har-monization, because if the level of harmonization increases we can expect that the levelof standardization also increases while the process complexity decreases. They wereoperationalized using instruments provided in the literature and depicted in Table 3.1.Business process standardization (BPS) was operationalized by a 3-item instrument de-veloped by Munstermann et al. [99], which assess the level of standardization of a setof business processes. This was a measurement model selected in Step 1 to be used forfurther validation. The second construct operationalized is business process complex-ity (BPC). It was operationalized with a 5-item instrument developed by Schafermeyeret al. [119], and described as the level of difficulty reported by project managers andoperators, during process standardization or execution.

Table 3.1: Description of constructs and measures used in the survey.

Concept (construct) Item Measure Reference

Business ProcessStandardisation (BPS)

BPS1The execution of the business process isstrongly standardised.

BPS2We have documented all actions of thebusiness process to a great extent.

[99]

BPS3During the execution of the process wefollow a well-regulated process cycle.

Business ProcessComplexity (BPC)

BPC1The employees executing the businessprocess need to be able to flexible adjustthemselves to the differing processsequences.

BPC2The set of inputs necessary for processexecution differ often. [119]

BPC3The business process is characterized byuncertainty.

BPC4 The business process is very complex.

BPC5A lot of information is needed to executethe business process.

After the survey questionnaire was complete, a pre-test was conducted using cog-nitive interviewing [26]. The questionnaire used for the pre-testing is included in Ta-ble 3.2. The pre-test included six academics with experience in business process man-agement (BPM) and two with experience in conducting research in social science, toevaluate the survey instrument. After the results of the pre-test were collected, thesurvey questions were further improved and an additional check was performed but

3.2. Methodology 41

Table 3.2: Questionnaire for pre-testing the survey instrument.

Surveyquestions

Pre-test questions

For questions1,2,3 and 4

1.When we asked questions about your experience inharmonization projects:1.1.Did you have a particular time period in mind?(i.e. in the last five years)1.2.How well do you remember this information?

For question 52.When we asked to evaluate your knowledge aboutprocess harmonization, how did you estimate your answer?

For question 63.What do the terms process harmonization and processstandardization mean to you?

For question 7

4.What do the terms level of similarity and level ofcompatibility mean to you?5.How sure of your answer are you?6.Were you able to find your first answer to the questionfrom the response option shown?

For questions 8and 9

7.What does the term standard mean to you?8.How sure of your answer are you?9.Were you able to find your first answer to the questionfrom the response option shown?

For question 1010. How did you get your answer?11.How sure of your answer are you?12.How did you feel about answering this question?

Generalquestion

13. Do you have comments about the wording or howsome specific question is described? Please provide anyadditional comments to improve this survey instrument.

only with informal experts’ interviews. The online survey was sent with a remindertwo weeks later, to increase the respondent’s rate. The data was collected after fourweeks for further analysis.

Finally, when the data was collected we assumed that responses from people whodid not participate in the survey are not different than those from people who didparticipate. This assumption was represented using the following hypothesis, wheremu represents the mean value of a group of observations in the sample:

H0 :µrespondents = µnorespondents

H1 :µrespondents 6= µnorespondents

This assumption was tested conducting a non-response bias analysis to evaluatethe extent to which our sample represents the population of this study and the biasesobserve in the responses due to non-response [3]. An extrapolation method was carriedover successive waves of the survey. We identified two waves of questionnaire returns.Respondents in the first (early) wave are those who replied after the survey was madeavailable for the first time. The second (late) wave includes respondents who repliedafter a reminder was sent, two weeks after the first invitation to fill the survey.


The analysis includes the comparison of descriptive statistics of the demographicdata and comparison of key variables [46, 123]. The demographics considered were:the origin of respondents divided per language (Dutch and International sample), theexperience in process harmonization projects (yes/no, and the time of experience),and the role of respondents in harmonization projects. The key variables include theindicators of process harmonization evaluated in the survey and the comparison wasconducted using an independent sample two-tailed t-test with a level of significanceof 0.05. We compared differences in their perception about the extent to which theindicators derived in our measurement model actually reflect the level of harmonizationof a process. Additionally, we evaluate that the assumptions of the t-test were metusing Levene’s Test for Equality of Variances. In both tests, two paired t-test andLevene’s test, the null hypothesis can be rejected if the p-value calculated is smallerthan 0.05.

3.2.4 Model assessment

Factor analysis was used in this step to suggest the number of dimensions underlyingthe level of harmonization of a business process. Data collected in the previous stepwas used as an input for factor analysis. It consisted of responses gathered from 119professionals in BPM, using both the Dutch and English questionnaire. The answersprovided were their opinion based on previous experience or knowledge on the topic.They evaluated to what extent the provided measures actually assess an aspect specificof process harmonization. Their input was provided using a 5 points Likert scale.

Content validity of the items generated and the dimensionality of the constructunder study are evaluated using exploratory factor analysis (EFA), which is conductedusing the software package for statistical analysis (SPSS). Five steps were followed toconduct this exploratory factor analysis as suggested by Hair et al. [60]: (1) Examinethe factor loading matrix; (2) identify significant loadings in the matrix; (3) assesscommunalities; (4) Re-specify the factor model if needed; and (5) Label the factors.

There are three EFA decisions which are very important for the outcome of theanalysis: (a) the factor extraction model used; (b) the number of factors retained; and(c) the method used to rotate factors, if more than one factor is retained [28]. Commonfactor analysis is the factor extraction model selected. It is more appropriate than thecomponent factor analysis considering that the goal of this step is to identify latentdimensions represented in the original variables and we do not have previous knowledgeabout the amount of specific of error variance [60].

The number of factors to retain was based on the following three criteria: (1) Thelatent roots or eigenvalues should be greater than 1; (2) The conceptual foundation,indicating the underlying dimensions of the concept based on the theory; And, (3)the scree test criterion, which indicates the maximum number of factors that mustbe retained as the value where the inflexion point occurs. An orthogonal rotationalmethod was selected to rotate factors. Varimax was the method selected because it“has proved successful as an analytical approach to obtaining an orthogonal rotationof factors” [60].

3.3. Results 43

3.2.5 Model validation

In this step the internal consistency and reliability of the initial model is assessed. Butinternal consistency is not sufficient; we also evaluated whether our model correlateswith other related measure and whether it behaves as expected. The related measuresidentified in the literature during the conceptual definition were used to assess ourinitial measurement model, considering its hierarchical structure.

Data of the related concepts previously defined (process standardization and processcomplexity) was also collected in the online survey of the previous step. Partial leastsquare (PLS) was chosen for the analysis of this measurement model for three reasons:(1) it makes fewer demands regarding sample size than other methods; (2) It is able tohandle both formative and reflective indicators; and (3) it is better suited for theorydevelopment than for theory testing. Smart PLS software was used to measure thescale of measurement of the first and second order constructs.

The repeated indicator approach with mode B on the process harmonization con-struct and inner path weighting scheme is used to estimate the hierarchical latentvariable model [7]. This approach consists of using the indicators twice: (1) for thefirst-order constructs and (2) for the second-order construct (BPH). Having specifiedthe measurement model in this way, the path coefficients between the first and sec-ond order constructs represent the loadings of the second order latent variable [7, 86].“This approach produces generally less biased, and therefore, more precise parameterestimates and a more reliable higher-order construct score.” [7].

The evaluation of the measurement model include: (1) an assessment of the first-order reflective constructs using the constructs and measures loadings, t-values, AVE,composite reliability and discriminant reliability; and (2) for the second order for-mative construct, an evaluation of the measures weights, significance of weights andmulticollinearity among measures. The evaluation of the convergent validity is doneusing the using the constructs and measures loadings and weights and t-values. Thevalues of factor loadings and AVE should be above 0.50 and composite reliability above0.70 [60]. The discriminant validity for the first-order constructs is evaluated followingthe criterion described by Fornell and Larcker [48]. The criterion is that the squareroot of AVE should be greater than the variance shared between the construct andother constructs in the model [11]. The variance inflator factor (VIF) test was usedto evaluate multi-collinearity among measures, which is mostly used by authors in theliterature when using formative indicators [109].

A nomological network is used to assess the validity of the multidimensional struc-ture [87]. The direct effect of an antecedent of process harmonization on each aspect ofthe concept is measured. If they are equal, we consider this as a support of the multi-dimensional construct [42]. The antecedent construct selected in this study is businessprocess complexity, using the construct provided by Schafermeyer et al. [119]. Thefinal endogenous construct used to build the nomological network is business processstandardization, using the operationalization proposed by Munstermann et al. [99].

3.3 Results

This section presents the results organized in five subsections based on the outputs ofthe different steps described in the process depicted in Figure 3.2. The output of thefirst step is a list of relevant aspects of process harmonization use to derived further


measures. The initial set of measures of process harmonization is generated in step 2,and they were used to build our initial measurement model. As a result of the thirdstep we gathered a set of empirical data to generate scores for each of the measuresincluded in the initial model. These scores are used for further evaluation of the validityand reliability or our measurement model. With these scores, a refined measurementmodel is generated in step 4, and further validated in step 5. The final result of theprocess is a revised and validated measurement model of process harmonization.

3.3.1 Relevant aspects

Table 2.6 shows the different measures and the aspects of a business process standard-ization identified in the literature. The list of aspects includes: activities, control flow,data, information technology, management and resources. Activities refer to the levelof standardization of specific steps in the process. Control-flow measures the level ofstandardization of the sequence of activities. Data measures the level of standardiza-tion of input and output data used in the process. Information Technology refers to thelevel of standardization of IT systems. Management measures the standardization ofthe process assessment. And resources refers to the level of standardization of humanresources involved in the process.

The aspects derived from the literature were compared with those identified duringthe workshop. Table 3.3 include a list of all the aspects of process harmonization iden-tified using the literature and two professional teams who participated in the workshop.Aspects identified by at least two sources were included.

Table 3.3: Aspects selected based on literature and practice.

AspectsSource

Literature Workshopteam1 Workshopteam2

1. Activities X X2. Control-flow X X X3. Data X X4. Information Technology X X X5. Management X6. Resources X X

Activities and control-flow are considered together, because in some cases was notclear to what extent a measure evaluates an individual activity or the collection of ac-tivities with a predefined order. Therefore, both activities and the relationship betweenthese activities are included in one aspect.

3.3.2 Initial measurement model

A first set of eight measures of process harmonization was gathered as a result of thisstep and is depicted in Table 3.4. Table 3.4 shows the measures related to each oneof the aspects selected. This initial set includes one indicator for Activities (IA1), tworelated to Data aspect (ID1 and ID2), three defined for Information Technology (IT1,IT2 and IT3) and finally, two related to Resources (IR1 and IR2).

3.3. Results 45

Table 3.4: Measures of process harmonization.

Aspects Item Measure

Activities IA1 The percentage of common activities in the process

DataID1 The number of different documents used as

input for the same processData ID2 The number of different output reports

Information technologyIT1 The number of different software applications

in used in the process

Information technologyIT2 The number of different supplier’s paid for

the software applications

Information technologyIT3 The amount of money paid for the software

applicationsResources IR1 The percentage of common roles in the processResources IR2 The number of different roles executing the same activity

Using this set of measures we built a measurement model which evaluates the levelof harmonization of a business process in an organization. This initial measurementmodel derived is depicted in Figure 3.3.

ID1

ID2

IA1

IR1

IR2

IT1

IT2

IT3

DATA

ACT

RES

IT

BPH

First order construct

Second order construct

Indicator

Figure 3.3: Initial hierarchical model (reflective first order and formative second order)

Figure 3.3 shows that the level of harmonization is composed by four different as-pects as identified before (data, activities, resources and IT). Each one of these aspectsis measured through a set of indicators. The indicators are represented as reflectivebecause within each aspect, they are expected to be correlated. As a result, a hierar-chical model was derived, which is reflective first order (for DATA, Activities (ACT),Resources (RES) and IT) and formative second order (business process harmonization-BPH). Empirical research is conducted in the next step to assess the number of differentdimensions underlying this model and its structure.


3.3.3 Empirical data

Table 3.5 shows the demographics of participants divided in two groups early and laterespondents. In total, 119 responses were gathered. They are composed by samplegroups: the Dutch sample and the International sample. The response rate for theDutch sample is 16.7%, with 50 complete surveys out of a targeted population of 300members of the BPM round table by the time the survey was conducted; And 2.23% forthe International sample with 69 responses out of a population of 3,092 members (1,313form the LinkedIn community and 1,779 from an international consultancy company).

Table 3.5: Participants demographics divided early and late respondents.

Demography CategoryRespondents (early) Non-respondents (late)Frequency % Frequency %

LanguageDutch 38 38,0 12 63,2English 62 62,0 7 36,8

ExperienceYes 53 53,0 13 68,4No 47 47,0 6 31,6

Experiencetime

None 47 47,0 6 31,6Less than 1 year 30 30,0 8 42,11 to 3 years 20 20,0 5 26,3More than 3 years 3 3,0 0 0,0

RoleProject manager 50 50,0 0 0,0Manager 2 2,0 4 21,1Process architect 20 20,0 4 21,1Business analyst 8 8,0 1 5,3Consultant 14 14,0 2 10,5Other 6 6,0 8 42,1

Our response rate of 16.7% in the Dutch sample slightly exceeds the 10 - 12% ratethat Hambrick et al. [62] describe as typical for surveys of executives. However, forthe International sample this response rate was significantly reduced to 2.23%. Eventhough the survey was implemented following guidelines such as the ones suggested byDillman [39], we expected to obtain a low response rate due to the subject of the survey.A difficulty to obtain a higher response rate is that not everyone in the sampling framemay be familiar with the subject of the survey. They are professionals in business pro-cess management (BPM) but not necessarily focused on process harmonization withinBPM. For the Dutch sample we have closer contact with the potential participantsand have more control over their expertise in BPM, compared to the Internationalsample. Therefore, potential respondents which are unfamiliar with the topic may notfeel interested or uninformed to provide enough input to the survey

Comparing the level of experience between both samples, we can observe that thepercentage of people with experience in harmonization projects increased by 15% in thelate response sample compared to the early response. However, the time of experiencewas reduced. It can be observed that in the late response sample none of the respon-dents had more than three years of experience. Significant differences were observedin terms of the role in both samples. The majority of early respondents were projectmanagers while in the late response sample these group is distributed among managersand others. It is possible that the term project manager is too specific and therefore,several terms for managers with the same type of function can be used. However, in

3.3. Results 47

terms of the role we can conclude that significant differences are registered in the earlyand late respondent’s sample.

After observing differences in demographic characteristics of both samples, we de-cided to evaluate to what extent these differences influence their responses. The de-scriptive statistics for all the responses are shown in Table 3.6 and Table 3.7.

Table 3.6 shows the descriptive statistics of each of the indicators collected fortwo groups (1- early and 2- late respondents). For each of the indicators, the mean,standard deviation and standard error of the mean was calculated. They are furtherused for the t-test.

Table 3.6: Group statistics of responses (ID1, ID2, IA1, it1, IT2 and IT3) for early (1) early respon-dents and (2) late respondents.

Resp. Statistic Value

Bootstrap a

BiasStd.error

95 %confidence intervalLower Upper

ID1 1 N 100Mean 3,22 0,00 0,12 2,98 3,46Std. Deviation 1,168 -,011 0,095 0,954 1,341Std. Error Mean 0,117

2 N 19Mean 2,79 0,00 0,23 2,33 3,27Std. Deviation 1,032 -0,030 0,147 0,717 1,286Std. Error Mean 0,237

ID2 1 N 100Mean 3,42 0,00 0,11 3,19 3,63Std. Deviation 1,093 -0,012 0,084 0,905 1,242Std. Error Mean 0,109


IA1 1 N 100Mean 3,50 0,00 0,12 3,26 3,72Std. Deviation 1,150 -0,011 0,095 0,954 1,316Std. Error Mean 0,115


IT1 1 N 100Mean 3,53 0,00 0,09 3,34 3,70Std. Deviation 0,881 -0,011 0,091 0,699 1,051Std. Error Mean 0,088




Resp. Statistic Value

Bootstrap a

BiasStd.error

95 %confidence intervalLower Upper

IT2 1 N 100Mean 2,34 0,00 0,10 2,13 2,54Std. Deviation 1,017 -0,010 0,064 0,879 1,142Std. Error Mean 0,102


IT3 1 N 100Mean 2,06 0,00 0,09 1,88 2,23Std. Deviation 0,862 -,008 0,060 0,739 0,971Std. Error Mean 0,086

2 N 19Mean 2,00 0,01 0,11 1,78 2,21Std. Deviation 0,471 -,029 0,115 0,224 0,649Std. Error Mean 0,108

IR1 1 N 100Mean 3,74 0,00 0,10 3,54 3,92Std. Deviation 0,939 -0,022 0,122 0,677 1,155Std. Error Mean 0,094


IR2 1 N 100Mean 3,37 0,00 0,10 3,16 3,57Std. Deviation 0,991 -0,011 0,080 0,838 1,146Std. Error Mean 0,099


aUnless otherwise noted, bootstrap results are based on 1000 bootstrap samples

The results depicted in Table 3.7 show that for all the indicators, the significancep-value) of Levene’s test is greater than 0.05, except for IT3 which is 0.007. Theseresults indicate that the assumption that the variances are equal cannot be rejectedand therefore, the results of the t-test for equal variance should be conducted. Onlyfor IT3 we considered the results for equal variance not assumed. The results of thet-test for all the indicators show a p-value greater than 0.05. Therefore, no significantdifferences are detected, and at 0.05 level of significance, the null hypothesis H0 thatearly and late responses are not different cannot be rejected, and we consider that thereis no response bias in the results obtained.

3.3.

Resu

lts49

[H]

Table 3.7: Results of independent sample t-test to compare the responses of (1) early and (2) late respondents, with respect to all the indicators evaluatedduring the survey.

Equalvarianceassumed?

Levene’s test for t-test for equality of meansequality of variances

FSignificance(p-value)

tdegreesoffreedom

Signif.2-tailed(p-value)

Meandifference

Std.errordifference

95 % confidenceinterval of the

differenceLower Upper

ID1 Yes 0,240 0,625 1,498 117 0,137 0,431 0,287 -0,139 1,000No 1,631 27,546 0,114 0,431 0,264 -0,111 0,972

ID2 Yes 0,480 0,490 -0,587 117 0,558 -0,159 0,271 -0,695 0,377No -0,617 26,535 0,543 -0,159 0,258 -0,688 0,370

IA1 Yes 0,190 0,663 -0,279 117 0,781 -0,079 0,283 -0,639 0,482No -0,303 27,511 0,764 -0,079 0,260 -0,612 0,455

IT1 Yes 1,443 0,232 -1,195 117 0,234 -0,259 0,217 -0,689 0,171No -1,291 27,311 0,208 -0,259 0,201 -0,672 0,153

IT2 Yes 3,673 0,058 0,520 117 0,604 0,129 0,249 -0,363 0,622No 0,586 28,608 0,562 0,129 0,221 -0,323 0,582

IT3 Yes 7,530 0,007 0,294 117 0,769 0,060 0,204 -0,344 0,464No 0,434 44,876 0,667 0,060 0,138 -0,219 0,339

IR1 Yes 0,394 0,531 -0,218 117 0,828 -0,049 0,227 -0,499 0,400No -0,262 31,183 0,795 -0,049 0,189 -0,434 0,335

IR2 Yes 0,425 0,516 0,852 117 0,396 0,212 0,249 -0,281 0,705No 0,838 24,975 0,410 0,212 0,253 -0,309 0,733


3.3.4 Model dimensions

The analysis is performed with 119 responses, which constitutes more than 14:1 ratioof observations per variable. This ratio is more than adequate, a minimum of 50observations and 5:1 ratio [60], for the calculation of the correlations between variables.The overall measure of sampling adequacy (MSA) of 0.761, and above 0.50 for eachindividual variable (0.775, 0.798, 0.793, 0.844, 0.679, 700, 729 and 840), indicate thatsufficient correlation exists among variables and therefore it is appropriate to proceedwith the factor analysis. An exploratory common factor analysis was conducted toexamine whether the number of dimensions conceptualized could be verified empirically.Table 3.8 shows that the total variance can be explained by three factors which canexplain 68.67% of the total variance. They also have an eigenvalue of 1 or more.

A three factor structure was considered for evaluation as suggested in the previousstep. The next step was to identify the factor structure matrix shown in Table 3.10.For a sample size of 119, factor loadings above 0.50 should be considered significantbased on a 0.05 significance level, a power of 80% an standard error assumed to betwice as those from correlation coefficients [60].

Table 3.8: Total variance explained.

Factor

Initial Extraction sums ofRotations sumsof squared load.(Total a)

eigenvalue squared loadingsTotal % of

varianceCum.%

Total % ofvariance

Cum.%

1 3,285 41,058 41,058 2,842 35,520 35,520 2,1762 1,209 15,107 56,164 0,817 10,207 45,728 2,0023 1,000 12,505 68,669 0,567 7,089 52,816 1,8704 0,778 9,731 78,4005 0,605 7,559 85,9606 0,436 5,454 91,4147 0,388 4,852 96,2658 0,299 3,735 100,000

Extraction Method: Principal Axis Factoring.aWhen factors are correlated, sums of squared loadings cannot be addedto obtain a total variance.

In the factor structure matrix depicted Table 3.9a, a clear factor structure cannotbe identified, considering that IA1 and IR2 show high loadings on two factors. Thissuggested the need to delete some of the factors and re-evaluate the structure (SeeTable 3.9b) .

To decide which factor should be removed we observed at the communality levelin Table 3.10a. The indicator which shows the lowest communality level is IT1 with0,525.

This can also be observed in the correlation matrix in Table 3.11, in which this indi-cator did not show a significant correlation with any of the other indicators. Therefore,IT1 was removed from the model and again the factor structure matrix and commu-nality level were assessed.

3.3. Results 51

Table 3.9: Factor Structure Matrix:

(a) Initial model.

(a)Component1 2 3

IT3 0,860 0,125 0,156IT2 0,859 0,197 0,075ID1 0,182 0,863 0,038ID2 0,274 0,802 0,189IA1 -0,070 0,614 0,536IR1 0,111 0,184 0,783IT1 0,149 0,060 0,707IR2 0,514 0,124 0,515

(b) Final model.

(b)Component

1 2 3

ID1 0,846 0,191 0,024ID2 0,802 0,280 0,161IA1 0,669 -0,0690 0,45IT2 0,216 0,872 0,021IT3 0,112 0,852 0,212IR1 0,189 0,067 0,892IR2 0,106 0,479 0,629

Rotation method: Varimax with Kaiser normalization

Table 3.10: Communality using a principal component analysis:

(a) Initial model.

(a) Initial Extraction

ID1 1,000 0,780ID2 1,000 0,753IA1 1,000 0,669IT2 1,000 0,782IT3 1,000 0,780IR1 1,000 0,659IR2 1,000 0,545IT1 1,000 0,525

(b) Final model.

(a) Initial Extraction

ID1 1,000 0,753ID2 1,000 0,748IA1 1,000 0,655IT2 1,000 0,807IT3 1,000 0,784IR1 1,000 0,836IR2 1,000 0,636

52C

hap

ter3.

Measu

resof

pro

cessh

armon

ization[H]

Table 3.11: Correlation matrix including eight indicators of initial measurement model.

ID1 ID2 IA1 IT1 IT2 IT3 IR1 IR2

ID1 Pearson Correlation 1,000 0,605 0,409 0,166 0,276 0,263 0,232 0,280Sig. (1-tailed) 0,000 0,000 0,035 0,001 0,002 0,006 0,001

ID2 Pearson Correlation 0,605 1,000 0,466 0,250 0,381 0,323 0,325 0,349Sig. (1-tailed) 0,000 0,000 0,003 0,000 0,000 0,000 0,000

IA1 Pearson Correlation 0,409 0,466 1,000 0,339 0,194 0,175 0,428 0,223Sig. (1-tailed) 0,000 0,000 0,000 0,017 0,028 0,000 0,007

IT1 Pearson Correlation 0,166 0,250 0,339 1,000 0,268 0,199 0,304 0,276Sig. (1-tailed) 0,035 0,003 0,000 0,002 0,015 0,000 0,001



IR1 Pearson Correlation 0,232 0,325 0,428 0,304 0,157 0,295 1,000 0,435Sig. (1-tailed) 0,006 0,000 0,000 0,000 0,044 0,001 0,000

IR2 Pearson Correlation 0,280 0,349 0,223 0,276 0,372 0,410 0,435 1,000Sig. (1-tailed) 0,001 0,000 0,007 0,001 0,000 0,000 0,000

3.3. Results 53

As a result, a three factor structure is proposed considering the clear structure, inwhich all the variables have high loadings only with a single factor and the minimumlevel of communality of all the factors is 0,636. The group of variables grouped inthree factors is label based on the type of indicators that they represent. A revisedmeasurement model is illustrated in Figure 3.4, composed by three factors: ACT-DATA, RES and IT. ACT-DATA is composed by the indicators IA1, 1D1 and 1D2;RES by IR1 and IR2 and IT by IT2 and IT3.

ID1

ID2

IA1

IR1

IR2

IT2

IT3

ACT‐DATA

RES

IT

BPH

First order construct

Second order construct

Indicator

Figure 3.4: Revised hierarchical model (reflective first order and formative second order)

3.3.5 Final measurement model

The repeated indicator Mode B PLS-SEM model used for our analysis is depicted inFigure 3.5. It shows how the indicators of each individual first-order construct (Data,Resources and IT) are repeated in the second-order construct (BPH). An exogenousconstruct, business process complexity (BPC), is linked to both first and second orderconstructs. And finally, a construct of business process standardization (BPS) is usedas a final endogenous construct, considering that its indicators (IS1, IS2 and IS3) canbe used as reflective indicators of BPH.

The results gathered from the assessment of the reflective first-order construct aresummarized in Table 3.12 and Table 3.13. Table 3.12 shows the loadings, t-values, aver-age variance extracted (AVE) and composite reliability values for all the indicators andfirst-order constructs. The results of convergent validity are satisfactory, consideringthat all the loadings are above 0.50, with a minimum value for the reflective indicatorsof 0.71(IR1). All the reflectively measured first-order constructs (ACT-DATA, RESand IT) showed satisfactory values for convergent validity and reliability, with an AVE<0.50 and composite reliability above 0.70. This suggests that the indicators accountfor a large portion of the variance of each latent construct.


ID1

ID2

IA1

IR1

IR2

IT2

IT3

0,228

0,131

0,169

0,995

ACT‐DATA

RES

IT

BPH

ID1

ID2

IA1

IR1

IR2

IT2

IT3

0,265

BPS

0,000

BPC

IS3IS2IS1

IC4IC3IC2IC1 IC5

0,915 0,727 0,714

0,836

0,980

0,887

0,718

0,896

0,923

0,726

0,554 0,518 0,785 0,9870,940

0,411

0,362

0,478

0,459

0,485

0,307

0,515

0,158

0,257

0,088

0,258

0,341

0,172

0,224

Figure 3.5: Revised hierarchical model (reflective first order and formative second order)

3.3.

Resu

lts55

[H]

Table 3.12: Results of the assessment of reflexive first-order construct: evidence of convergent validity

Constructs and Measures Loadings t-value AVECompositereliability

Activities and Data (ACT-DATA) 2.3466 0.6600 0.8529

IA1-The percentage of common activities in the process. 0,7366 11.2811ID1-The number of different documents used as input for thesame process.

0,8485 43.6912

ID2-The number of different output reports. 0,8469 22.4686

Resources (RES) 2.4602 0.7228 0.8389

IR1-The percentage of common roles in the process. 0,8165 9.4556

IR2-The number of different roles executing the same activity.0,8826 28.7927

Information Technology (IT) 2.2413 0.8339 0.9094

IT2-The number of different supplier’s paid for the softwareapplications.

0,9066 29.8478

IT3-The amount of money paid for the software applications. 0,9197 69.6419

56C

hap

ter3.

Measu

resof

pro

cessh

armon

izationTable 3.12 – Continued from previous page

Constructs and Measures Loadings t-value AVECompositereliability

Business Process Standardization (BPS) 6.1067 0.5462 0.7801

IS1-The execution of the business process is strongly standardized. 0,8048 13,8649IS2-We have documented all actions of the business process to agreat extent.

0,5942 3,959

IS3-During the execution of the process we follow awell-regulated process cycle.

0,7987 10,9688

Business Process Complexity (BPC) 6.8837 0.5085 0.8314

IC1-The employees executing the business process need to beable to flexible adjust themselves to the differing processsequences

0,5117 3,9149

IC2-The set of inputs necessary for process execution differ often 0,5046 3,6915IC3-The business process is characterized by uncertainty 0,8210 25,2406IC4-The business process is very complex 0,8066 23,613IC5-A lot of information is needed to execute the business process 0,8374 28,1152

3.4. Conclusions 57

Table 3.13: Correlation of latent variables and evidence of discriminant validity.

Construct BPC BPS DATA IT RES

BPC 0,7131BPS 0,5761 0,7391DATA 0,4686 0,3985 0,8124IT 0,4170 0,3590 0,3699 0,9132RES 0,3547 0,4289 0,4575 0,4222 0,8502

The R-squared of the measurement model of 0.995 shows that the formative higherorder construct (BPH) is explained by its components (ACT-DATA, RES, IT). Lookingat the t-values at the first-order construct level, all the weights are significant at a 0.05level (t >1.96).

Moreover, the variance inflator factor (VIF) used to test multi-collinearity amongmeasures provided values between 1.20 and 1.96 which is below the threshold of 3.Table 3.13 shows the correlations between first-order construct and the square root ofaverage extracted variance (AVE) highlighted in bold. It shows evidence of discriminantvalidity, because the squared root of AVE for all the first-order constructs is greaterthan their correlation with other constructs in the model.

Collectively, the results have shown that the entire measurement model derivedsatisfy all the requirements for validity and reliability.

3.4 Conclusions

The result of this study is a validated measurement model to assess the level of harmo-nization of business processes in an organization. This measurement model providesthe necessary information to answer RQ3. What measures can be used to quantify thelevel of process harmonization of a business process?

A set of seven measures can be used to evaluate the level of process harmonization inan organization. These measures assess four different aspects of process harmonization,including: Data, Activities, Resources and Information Technology. This set includestwo indicators related to Data aspect (‘ID1: the number of different documents usedas input for the same process’ and ‘ID2: the number of different output reports’), oneindicator of Activities (‘IA1: the percentage of common activities in the process’), twoindicators related to Resources (‘IR1: the percentage of common roles in the process’and ‘IR2: the number of different roles executing the same activity’) and two relatedto the Information Technology aspect (‘IT2: the number of different supplier’s paid forsoftware applications’ and ‘IT3: the amount of money paid for software applications’).

These seven measures were used to build a hierarchical measurement model, whichis reflective in the first order and formative second order. The first order constructsinclude relevant aspects of process harmonization: Data combined with Activities(ACT-DATA), Resources (RES) and Information Technology (IT). The second or-der construct is process harmonization which is formative with respect to the differentaspects. The interpretation of that model is that the concept of process harmonizationhas different dimensions (that we call aspects), which can be assessed using differentmeasures. Results discussed throughout this chapter provide evidence that the modelderived is a valid and reliable measurement model for process harmonization.


Our measurement model has certain strengths but also exhibits limitations. Asignificant strength of our construction process is the multi-method approach followed,especially in the conceptualization step. It enriches the construct providing views fromliterature together with practice. The literature provided focus on the main aspects ofthe construct that were used for further operationalization. Using these initial set ofaspects, practitioners defined measures from their practical experience.

A limitation of this study is that we do not have a large number of indicators peraspect. This can lead to an under-specification of aspects that must be considered inthe assessment of process harmonization. The measurement model cannot be evaluatedcontaining only one indicator per aspect. Therefore, in our current study a maximumnumber of four aspects could be identified with the initial set of indicators generated.This is the case with the indicator IA1 which was theoretically defined as an indicatorof Activities, and during the confirmatory phase was further combined with indicatorsof Data. It was not a surprise that it was combined with Data considering that thecorrelation between the harmonization of the steps followed in a process and the datainput or output used, is stronger than the correlation of data with the number ofresources used in the process or IT systems.

By conceptually developing and empirically validating the process harmonizationconstruct, this study conveys an important message that effective harmonization initia-tives requires an evaluation of the level of process harmonization achieved at differentstages in the process to be able to link these measures with specific improvements inthe organization. Our measurement model of process harmonization can be used byresearchers and company executives to guide future research and practice. Researcherscan use this systematically developed and validated measurement model as a startingpoint in the examination of the effects of process harmonization on business perfor-mance, or the link between organizational factors and the level of harmonization. Inpractice, this measurement model can assess the ratio of input and output of a singleprocess harmonization program. This may serve to justify investments in these typesof improvement programs. The conventional approach of simply investing in IT doesnot automatically lead to success of harmonization initiatives. One of the results ofour research is that IT related measures only evaluate one aspect of the level of processharmonization.

Chapter 4

Relationship betweencontextual factors and processharmonization

“Ignorance more frequently begets confidence than does knowledge: it is those whoknow little, not those who know much, who so positively assert that this or that

problem will never be solved by science”Charles Darwin

Following the contingency theory that frames our research, in this chapter we elab-orate on the type of relations between contextual factors and the level of harmonizationof different aspects of business processes in an organization. We argue that contextualfactors are determinants of the level of harmonization of a process. We also believethat the type of effect may differ for different aspects of the process.

We use the results gathered from the previous two chapters as an input to explorerelations and to develop a more unified view of the role of contextual factors in processharmonization. Several limitations of the current literature in process harmonizationare addressed in this chapter. First, we explore different aspects of process harmoniza-tion. Second, we include characteristics of one of the factors (organizational structure)for a deeper understanding of the specific aspects playing a role in process harmoniza-tion and the type of influence that they exert. Finally, we explore multiple factorstogether.

Therefore, the final study in this thesis aims to answer the following research ques-tion: RQ4 - How do contextual factors influence the level of harmonization of eachindividual aspect of process harmonization? A multiple case study is conducted toexplore these relations, becuase it provides a variety of empirical evidence to derivepropositions. We use data from six cases to further derive a set of propositions ex-plaining the relationship between contextual factors identified as relevant and the levelof harmonization of each individual aspect of the process. Recognizing multiple as-pects of process harmonization opens the possibility that contextual factors may exerta different influence on each individual aspect. In practice, this knowledge provideguidance to formulate the appropriate strategies for exploiting the expected benefits ofprocess harmonization.

60 Chapter 4. Relationship between contextual factors and process harmonization

The chapter proceeds as follows. The next section describes the initial conceptualmodel that is used to explore possible relations, defining the factors and aspects ofprocess harmonization which are included in the analysis. Section 4.2 presents themethodology, providing details of the case study design, covering the selection of cases,data collection and a description of the data analysis. The results including a set ofpropositions derived are shown and discussed in Section 4.3. An evaluation of valid-ity and reliability aspects are described in detail in Section 4.4. Finally, Section 4.5concludes the chapter, addressing the limitations of this study and raising additionalquestions for future research.

4.1 Conceptual background

The conceptual model of this study is based on the work done in Chapter 2 andChapter 3. The empirical analysis described in this chapter is based on the conceptualmodel for explaining the determinants and effect of process harmonization depictedin Figure 2.3. The measures derived in Chapter 3 are used to assess the level ofharmonization of different aspects of the processes studied. Figure 4.1 summarizes thelist of contextual factors and aspects of process harmonization that are included in ourassessment.

Level of harmonization

Data

Resources

Activities

Information technology

Product type

IT Governance centralization

Maturity level

Number of different locations

Number of mergers and acquisitions

Contextualfactors

?

Rule observation

Job codification

Specificity of job

Written communication

Participation in decision making

Hierarchy of authority

Departmental participation

Level of process structuredness

Organizational structure centralization

Organizational structure formalization

Figure 4.1: Initial model for exploring the type of relations between contextual factors and the levelof process harmonization

The list of contextual factors included in this model are based on leading articles onprocess harmonization, which were gathered by means of an structured literature re-view described in Chapter 2. Therefore, it is an appropriate starting point to assess the

4.1. Conceptual background 61

link between these factors already identified in the literature and the different aspectsof process harmonization. Only the relations between contextual factors and the levelof harmonization are evaluated in this chapter, considering that the effect of processharmonization is well acknowledged in the literature [98, 104, 106, 146, 149]. Fromthe initial set of contextual factors summarized in the conceptual model depicted inFigure 2.3, three of them (‘cultural differences’, ‘different regulations’ and ‘power dis-tance’) are excluded from the analysis because we did not have the necessary resourcesto conduct the extensive data collection required. The reasons for their exclusion arethe following: all the companies included in the sample are large organizations, andthe people interviewed are all located in one country. Even though different culturalbackgrounds could be identified in the people interviewed, we do not have the neces-sary variety to compare cultural differences with our sample. Considering that all theorganizations in our sample are large and operate in different domains, they have alarge number of different regulations that cannot be compared among them. In thiscase, only the number of regulations will not provide any insights about this contextualfactor. The same justification applies for the exclusion of ‘power distance’. As a re-sult, the list of contextual factors that are empirically evaluated in this study includes:number of different locations, IT governance centralization, product type, maturitylevel, number of mergers and acquisitions, organizational structure and level of processstructuredness, as depicted in Figure 4.1.

Figure 4.1 shows two different dimensions of the organizational structure that werenot differentiated in the original model. In the original model (See Figure 2.3) onlythe bureaucracy of the organizational structure was associated with the level of processharmonization. However, for a deeper understanding of the dynamics of this relation,we decided to explore the constitutive dimensions of bureaucracy identified in the liter-ature. Although many studies in organizational research have generated different andsometimes divergent indicators to tap the same concept, the leading reports are thoseof Hage and Aiken [59], Lawrence and Lorsch [81], Brewer [13], Perrow [105], Blauand Schoenher [9] and Mintzberg [94]. Therefore, we choose the two most highlighteddimensions of organizational structure defined by Hage and Aiken[59] and Perrow [105]and described in Pennings[103]. The aspects of centralization used in our model in-cludes: personal participation in decision making, hierarchy of authority and depart-mental participation in decision making. Four aspects of the organizational structureformalization are assessed, including: job codification, job specificity, rule observationand written communication. They are all included in the model as shown in Figure 4.1.We decided to use the measures of centralization and formalization provided by theseauthors because they have been used in previous research and they have shown ade-quate levels of validity and reliability.

Finally, four aspects of process harmonization are included in the analysis: Activi-ties, Data, Resources and Information Technology. Therefore, indicators of control-flowand activities are considered together, and no indicator of management was includedin the analysis. A detailed description of the different factors and the measures of thelevel of process harmonization are provided in the next section.


4.2 Methodology

A multiple case study methodology was adopted in this study to identify relationsbetween contextual factors and the level of harmonization. A case study approach isappropriate in this study because we do not only want to investigate what factors exertan influence in the level of harmonization, but also the mechanisms underlying theserelations. It supports a deep and detailed investigation that is necessary to answerhow these factors exerted an influence on the harmonization level of a process in anorganization [114].

A multiple case study methodology was chosen because it enable comparisons thatclarify whether an emergent finding is simply idiosyncratic, related to a single case orconsistently replicated by several cases [44]. The propositions derived using multiplecases are more deeply grounded in a variety of empirical evidence [45].

The outline of the different steps followed to conduct a multiple case study is de-picted in Figure 4.2. It is an iterative process which is divided in three phases: (1)Sampling, (2) Data collection and (3) Data analysis. The following subsections describein detail each one of the seven steps shown in Figure 4.2.

Sampling Data Collection

Selection of cases

Step 1

Preparation of cases

Step 3

Within case analysis

(for each case)

Step 6

Data Analysis

Execution of interviews

Step 4

Study of additional

documentation

Step 5

Cross‐case analysis

Step 7

Definition of information sources

Step 2

Figure 4.2: Approach for conducting a multiple case study

4.2. Methodology 63

4.2.1 Selection of cases

The sample consists of six comparative cases representing a diverse theoretical sam-pling [43], in which cases are companies selected to produce contrasting results forpredictable reasons. Even though there is no ideal number of cases, a number between4 and 10 cases is considered appropriate for this type of research [43].

Multiple cases can be regarded as equivalent to multiple experiments. The morecases that can be marshaled to establish or refute a theory, the more robust are the re-search outcomes [45, 114]. However, it was also a convenience sample with accessibilityas a deciding factor. In the current study, attempts were made to find a reasonablebalance about the number of cases necessary to provide enough variety for patternsidentification. We contacted eight companies in total, but only six were included inour study considering the limitations in the number of resources available to conductthe interviews (a single researcher) and the number of companies willing to cooperate.

Table 4.1 provides a description of the six companies studied. The variety in thesample is regarding the context in which the different companies investigated operate,including: industry type, company age and company size. The industries in which theyoperate include: (1) manufacturers of goods such as vehicles and cleaning equipment,and (2) service providers, such as healthcare and logistics providers. The third col-umn indicate the company’s age, which is important in our study to replicate differentcategories in the sample, for instance different values of number of mergers and acqui-sitions. The fourth column is used to show that only large companies were selected,because harmonization is not relevant for small organizations. Large companies wereselected from various industries and different age, based on the premise that companiesthat differ in terms of these four properties may also have differences in their contex-tual factors. In this case, with a variety in the sample we can obtain variety in thefactors to be able to easily observe contrasting patterns in the data [45]. Our researchfocused only on large organizations because the relevance of process harmonization ishigher in this type of organizations [66]. The last three columns indicate the type ofprocesses assessed in each company. It differs per company based on the availability ofthe information about some specific processes.

Table 4.1: Description of the sample.

Case IndustryCompany Company Processes evaluated

age size Standard Routine Nonroutine

C1 Automotivemanufacturer

84 Large X X X

C2 Locomotivemaintenance services

75 Large X X X

C3 Healthcare equipmentmanufacturer

18 Large X

C4 Logistic services 44 Large X X

C5 Cleaning equipmentmanufacturer

143 Large X X

C6 Healthcare services 193 Large X X


Within the wide range of processes that can be considered in a company, we decidedto select a management process. A management process defines capabilities for direct-ing, linking individuals and groups across the company and coordinating work flows todeliver their offering to customers [100]. This type of processes is specially importantin our case because higher levels of standardization can be achieved by strengthen-ing its capabilities [55]. Some examples of management processes in a company areplanning, purchasing and human resources management. The purchasing process wasselected to conduct our study for three reasons. First, purchasing is dictated by avail-able expertise and connections in practice, and our research team had the advantage ofhaving access to these connections. Second, the purchasing process has gained a criti-cal role in the global environment. For multinational companies a critical competenceis to source globally due to the challenges involved in the coordination of activitiesin multiple locations. Challenges arise from having multiple subsidiaries embedded indifferent cultural and regulatory systems [53]. In this scenario, the standardization ofpurchasing process can bring new technologies and capabilities to the value chain, andalso new cost drivers for the products.

Third, within the purchasing process we found all the different process types neces-sary to fill the theoretical categories that will be compared in this study. Lillrank [84]classified a process based on its level of structuredness, input and output variety, usingthree categories: standard, routine and non-routing. Considering our assumption thatthe role of contextual factor may differ for very standard processes compared to lessstandard ones, we decided to select three sub-processes within procurement (represent-ing each category defined by Lillrank [84]), to conduct our analysis. Using the pur-chasing process model defined by Weele [137], the three sub-processes selected withinpurchasing include: (1) ordering, to represent the standard process; (2) supplier’s se-lection, to represent the routine process; and (3) determining the specifications, torepresent the non-routine process. During data collection, the sub-process of deter-mining the specifications was discard because it can only be assesses by interviewingmultiple employees from operational areas located in different countries, and it wasunfeasible with the resources available for the project.

4.2.2 Definition of information sources

Two types of information sources are used: documents and interviews. They are partof the methods mostly used for data collection [75] and it increase the reliability ofdata by using multiple sources [43, 140]. The documentation used include businesspresentations and process descriptions. The benefit of this type of documentation isthat it provides precise details and can be reviewed repeatedly if necessary [56].

The interviews are the second source of information used. They have the bene-fit to specifically focus on the case study topic [56]. The interviews were conductedusing a pre-defined questionnaire, which is included in Appendix D. This is an in-strument prepared for use by an interviewer in a personal interaction which containsa predetermined set of questions to conduct a semi-structured interview. This typeof questionnaires provides uniform information and assures the comparability of data,which is required for this study with multiple cases.

In the first part of the questionnaire, we evaluate characteristics of the sample,including: company name, industry type, company size and company age. The com-pany size is measured based on the definition given by the European Commission’s

4.2. Methodology 65

recommendation 2003/361/EC [27]. They use three criteria to define the companysize: (1) the number of employees, (2) turnover and (3) the annual balance-sheet inEuros, evaluated in questions 3a, 3b and 3c. Using their recommendation we classifythe organizations in micro, small, medium and large.

The second part of the questionnaire (questions 5-14) assess the contextual fac-tors under study. Table 4.2 indicates which specific questions were used to evaluateeach factor, the type of questions used. Two types of questions are included in thequestionnaire: close-ended and combined. The close-ended questions were built usingcategories from the literature which are indicated in Table 4.2, and defining multiplechoices. Specific information is elicit via specific close-ended questions. The advantageof these type of questions in our case is that it facilitate the comparability of data andwith previous literature. The combined questions start with a close-ended questionand based on their choice we ask for additional explanation in an open-ended format.This is the case for level of process structuredness and product type.

The last column of Table 4.2 comprises the categorization of the constructs used inthe questionnaire.

The number of locations and number of mergers are evaluated through direct ques-tions. Also, IT governance and product type were evaluated through direct questioningusing a classification provided in the literature. The IT governance is classified incentralized and decentralized [12]. Boynton et al. [12] suggests to make a distinctionbetween who performs the function and who make the IT decisions. He suggests that“IT governance is not a concern with the location and distribution of IT resourcesthemselves, but rather with the location, distribution and pattern of managerial re-sponsibilities and control that ultimately affect how IT resources are applied and thenimplemented” [14]. The product type was classified in: divided or combined. Dividedare those cases in which the responsibilities for the execution of supplier selectionan ordering process is divided in different roles per process type (direct versus indirectproducts, products versus services). The product types defined in this study are clearlydistinguished in purchasing literature [137].

For the operationalization of the maturity level, we use the five levels of the Capa-bility Maturity Model (CMM) proposed by Humphrey [71]: (1) Initial, (2) Managed,(3) Standardized, (4) Predictable and (5) Innovating. The Capability Maturity Modelwas originally developed for software development processes. Afterwards, it was suc-cessfully adopted for analysing the state of a wide variety of processes, and even referas the standard framework for business process management by Davenport [37]. Threequestions were used in evaluate each level of the model, which are included in Ap-pendix D.

Organizational structure is a complex construct identified in the literature as multi-dimensional [79, 103]. The questionnaire presented by Pennings [103] is used to evaluatethe level of centralization and formalization of the organization. The number of itemsused to evaluate each one the different aspects of the index, are indicated in Table 4.2and the specific questions are included in Appendix D.

The level of process structuredness is assessed in the last part of the questionnaire(Questions 12-14). The evaluation is performed evaluating different aspects of theprocesses defined by Lillrank [84] to classify a process as standard, routine or non-routine. These questions are used to confirm our assumption that ordering can beclassified as a standard process and supplier’s selection as routine. Initially, the differentprocess variants were distinguished using an open-ended question, and later multiple


Table 4.2: Contextual factors and classification scheme.

Contextual factorsQuestionnumber

Questiontype

Categorization[reference]

Number of differentlocations

5 Close- Ordinal scaleended

IT governancecentralization

10 Multiple Centralized andchoice decentralized [14]

Product type11a , 11b Combined Divided and combined

(for product versus serviceand direct versus indirect)[137]

Maturity level9a-9e Multiple Initial, repeatable, defined,

choice managed or optimized [71]


6 Close- Ordinal scaleended

Organizational structure centralizationParticipation in decision making 7a Multiple 5 items Likert scale [59]

choiceHierarchy of authority 7b Multiple 4 items Likert scale [59]

choiceDepartmental participation 7c Multiple 5 items Likert scale [103]

choice

Organizational structure formalizationJob codification 8a Multiple 4 items Likert scale [59]

choiceRule observation 8b Multiple 4 items Likert scale [105]

choiceSpecificity of job 8c Multiple 4 items Likert scale [1]

choiceWritten communication 8d Multiple 5 item [103]

choice

Level of processstructuredness

14 Combined Non-routine, routine,standard [84]

4.2. Methodology 67

choices were provided to evaluate each process type considering the process variantsidentified. Finally, the level of harmonization of each aspect of the process (activities,data, resources and information systems) is evaluated (question 15) using the indicatorsdefined in Chapter 3.

A combination of open-ended and close-ended questions used for an interview areknown as a semi-structure interview. Semi-structured interviews have been used in thiscase study motivated by the fact that: (1) the type of inquiry is exploratory and theinterviews should allow unexpected information, and (2)some specific answers can beanticipated and categorize to facilitate comparability of data with previous research.While choosing the right type of interview is crucial for data collection, the selectionof the proper interviewees is also critical [33].

4.2.3 Preparation of cases

The preparation of cases consists of initiation of contact with the companies previouslyselected and the identification of candidate interviewees. The initiation of contactwith candidate companies were made via the professional network of the researchersinvolved in the project. The contact person were typically managers or directors inprocurement. An initial contact was made via e-mail, explaining the purpose of thestudy and expectations from the project in terms of required information and effort.The idea of this first approach was to raise the contact’s person interest in the study.In case of interest, a first appointment was arranged with the contact person to ensureconfidentiality and to address the resources needed for the following steps. During themeeting, we also described the type of knowledge required for the appointment of keyinformants. The outcome of this first meeting was an approval to start the project anda list of candidates with their respective contact data.

Considering the study’s research questions, the potential key informants have therole of managers, buyers or planners in procurement. Candidate interviewees wereselected based on their experience in the process. Information about contextual factorswas provided by top managers and directors, while specific information about theprocess variants was provided by mid-level managers directly responsible or involvedin the process.

The next step was to contact each one of the candidates by e-mail or telephone toschedule an interview with each one. Before the interviews, a selection of questionswere assigned for each candidate to ensure that the interview could take place in onehour and that rich information can be gathered from proper sources. Table 4.3 presentsa summary of the different roles selected for interviews described per company. Thefirst column refers to cases instead of companies’names due to confidentiality concerns.For each case, different roles can be described as key informants. The second columnpresents the roles of key informants with a general knowledge of the organization andknowledge of procurement. The next three columns are used to present the roles ofkey informants, which are people responsible for one of the process types specified:ordering, suppliers selection and contracting, and determining specifications.

For this case study design, an average of six interviews were conducted per case.The number of interviews performed depends on the number of different roles executingthe same process, i.e. if the supplier selection is performed by a direct procurementmanager and indirect procurement manager, then both were included for interviews.


Table 4.3: Selected cases and informants per process type.

Case #Informants per section:

General Ordering processSupplier’sselection

Determiningspecifications

C1 6

R1A (Directorprojects) andR1B (commodityunit manager)

R1C (commodityunit manager)and R1F(commoditymanager)

R1D, R1E(Commoditymanagers)

R1D(Commoditymanager)

C2 8R2A (strategicbuyer)

R2C(procurementmanager)

R2B (managerprocurement andlogistics), R2D(project leaderprocurement ),R2E (categorysourcing team),R2F (supplychain manager)

R2B (managerprocurement andlogistics) andR2D (projectleaderprocurement )

C3 5

R3A and R3D(trade-inmanager), R3E(projectmanager)

R3A (trade-inManager)

—R3B (logisticsmanager) andR3C (planner)

C4 4

R4A (land freightmanagers), R4Cand R4D (oceanfreightsupervisors)

R4A (landfreight manager)and R4C (oceanfreightsupervisor)

R4A (landfreight manager),R4B (oceanfreight manager)and R4D (oceanfreightsupervisor)

—

C5 2

R5A (EUpurchasingmanager) andR5B (materialsmanager)

R5B (materialsmanager)

R5A (EUpurchasingmanager)

—

C6 5

R6A(procurementmanager), R6B(senior buyer),R6D and R6E(applicationspecialists)

R6C (seniorbuyer)

R6B (seniorbuyer)

—

4.2. Methodology 69

4.2.4 Execution of interviews

The first information source used was the interviews, because they provide a generalpicture of the process and the company that help to even refine the original ques-tionnaire if necessary. A single researcher was in charge of conducting the interviews.Individual interviews were conducted instead of group interviews, because the type ofquestions included in the questionnaire does not require a group discussion and wealso wanted to avoid bias. The highly experienced informants used for the interviews,are people from different hierarchical levels who can provide different perspectives ofthe problem [45]. We gain more insights having different individual perspectives inthis study, especially with answers about perceptions. Individually interviewed peoplemight feel more comfortable to provide their perception of some topics.

The duration of each interview was restricted to one hour due to the availabilityof the resources. The process of conducting the interviews in the first case was timeconsuming and sometimes additional interviews were scheduled to complete the ques-tionnaire. Therefore, we decided to record the interviews in the other cases to increaseefficiency in the process. The audio recording is one of the methods most commonlyused in case study research [115].

The first interview was always conducted with a contact person with general knowl-edge of the processes under study, the organization and procurement department. Heprovided contextual information about the company and the process variants. Also thecontact data of candidate key informants was provided in the first meting. An e-mailwas sent to each respondent with a description of the research topic and the type ofgeneral information that would be gathered during the interview. In some cases thefull questionnaire was requested. Interviewees suggested additional documents that cansupport some specific information. They sent these documents after their interview.

4.2.5 Study of additional documentation

The process of data collection is iterative, not only during data collection itself but alsoafter data analysis, as shown in Figure 4.2. During data collection is possible that newideas or patterns emerge in the process that require adjustments in the process. In casestudy research there is an overlap between data collection and data analysis [140]. Thedata analysis in this study is divided in two parts: within-case and cross-case analysis.

The documents collected after the interviews included a wide variety of businessreports and process descriptions. The business reports include information such as: (1)a detailed description of the business, their products or services, (2) the organizationalstructure describing the different organizational units and their interrelation, (3) alist of legal or operational regulations specific for their industry, and (4) scorecardtemplates. The process descriptions were provided when available in one of two formats:(1) process models using BPMN notation, or (2) using a standard template to describeprocedures in the company. The type of documents collected differ among cases, anddepend on their availability in each case.

These documents were provided by the same key-informants after the interviewswere conducted. This type of documentation enriched the data obtained during theinterviews and provide a detailed description of specific quantitative data required [56].The additional documentation is also used to understand the underlying dynamics ofthe relations identified in the cross-case analysis. The following subsections present adetailed description of the process followed for data analysis in this multiple case study.


4.2.6 Within-case analysis

The objective of data analysis is to find relationships and contrasts between vari-ables [56]. The importance of within-case analysis is driven by the need to gain a richfamiliarity with each case to allow the emergence of unique patterns for preliminary the-ory generation [43]. The techniques used for within-case analysis in this study involvedcoding and tabulating all the quantitative data available, using the the classificationscheme provided in Table 4.2. The clear definition of constructs a priori assisted us inconverting data from different sources to measure the constructs in each case.

Once all individual data was tabulated, we looked for convergence of informationamong different respondents. When differences in quantitative values were identified,we asked again each one of the respondents to answer the same question in order toavoid a misunderstanding of the original question. If after the second attempt thedifferences still persisted, we decided to delete the data for considering it not reliablefor further analysis.

The criteria to define a significant difference in the values of respondents varyamong factors. There are two types of answers that we expect: perceptions and facts.The answers to : number of different locations, IT governance centralization, producttype, and number of mergers are factual. However, the evaluation of maturity level,organizational structure centralization and formalization are based on the respondents’perception. Differences in the responses are only expected for the answers based onperceptions. Therefore, the criteria to define a difference among responses is onlyspecific for these factors (maturity level and organizational structure). Differences oftwo or more levels identified in the maturity level are considered significant and neededconfirmation. For differences of one level, we use an average to estimate each score.

Additionally, the scales used to measure the indexes for the organizational structurecentralization and formalizations were taken from different literature sources and theyall use Likert scales. However, the number of levels of their scales differ among items.For instance, the participation in decision making is measured using a 5-point Likertscale, while the hierarchy of authority use a 4-point Likert scale. We decided to unifythe scales to get uniform statements that can be translated to general propositions.Therefore, for the analysis and categorization of these responses we are more interestedon the direction than in the strength of their opinion, considering that the strengthis more likely to differ among individuals [22, 83]. Therefore, in a 5-point Likertscale, from completely disagree to completely agree, both scores of 1 and 2 meandisagreement. The difference between them is just the strength of their disagreement.The same occurs for levels 4 or 5 which can be interpreted as different strengths ofagreement. As a result, we will categorize these results considering only two levels:‘low’ (for the two lowest values of the scale) and ‘high’ (for the two highest values).Scores in the middle point of the 5-point Likert scale are included in the category ‘high’,because previous research in the topic conclude that “a four point scale without a mid-point appears to push more respondents towards the positive end of the scale” [52, 145].In this case, differences in responses among interviewees are considered significant ifone is ‘high’ and the other is low. In this case additional confirmation is required.

Finally, the end result of the within-case analysis is a concise description of thecontextual factors and the level of harmonization of each individual case. In order toconduct the cross-case analysis, these individual factors are compiled across organiza-tions.

4.2. Methodology 71

4.2.7 Cross-case analysis

The cross-case analysis is performed after the within case analysis. After each singlecase was analyzed using a within case analysis, it was also included for the cross-casecomparison. In this way we could also identify if more cases were needed to have enoughvariety in the sample to make the comparisons. The technique used for cross-caseanalysis is pattern matching. We defined different categories for each contextual factor.Next, we grouped the cases per category, aggregated at a process level and separatedper process type (supplier’s selection and ordering). Finally, we look for patterns in thevalues of the indicators of process harmonization. We use pattern matching because itis a valuable technique to increase internal validity [140] by increasing the confidence ofmaking inferences using multiple data sources. To illustrate this procedure, a specificexample is shown in Table 4.4, describing the relations identified in the effect of writtencommunication in the level of process harmonization.

IndicatorsWritten communication

RelationsLow HighC5 C2 C6 C1 C3 C4

2.50 4.00 4.00 4.50 5.00 5.00

IA1 60.00 68.33 75.00 85.00 87.50 88.75 PositiveID1 11.00 7.00 2.50 4.50 4.00 8.00 PositiveID2 1.50 1.00 1.00 2.25 3.00 2.00 —IR2 4.00 4.00 6.00 4.25 3.50 3.50 —IT1 1.50 3.25 1.50 1.50 1.00 1.25 —

(a) Aggregated at process level

Indicators


RelationsLow HighC5 C2 C6 C1 C4

2.50 4.00 4.00 4.50 5.00

IA1 60.00 70.00 75.00 75.00 75.00 PositiveID1 3.00 4.33 4.00 10.00 6.00 NegativeID2 1.00 1.00 1.00 5.00 2.00 —IR2 4.00 3.00 4.00 6.00 3.00 —IT1 1.00 3.00 2.00 1.00 1.00 —

(b) Supplier’s selection process



2.50 4.00 4.00 4.50 5.00 5.00


(c) Ordering process

Table 4.4: Relations identified with pattern matching using the written communication as a criteriafor clustering of cases.


Written communication is one of the aspects of organizational structure formaliza-tion that are evaluated in this study. Data is presented aggregated at a process level(Table 4.4) and divided per process type: supplier’s selection process (Table 4.4b) andordering process (see Table 4.4c). The level of written communication is measuredusing a 5-point Likert scale and for analysis purposes it is categorized in two levels:‘low’ (<3) and ‘high’ (>=3), as described in subsection 4.2.6. For the analysis usingthe aggregated values at a process level, one case (C5) is classified as ‘low’, with anaverage level of written communication of 2.50. The other cases are classified as ‘high’considering that their average values of written communication are higher than 3 (C1,C2, C3, C4 and C6). Next, we compare the values of the indicators of process har-monization (IA1, ID1, ID2, IR2 and IT1) between the two clusters to identify specificpatterns. In this case we found two relations between written communication and:(1) the level of harmonization of activities (IA1) and (2) the level of harmonization ofdata (ID1). For instance, if all the values of IA1 of the companies in the cluster ‘low’(60.00) are lower than the values in the cluster ‘high’ (70.00 and 75.00), then we canconclude that there is a pattern in the data and the type of relationship is positive (fora low level of written communication there is a low level of harmonization of activities.The rules used to define the type of relations between factors and indicators of processharmonization are summarized in Table E.1.

The cross-case analysis of the level of structuredness is performed differently thanthe other factors, considering that for each case we have two different values for eachindicator (one for the ordering process and another for the supplier’s selection process).Therefore, we need to evaluate the level of their differences and compare them amongcases. The level of their differences is calculated using the percentage of the differencebetween the value of each indicator for the ordering process and the value for thesupplier’s selection process. With this measure we want to find the strength anddirection of the differences. For instance, if the percentage of the differences in theindicator IA1 is positive for all the cases in the sample, we can conclude that therelation is positive. The reason is that the ordering process is a standard process whichrepresents a higher level of process structurednes than the supplier’s selection processthat is a non-routine one. In this case the relation is positive because the processwith a higher level of structuredness always has a higher level of standardization ofthe activities. Additionally, we evaluate whether the relations identified using theaggregated values hold for both process types as well.

It can be observed in Table 4.4b that case C3 is not included in the analysis for thesupplier’s selection process because this process type was not assessed in this specificcase as shown in Table 4.1. Appendix E presents all the relations identified with thecross-case analysis, organized per factor. Results were aggregated at a process leveland separated by process type (ordering and supplier’s selection). The next step isto compare the findings. We review the literature to sharpen the insights identified,and finish describing a set of relevant research propositions [43]. Propositions are onlyderived when relations were identified and showed theoretical support. Otherwise, wedo not derive propositions to indicate absent of relations, because it may be the casethat they are not found due to our limited sample size.

4.3. Results 73

4.3 Results

As a result of the within and cross-case analysis described in the previous section,all the relations identified are summarized in Table 4.5. The relations are shown in amatrix, in which the contextual factors are included in the vertical axis and the aspectsof process harmonization are included in the horizontal axis. If a relation between afactor and one of the aspects was found, the type of relation (positive or negative) isindicated in the intersection.

Table 4.5: Relations identified using aggregated data at a process level.

Factors Activities Data ResourcesInformationTechnology

Number of different locationsIT governance centralization NegativeProduct type Negative Negative NegativeMaturity level Negative PositiveNumber of mergers and acquisitions NegativeOrganizational structure centralization- Participation in decision making Positive- Hierarchy of authority- Departmental participationOrganizational structure formalization- Job codification Positive- Rule observation- Specificity of job Positive- Written communication Positive PositiveLevel of process structuredness

In this section we present a detailed analysis of the relations found, using an inte-grated approach [56]. In this integrated approach, we explain the relations organizedaround the contextual factors, using the underlying data supplied by each case study.This approach is very efficient from the reader’s side who can go directly to thosefactors of interest. First, we provide an explanation of the type of relations identi-fied. Second, we compare our findings with previous literature to generate theory withstronger internal validity [43]. Third, we derive theoretical propositions that can befurther used to propose a modified model for explaining the determinants and effect ofprocess harmonization.

4.3.1 Effect of different locations

No relations were found even though our data indicate a diverse sample in terms ofthe number of locations (See Table 4.6). This contradicts what we expect from theory,in which an increasing number of locations is associated with a lower level of processharmonization in general [134]. It would be expected that from the information systemsperspective, more locations create a more diverse and complex systems landscape. Thisshould be reflected in the IT aspect of process harmonization.

However, it seems that the relation between number of locations and process har-monization is not a clear-cut issue. Tregear [134] discussed that the significant effectof this factor may be caused by differences in local market imperatives such as na-


tional or regional culture, local market conditions, customer expectations, legislativerequirements and more. Tregear [134] argues that in multinational companies differentvariations of the processes may arise because the interaction between individuals per-forming similar tasks is reduced. However, we consider that this effect is reduced forlarge organizations with a high levels of maturity and formal procedures such as thecompanies in our sample. In our case, personal differences can exert a lower impactin the harmonization of processes which are less dependent on a individual way ofworking.

Moreover, the way in which this factor was measured may exert an influence in theresults obtained. First, we consider in this study the number of locations in which theprocurement department is established and not the number of operational facilities ofthe company. The reason for that is that the unit of analysis is an internal process inprocurement and we are interested in the number of locations in which these processesare executed. Both ordering and supplier’s selection were evaluated from the perspec-tive of the procurement department only. Second, only in two cases (C3 and C5),the processes were assessed considering international locations. If indeed the effect ofnumber of locations is more associated with differences in countries or regions, thenadditional data is required to assess the relations between process harmonization andthis factor.

Table 4.6: Contextual factors of companies studied

CaseNumber oflocations

ITgovernance

Product type Maturity levelNumber ofmergers

C1 3 centralized Combined 5 2C2 3 centralized Combined 4 3C3 2 centralized Divided 5 1C4 50 decentralized Divided 4 2C5 7 centralized Combined 5 1C6 1 centralized Combined 3 1

4.3.2 Effect of IT governance centralization

A negative relation was identified between the centralization of IT and the level ofharmonization of activities. This means that the level of harmonization of activities islower when the IT governance is centralized. However, we do not have a theoreticalexplanation for this finding and by analyzing the data we suspect that this is theeffect of an spurious relation. This negative relation is driven by the effect on thesupplier’s selection process only and it is the effect of the product type (product versusservices) and not of the ‘IT governance. We have two reasons to believe to supportthis statement. First, when comparing Table E.3 and Table E.5, we observe thatthe clusters defined for the analysis of IT governance centralization and product type(products versus services) are the same for the supplier’s selection process, but notfor the ordering process. Analyzing the data for the supplier’s selection process, wecan see that the values for a decentralized IT governance (C4) is the highest valuewithin the sample (75). We cannot define a negative relation because there are othercases in the sample with the same value with a centralized IT governance (C1 and C6).However, in the ordering process it is clear that this relation doesn’t hold. Second,

4.3. Results 75

a theoretical support can be provided to explain these relations for product type, asshown in the next subsection. Therefore, we conclude that this outcome is not theeffect of IT governance centralization but of the product type.

Based on the previous analysis, we can conclude that our findings do not yieldsignificant relations, which is unexpected for three reasons. First, previous researchshowed evidence of positive relations between IT governance centralization and thelevel of harmonization of business processes [10, 55, 112]. We dealt with the classiccentralized/ decentralized dichotomous classification of IT governance structure. Fromthis classic research stream, most authors agreed that a centralized form providesmore control over standards, while a decentralized form increases the customization ofsolutions for each business unit [4, 14, 139]. Similarly, Ross et al. [112] considered theIT governance centralization as one of the characteristics of the four business operatingmodels. It shows that the IT decisions are made centrally for models with a high level ofprocess standardization, while IT decisions are made within business units for modelswith a low level of standardization.

Second, our sample size and mix is sufficient to relate patterns using only two ITgovernance forms (centralized and decentralized). We based this statement in the factthat Brown and Magill [16] also used six companies to explore relations between tenantecedents and four IT governance forms. The interacting antecedents evaluated in-cluded: corporate vision, corporate strategy, overall firm structure, culture-businessunit autonomy, strategy IT role, senior management of IT, satisfaction with manage-ment and technology, satisfaction with use of technology, strategic grid of current/future applications and locus of control for systems approval/priority. As a result,they gathered patterns to define profiles to provide a predictor model for IT gover-nance structures [14].

Third, we evaluated the IT governance structure at the organizational unit levelinstead of the corporate level. Brown [15] presents evidence that factors at the busi-ness unit level might provide stronger explanations of the governance arrangementswithin individual units. Therefore, we are satisfied that our focus at the unit level isappropriate for this study.

However, it is possible that in practice there are other factors not considered in ouranalysis that interact with the centralization of IT governance and play a role in ourexpected relations. Also, in the initial model (See Figure 4.1) we do not consider rela-tions between the different contextual factors included in the model. This assumptionof non-interacting factors may be unrealistic with respect to the IT governance. Thisis supported by the findings of Sambamurthy and Zmud [118], which evaluated theformation of the IT governance modes at the enterprise level, using eight cases. Theirstudy found that strong support for ‘multiple contingencies’ theory. This theory arguesthat “contingency forces interact with each other by either amplifying, dampening, oroverriding their mutual influences on the IT governance mode” [118].

4.3.3 Effect of product type

The distinction between roles working with direct (production related) and indirect(non-production related) products is constant in all the cases in our sample as shownin Table E.4. This is expected considering that the differences between direct andindirect spent is well known at the procurement community. The distinction emergedas part of an attempt to bring more attention to the products that has a more direct


influence in the final product spent and also have a greater role to support the businessstrategies. In this sense, the benefits of integrating these proces types does not seem tooffer significant advantages compared to the disadvantages, such as less specializationin the sourcing of critical commodities due to invest more time in indirects.

However, this is not the case when it comes to products versus services. In oursample, we found two cases (C3 and C4) in which a distinction is made in the rolesmanaging products and services. In both cases, the services are part of their businessoffering and therefore they have a strategic role in their operations. For instance, caseC3 is a manufacturer with a special process in which they refurbish medical systems.As part of their offering, they need to provide the services related to de-installation andinstallation of these specialized systems, and even the construction or adaptation of thearea where their systems will operate. These services are then part of their businessoffering. The same applies to case C4, which is a logistic service provider. Theirprocesses in procurement (of case C3 and C4) are directly linked with the operationalpart for the offering of their services to the external customer. In this case, it iscritical for them to distinguish products and services, because it implies that theyassign different priorities to each one of these two categories.

As a result of the cross-case analysis we found three relations between process type(only referred to services versus products) and the level of process harmonization. Allof them negative with respect to the harmonization of activities, resources and IT. Thenegative relations found can be explained in the situation in which the services arepart of the business offerings for all the companies that divide the roles for productsand services. For the referred companies (C3 and C4), the reasons to keep them apartare the same that holds for the distinction between directs and indirects, consideringthat they need a different focus in order to support their business strategy. Therefore,the advantages of their integration are not significant compared to the advantages ofproviding different priorities to support their business offerings. These arguments leadto the following propositions:

Proposition PROD1: The harmonization of activities is high when different rolesare assigned to manage products with different strategic focus in the organization.

Proposition PROD2: The harmonization of resources is high when different rolesare assigned to manage products with different strategic focus in the organization.

Proposition PROD3: The harmonization of IT is high when different roles areassigned to manage products with different strategic focus in the organization.

4.3.4 Effect of maturity level

While past research provides evidence of positive relationships among the level of ma-turity of an organization and the level of standardization [111], such relations may nothold true for all the different aspects of process harmonization. The cases included inour study showed medium and high levels of maturity and two different types of rela-tions were identified: (1) a negative relation with respect to data, and (2) a positiverelation with respect to resources.

A negative relation with respect to data means that in our sample we observe thatcompanies with a higher level of maturity have lower levels of harmonization of data.Even though this relation was only confirmed in the ordering process, Table E.6b showsthat for the supplier’s selection, the same negative relation is observed expect for onecase (C5). This finding contradicts what is theoretically expected, considering that

4.3. Results 77

higher maturity implies also more control and more standard procedures. Therefore,there is no satisfactory explanation for this finding, and no proposition can be is derived.

The positive relation with resources indicates that for higher levels of maturitywe observed higher levels of harmonization of resources. Since the process is morecontrolled and standard, the company relies less on the individual efforts such as in theinitial levels of maturity. Therefore, different roles can manage the different processvariants without affecting the general performance. No significant differences werefound among the scores provided by the different interviewees. Based on our findings,we derive the following proposition:

Proposition MAT1: For organizations with a high level of process maturity, theharmonization of resources is high.

4.3.5 Effect of mergers and acquisitions

Our findings reveal that companies with a higher number of mergers and acquisitionsshow a lower level of harmonization with respect to IT. This finding is in line with thearguments presented in section 1.1 in which we indicated that mergers and acquisitionsare one of the business strategies that impose new challenges in terms of harmonization.Mergers and acquisitions would increase the level of complexity in cases were differentprocesses and information systems exist, and interfaces become a necessity to fullymanage the complicated system landscape [117]. This is known in practice as thelegacy problem. After merged, a trade-off exists in which companies need to balance theneed for information exchange within their value chain and the investments necessaryto unify systems among companies and business units [112]. Based on the presenteddiscussion and findings we derive the following research proposition:

Proposition MER1: Organizations with a high number of mergers and acquisitionsare less harmonized in terms of their IT systems.

4.3.6 Effect of organizational structure centralization

Three items were used to evaluate the level of centralization of the organizational struc-ture: (1) hierarchy of authority, (2) departmental participation and (3) participationin decision making. Only high levels of organizational structure centralization wereobserved among all companies in the hierarchy of authority and departmental partici-pation, as shown in Table 4.7. Therefore, these two items were not assessed for a lackof variety in our sample that does not allow to observe any patterns in the data.

The next step was to conduct a within-case and cross-case analysis for the onlyitem that presents enough variety in the sample, the participation in decision making.As a result of the within-case analysis, no significant differences were found in theevaluations provided by the interviewees, except for case C1 in which differences wereidentified. We argue that the differences found in C1 are due to differences in theorganizational level of the interviewees. They explain that in the company the answerto these questions could significantly differ based on the level under analysis. Forinstance, when we asked about the participation in decision making related to aninterviewees’ organizational level or lower, he/she always participate in the decisionmaking, while for a higher level the participation is null. This could influence the finalevaluation because respondents with higher hierarchical levels assigned higher scoresto this item than respondents with lower level. In the literature we also find that this


Table 4.7: Organizational structure assessment of companies studied

Factor C1 C2 C3 C4 C5 C6

Organizational structure centralization:Participation in decision making High Low High High High HighHierarchy of authority High High High High High HighDepartmental participation High High High High High HighOrganizational structure formalization:Job codification High High Low Low Low HighRule observation High High Low Low Low LowSpecificity of job High Low High High Low HighWritten communication High High High High Low High

aspect must be taken into consideration when we evaluate questions which are aboutperceptions, in which even personal differences can bias the estimates of perceptualagreements [72]. In this case, we intentionally include people from different levels tohave better representativeness in the company sample.

As a result of the cross-case analysis for the participation in decision making, wefound a positive relation between the organizational structure centralization and thelevel of harmonization of IT. This pattern was found not only for the aggregated dataat a process level, but also for each one of the two process types. This proposes that forhigher levels of centralization of the organizational structure we can find higher levelsof harmonization of IT. This relation is expected considering that when decisions arecentralized, for instance by headquarters, the investments in IT systems are more uni-fied and oriented to provide global solutions. This centralized view of the organizationas a system allow the managers to find global IT solutions to support their manage-ment processes instead of trying to optimize one organizational unit, which is the casewhen the organizational structure is decentralized [55].

Additionally, we want to explain the relations in terms of the participation in deci-sion making and not only in terms of the general construct of organizational structurecentralization. The participation in decision making represents the extent to whichpeople in an organization participate in decisions about the allocation of resourcesand the determination of organizational policies [59]. It is also expected that whenmore people in the organization can participate in decisions about the allocation ofresources such as money for an investment in a new IT system, the managers canmake better and more informed decisions that guides to a more global optimization oftheir resources, instead of a local view. Based on this findings and our arguments, thefollowing research propositions is derived:

Proposition ORGC1: The harmonization of IT is high in organizations with a cen-tralized organizational structure.

Proposition ORGC2: For high levels of participation in decision making, the har-monization of IT is high.

4.3. Results 79

4.3.7 Effect of organizational structure formalization

The organizational structure formalization is the factor that shows more relations withthe level of harmonization of the process harmonization, as summarized in Table 4.5.All of the relations found with the analysis using the aggregated values at a processlevel were positive as expected from the literature in contingency theory.

We found that the level of formalization has an influence on all of the differentaspects of process harmonization included in this study (activities, data, resources andIT). The type of effect appears more significant for a standard process such as orderingcompared to the supplier’s selection process, if we consider the number of relations perprocess type as depicted in Table 4.8. It is intuitive to expect that the formalization ismore relevant in more standard processes, in which most of the work can be codifiedusing specific rules that give guidelines for the people performing the job. In lessstandard processes this formalization is not always feasible because the options in theprocesses are not always expected, requiring the development of new knowledge in theproces.

In this study we investigated the organizational structure formalization using fourindividual items: job codification, rule observation, specificity of job and written com-munication. Next we present a brief description of each item together with argumentsto explain the relations identified per item.

The job codification represents “how many rules define what the occupants of posi-tions are to do” [59]. One positive relation was identified between the job codificationand the harmonization of resources. This relation can be explained considering thatmeasures the degree of work standardization from the resources perspective, in whicheveryone has a specific set of tasks to perform and they are clearly defined per role.

No relations were found for rule observation when the data was aggregated at aprocess level. The rule observation “is a measure of whether rules defined are employedand enforced” [59].

The specificity of job refers to “the degree to which procedures defining a job arespelled out” [1]. A positive relation between specificity of job and the level of har-monization of activities was found. The argument provided to explain the relationbetween job codification and the harmonization of resources, also holds in this case.The job specificity shows the degree of work standardization from the process perspec-tive, instead of resources. In this case, procedures are defined to explain the detailsof the process and not view from the resources responsible to perform the process.Therefore, it is expected that for a more detailed and specific description of the tasks,the variation in the activities and control-flow is reduced. This positive relation withrespect to activities holds for both process types (See Table 4.8). Additionally, a posi-tive relation was found with respect to the harmonization of data, that only holds formore standard processes.

The last item of the organizational structure formalization explored in this study isthe written communication. Written communication refers to “the frequency of writtencommunication as reported by individuals” [103]. Our findings revealed a positive rela-tion between this factor and the harmonization of activities and data. These relationscan be explained if we consider that a high level of written communications also impliesa high level of information exchange among individuals in the organization. With ahigh level of information exchange, the definition of rules to approach unexpected situ-ations in a process become more unified. This leads to a higher level of harmonization


of activities, control-flow and data.Summarizing our previous findings and arguments we derived the following propo-

sitions:Proposition ORGF1 - The level of process harmonization is high in organizations

with a high level of organizational structure formalization.Proposition ORGF2 - The harmonization of resources is high in organizations with

a high level of job codification.Proposition ORGF3 - The harmonization of activities is high in organizations with

a high level of job specificity.Proposition ORGF4 - The harmonization of activities is high when the level of

written communication is high.Proposition ORGF5 - The harmonization of data is high when the level of written

communication is high.

4.3.8 Effect of level of process structuredness

The results of the cross-case analysis for the level of process structuredeness of theprocesses evaluated is summarized in Table E.15 in Appendix E. Table E.15a showsthe differences in percentage between the values of the indicators for the orderingand supplier’s selection process. No specific pattern can be observed in terms of thedirection of the relations, considering that for all the indicators we found positive andnegative values. However, for the majority of the cases the differences were positiveindicating that the level of harmonization of any of the aspects evaluated (activities,data, resources and IT) is consistently higher for ordering than the supplier’s selectionprocess as expected.

Next, we analysed the strength of their differences using an average of the absolutevalues of their differences. The average values are shown in Table E.15a. Differ-ences were found in all the cases with respect to Data and IT, being ‘data’ the aspectthat higher differences is ‘data’. This finding is supported by the literature whichindicates that the harmonization of data is one of the biggest challenges in terms ofintegration[112]. The aspect of ‘activities’ appears to be the less significant in termsof the strength of their differences. This finding is unexpected considering that thelevel of structuredness by definition is more oriented to the control-flow aspect of theprocess. However, it is possible that in order to observe significant differences in thesevalues, more extreme cases are needed such as the comparison between a standard anda non-routine process.

The last part of the analysis consisted on evaluating whether the types of relationsidentified in the cross-case analysis performed per contextual factor, hold for bothprocess types or are mainly driven by one of the processes. A summary of theserelations are depicted in Table 4.8. The type of relation are identified (negative orpositive) and the aspects with respect these relations hold (indicated in a parenthesis).Table 4.8 shows that the majority of the relations that were found in the analysisusing the aggregated data hold also for the ordering process. In one case, oppositedirections (one negative and one positive) were found for the different process types, inthe relations between the level of written communication and the level of harmonizationof data. In this case, the relation that holds for the ordering process was stronger inthe direction, considering that it holds also when using the aggregated data.

These findings suggest that the effect of some specific contextual factors differ per

4.3. Results 81

Table 4.8: Summary of the relations identified aggregated at a process level and per process type(ordering and supplier’s selection).

FactorAggregated atprocess level

Supplier’sselection

Orderingprocess

Number of different locations — — —

IT governance centralization Neg.(Act.) — —

Product type Neg.(Act.) — —Neg.(Res.)Neg.(IT)

Maturity level Neg.(Data) — Neg.(Data)Pos.(Res.) Pos.(Res.)

Number of mergers and acquisitions Neg.(IT) Neg.(IT) Neg.(IT)

Organizational structure centralization- Participation in decision making Pos.(IT) Pos.(IT) Pos.(IT)

- Hierarchy of authority — — —

- Departmental participation — — —

Organizational structure formalization- Job codification Pos.(Res.) — Pos.(Res.)

Neg.(Data)

- Rule observation — — —

- Specificity of job Pos.(Act.) Pos.(Act.) Pos.(Act.)Pos.(Data)

- Written communication Pos.(Act.) Pos.(Act.) Pos.(Act.)Pos.(Data) Neg.(Data) Pos.(Data)


process type and the type of effect is different with respect to different aspects ofprocess harmonization. We can conclude that the level of process structuredness hasa more significant influence in the harmonization of data than in the other aspects.Additionally, we can conclude that also in the majority of the cases the relation betweenthis factor and the different aspects of process harmonization is positive. Based on thisfindings we derive the following propositions.

Proposition STRUCT1: the effect of contextual factors in the level of harmonizationof more standard processes is more significant than for routine or non-routine processes.

4.4 Validity and reliability

This section address the strategies applied to enhance validity and reliability of thestudy and its results. Our study places special emphasis on construct validity, internalvalidity, external validity, and reliability, which are considered the most commonly usedcriteria for judging the quality of research designs in case study research [147].

Construct validity refers to “the extent to which we establish correct operationalmeasures for the concepts being studied” [140]. The tactics applied in this study toguard construct validity during the data collection include the use of multiple sources ofevidence and establishing a chain of evidence. The sources of evidence were individualsfrom different organizational levels and functions from all the companies. Additionally,the structured questionnaires used for data collection were based on previously usedoperationalization of the key constructs.

Actions to increase internal validity are relevant in our research, considering that wewant to derive relations between different variables. The cross case analysis conductedin this research using pattern matching seek to increase the internal validity of ourfindings. Pattern matching is one of the tactics proposed in Yin [147] that seek forconfirmation from multiple data sources and the reduction of peoples’s tendency toconclude based on a limited data [43]. Pattern matching is suggested to increase theconfidence of making inferences using the data available. Additionally, we contrastedthe results with the relations expected based on previous literature.

Table 4.9 corroborate which relations identified with pattern matching are sup-ported by previous literature. It is important to emphasize that we cannot contrastwith relations with respect to each specific aspect of process harmonization but withthe expected relations with the level of process harmonization in general. That isthe reason why in our comparison the literature refers only to general relations. Ad-ditionally, we presented the propositions derived to a group of three professionals inprocurement to evaluate wether they could be intuitively explained and sound withrespect to what they expect in practice. they corroborate that the propositions weresound.

The generalization of the case study so that it can contribute to theory is im-portant [114]. The external validity is concerned with the extent to which a study’sfindings can be generalized beyond the study itself [124]. The claim that our findingscan be generalized is based on the replication logic. We use multiple cases, six casesand two different process types per case (supplier’s selection and ordering), to supportthe propositions derived. Each case can be viewed as a different experiment, in which“the more cases that can be marshalled to establish or refute a theory, the more robustare the research outcomes”. Therefore, replication can be claimed in this case [114].

4.4. Validity and reliability 83

Table 4.9: Assessment of internal validity to corroborate the relations found with previous literature.

FactorRelationsfound

Expectedrelations Finding

Number of different locations — Neg. [134] —

IT governance centralization Neg.(Act.) Pos. [10, 55, 112] Not supported

Product type Neg.(Act.) — —Neg.(Res.)Neg.(IT)

Maturity level Neg.(Data) Pos. [111] Partially supportedPos.(Res.)

Number of mergers and acquisitions Neg.(IT) Neg. [134] Supported

Organizational structure centralization- Participation in decision making Pos.(IT) Pos.[55, 126] Supported

- Hierarchy of authority — — —

- Departmental participation — — —

Organizational structure formalization- Job codification Pos.(Res.) Positive[55, 126] Supported

- Rule observation Positive[55, 126] —

- Specificity of job Pos.(Act.) Positive[55, 126] Supported

- Written communication Pos.(Act.) Positive[55, 126] Supported

Level of process structuredness Pos.(Act.) Positive[55, 126] Partially supportedNeg.(Data)


Reliability is the extent to which a study’s procedure can be repeated with thesame results [147]. We kept a database to establish a chain of evidence during the datacollection process. This database contains a detailed description of all the cases.

4.5 Conclusions

The goal of this chapter is to examine the relationship between contextual factorsand the level of harmonization of different aspects of process harmonization. Theinitial model specifying the contextual factors and the aspects of process harmonizationincluded in the analysis are depicted in Figure 4.1. These relations have been tested inprevious research [8, 97, 99, 119, 146] only between the contextual factors and the levelof harmonization in general, but not with respect to specific aspects of harmonization.In this study we empirically tested these relations using a multiple case study approachwith six cases.

Overall, our findings support previous arguments of contingency theory, consideringthat we found relations between most of the factors evaluated. We suggest that therole of the contextual factors differ per process type, being more significant for morestandard processes than less standard ones. Additionally, we found that the level offormalization of the organizational structure has a significant role when harmonizationinitiatives are conducted in an organization, which has an influence in all the differentaspects of process harmonization. The indicators that show more patterns in the datawere the indicators of activities and IT. A summary of the different types of relations(positive and negatives) is depicted in Table 4.9 with a corresponding validity checkprevious literature.

Twelve different propositions were derived as a result of the relations identifiedin this empirical study. Relations were found for all the factor analyzed, excludingthe number of locations. Additionally, the relations of two items of the organizationalstructure centralization (hierarchy of authority and departmental participation) werenot explore for lack of variety in the sample.

These findings build in the contingency theory and the research stream on processstandardization. Identifying the organizational structure formalization as the mostsignificant factor in the relations identified, we confirm the findings in contingencytheory which suggests the way in which organizations operate must fit to the degree ofuncertainty in their environment [40, 92]. Additionally, this study contributes to theexiting theory in process harmonization by shedding light on the specific aspects of aprocess that can be influenced by contextual factors.

Our results provided important managerial guidelines for conducting harmonizationinitiatives. One managerial implication is that it is not always optimal to impose onestandard process, considering that there is a trade-off involved in this decision. Ourfindings offer a good basis for business managers to better understand which contextualfactors are relevant when conducting harmonization projects and the specific aspects ofprocess harmonization that they influence. Also, the way in which managers structuretheir organization is critical to guarantee the expected gains of process harmonization.When new investments in IT are the ones leading harmonization initiatives, then itis recommended to centralize the decision making. The recommended centralizationin this case provide a global overview of the organization to avoid searching for localsolutions and investments in IT systems have to be included in the analysis of thetrade-off. However, when a harmonization initiative is mainly process driven, then

4.5. Conclusions 85

the most significant aspect to consider is the organizational structure formalization.In this case, the investments should focus on defining clear procedures in which theactivities to be performed in each process are clearly specified and codified. In this case,investments in resources to build and maintain such procedures have to be consideredin the evaluation of the trade-off.

Some caution must be exercised when generalizing our findings because the samplesize was small, particulary when subdivided into various categories. We also evaluatedonly managerial processes in large organizations. With that caveat stated, we foundsome results that could be explored deeper in future research.

Chapter 5

Conclusions

“I do not feel obliged to believe that the same God who has endowed us with sense,reason, and intellect has intended us to forgo their use”

Galileo Galilei

Business process harmonization has become a necessity for globally operating com-panies, considering that it facilitates the coordination of activities and informationexchange that takes place in their value chain. However, observations from practice re-vealed that the expected benefits of process harmonization are hardly exploited. Theseobservations suggest that there is a trade-off faced by companies, which consists ofmaking a choice between the extent to which a process should be harmonized and theinvestment in resources necessaries to achieve these harmonization level. Academicresearch has not yet been able to provide an adequate answer to the factors that in-fluence this trade-off. Therefore, knowledge of these influencing factors needed furtherinvestigation.

Using insights from the contingency theory, we developed an overall research frame-work which proposes that the level of harmonization that can be achieved is dependenton contextual factors of the organization and its environment. Against this backgroundour leading research question is: What contextual factors should be considered to decideon the right level of harmonization of a business process in an organization?

We have divided this research question in four sub-questions to extend our under-standing of the contextual factors and their effect on the trade-off previously described.In this final chapter, we summarize the results of the four sub-questions previously men-tioned and we use the collective results to answer the overall research question. Wereflect on the theoretical and practical implications of this research. We conclude thischapter reflecting on the limitations of this dissertation and suggesting avenues forfuture research.

5.1 Main conclusions

An overview of the most important findings of this research is provided in this section.They are organized around the four research questions that lead to an answer to ouroverall research question.

88 Chapter 5. Conclusions

RQ1-What contextual factors determine the level of harmonization of abusiness process?

Eleven contextual factors were identified in previous studies by means of a litera-ture review described in Chapter 2. They were summarized in a conceptual model fordetermining the effect of contextual factors on the level of process harmonization andsubsequently on business performance. The contextual factors identified were classi-fied in three categories (immediate, internal and external) based on the level of theorganization that they characterized.

Immediate factors are contextual factors that describe a process that is being har-monized. The immediate factors identified were: the level of process structurednessand personal differences in terms of experience and knowledge of the the resources.Internal factors are those that describe the internal environment of an organization.Six internal contextual factors identified were: the number of locations, level of cen-tralization of IT governance, differences in product type, maturity level, organizationalstructure and number of mergers and acquisitions. External factor are those beyondthe control of an individual organization. We identified three that are relevant toour study: cultural differences, different regulations and power distance with otherorganizations in the business network in which the organization operates.

RQ2. Which aspects of a business process are relevant to measure thelevel of process harmonization?

Six relevant aspects [5] were identified through a literature review described inChapter 2. We looked for measures of process harmonization to identify which aspectsof the process they assessed. The aspects found include: activities, Control-flow, data,resources, Management and information technology. Activities is a behavioral aspect ofa business process which describes the set of activities and specific order in which theyare executed in a process. Control-flow was later combined with activities consideringthat they both refer to a behavioral aspect of a process. Data is an informationalaspect which describes the information that is involved in a business process, howit is represented, and how it is propagated among different activities. Resources isan organizational aspect which describes in which way resources are involved in thebusiness process. Management refers to an organizational aspect concerned to the rulesto control process performance. The last aspect, information technology, is treated asan organizational aspect in this research instead of an informational one. The reasonis because it refers to the way in which IT systems are involved in the process and noton the specific information exchanged through those systems.

These aspects were used as an input in the operationalization of process harmo-nization that is performed in Chapter 3 to answer the next research question.

RQ3. What measures can be used to quantify the level of harmoniza-tion of a business process in an organization?

Seven measures were empirically tested and validated using survey data collectedfrom 119 professionals in business process management. They were developed to assessall the relevant aspects of process harmonization previously defined: activities, data,resources and information technology. For activities, the measure proposed is ‘thepercentage common activities in different variants of a process (IA1)’. Data is evalu-ated through two measures: (1)‘the number of different input data (ID1)’ and (2)‘thenumber of different output data (ID2)’. The aspect of resources is evaluated with two

5.1. Main conclusions 89

measures: (1)‘the percentage of common roles in a process (IR1)’ and (2)‘the numberof different roles executing the same activity (IR2)’. And the aspect of informationtechnology is measured using ‘the number of different software applications in use inthe process (IT1)’. These measures were used to build a hierarchical measurementmodel described in Chapter 3.

RQ4. How do contextual factors influence the level of harmonizationof each individual aspect of process harmonization?

We identified significant relations in six out of eight contextual factors empiricallytested through multiple-case studies with six cases. The contextual factors examinedwere: (1) product type, (2) maturity level, (3) number of mergers and acquisition, (4)organizational structure centralization, (5) organizational structure formalization, (6)level of process structuredness, (7) number of different locations and (8) IT governancecentralization. Table 5.1 presents an overview of the propositions which explain all therelations identified. Next, we provide a discussion about the relations identified betweenthese eight contextual factors and the four relevant aspects of process harmonizationpreviously defined.

The first factor refers to product type. We found that when companies assign dif-ferent roles to manage products with different strategic focus in the organization, thelevel of harmonization of activities, resources and IT is higher. These relations can beexplained because the advantages of integration of different products (products versusservices) are not significant compared to the advantages of providing different prioritiesto support their business offerings. This is relevant for companies in which services arepart of their business offerings and demand more focus.

We found that organizations with higher levels of maturity have higher levels ofharmonization of resources. However, it was surprising not to find relations with respectto the harmonization of activities, considering that by definition of maturity processesare more controlled and standard.

In line with theory, we found that organizations with a higher number of mergersand acquisitions are less harmonized in terms of their information technology aspect.It was expected that more mergers and acquisitions increase the level of complexity inthe systems landscape [117] due to the presence of different systems to support similarprocesses. This contextual factor is relevant in the context of process harmonization,considering how companies depend nowadays on their systems to manage and controllarge amount of data and the need to unify systems among companies and organiza-tional units to facilitate their information exchange.

Two dimensions of organizational structure were differentiated in our study to pro-vide better insights into the underlying mechanisms of this factor that play a rolein process harmonization. Both, centralization and formalization, are the most high-lighted dimensions of organizational structure in the literature [59, 103, 105]. In thisstudy, we considered the effect of each one of these two dimensions independently.

Organizational structure centralization was represented using three items: personalparticipation in decision making, hierarchy of authority, and departmental participationin decision making. We found that organizations with higher levels of organizationalstructure centralization show higher levels of harmonization of IT. This conclusion wasderived based on a positive relation identified between personal participation in decisionmaking and IT. This relation is expected considering when decisions are centralized,managers have a global view of the company and their investments are more oriented


Table 5.1: Research propositions derived as a result of this research.

Factor Proposition

Product Proposition PROD1: The harmonization of activities is highwhen different roles are assigned to manage products withdifferent strategic focus in the organization.

type

Proposition PROD2: The harmonization of resources is highwhen different roles are assigned to manage products withdifferent strategic focus in the organization.

Proposition PROD3: The harmonization of IT is high whendifferent roles are assigned to manage products with differentstrategic focus in the organization.

Maturity Proposition MAT1: For organizations with a high level ofprocess maturity, the harmonization of resources is high.level

Mergers and Proposition MER1: Organizations with a higher number ofmergers and acquisitions are less harmonized in terms of theirIT systems.

acquisitions

Org.Struct.Proposition ORGC1: The harmonization of IT is high inorganizations with a centralized organizational structure.

centralization

Proposition ORGC2: For high levels of participation indecision making, the harmonization of IT is high.

Org.Struct. Proposition ORGF1: The level of process harmonization ishigh in organizations with a high level of organizationalstructure formalization.

formalization

Proposition ORGF2: The harmonization of resources is high inorganizations with a high level of job codification.

Proposition ORGF3: The harmonization of activities is high inorganizations with a high level of job specificity.

Proposition ORGF4: The harmonization of activities is highwhen the level of written communication is high.

Proposition ORGF5: The harmonization of data is high whenthe level of written communication is high.

5.1. Main conclusions 91

to provide global solutions. In this case, investments in IT systems are more unified toprovide IT solutions to support business processes, instead of looking for investmentsin each individual organizational unit. However, we were not able to find relations forthe other two items due to lack of variety in our sample. All the companies in thesample scored high in both items.

Four items were used to assess the relations between organizational structure formal-ization: job codification, rule observation, specificity of job and written communication.We found positive relations between formalization and the level of harmonization ofactivities, data and resources. We observed that the level of harmonization of resourceswas high when the level of job codification was high. Higher levels of harmonization ofactivities were found when the job specificity and written communication were high.And the harmonization of data was high for organizations with a high level of writtencommunication. The only item that did not show any relations was the rule obser-vation. The only aspect of harmonization that was not related to formalization wasIT.

The only immediate factor included was the level of process structuredness. Ourfindings suggest that the effect of contextual factors differs per process type and thetype of relations (positive or negative) may also differ with respect to different aspectsof process harmonization. The data aspect showed more differences in the type ofrelations between different process types. Also, the majority of the cases showed apositive relation with respect to the level of structuredness.

No relations were found between the number of different locations and any aspectof process harmonization. This finding contradicts the expected negative relation sug-gested in the literature [134]. We may expect that more locations would create amore diverse and complex system landscape, which would negatively influence the in-formation technology aspect. However, we consider that the influence of this factoris dependent on the interaction with other factors such as cultural differences, legalrequirements, or even the level of formalization of each organization. For instance,the effect of multiple locations in one country or region with similar conditions can beless than the effect when the same number of locations are in different countries withcomplete different cultural backgrounds. The underlying reason is that in multina-tional companies different variations may arise due to the limited interaction betweenindividuals performing similar tasks [134].

No significant relations were found between the level of centralization of IT gover-nance and the level of harmonization. This is unexpected considering that previousresearch has proven significant relations with respect to process harmonization in gen-eral [10, 55, 112], but we could not identify any relations with respect to each individualaspect. We revised the appropriateness of our methodology to understand the source ofthis unexpected finding: (1) we dealt with the classical centralized/ decentralized clas-sification of IT governance structure typically used in previous studies; (2) we verifiedthat the sample size and mix was sufficient to relate patterns; (3) the unit of analy-sis was the organizational unit level, which is appropriate for our analysis to providestronger explanations of the governance arrangements [15]. Therefore, we consider thatit is possible that our assumption of non-interacting factors (between contextual factorsin our study) may be unrealistic with respect to the IT governance centralization. Wemay be ignoring some interacting effects or some factors which were not included inour analysis.

To conclude, all the relations described in this section explain how the different


factors determine the level of harmonization of a business process. They are used toanswer the overall research question.

The overall research question is: What contextual factors should be consid-ered to decide on the right level of harmonization of a business process inan organization?

Returning to the outset of this research, the question of whether contextual factorsdetermine the extent to which a business process can be harmonized has to be answeredin positive terms. The results of this research confirm the concept of fit suggested bycontingency theory, showing how contextual factors are determinants of the extent towhich a process can be harmonized.

The contextual factors that showed relations with the different aspects of processharmonization are considered relevant to build propositions for future research direc-tions. Six relevant factors were identified including: product type, maturity level,number of mergers and acquisitions, organizational structure centralization, organi-zational structure formalization and the level of process structuredness. In contrastto the theory suggested by other authors, the centralization of IT governance did notshow any significant relation with any aspect of process harmonization. This surprisingfinding may be explained based on our limitations which are explained in Section 5.4.

5.2 Theoretical contributions

We can distinguish three contributions of this dissertation to theory in business processharmonization.

First, we developed a conceptual model that presents a structured overview of thecontextual factors that determine the extent to which a process can be harmonized,relevant aspects of process harmonization, performance benefits that can be gainedthrough process harmonization, and relations between these concepts. For the devel-opment of our model we used contingency theory as a guiding theory to emphasize thatthe extent to which a process can be harmonize depends on contextual factors of theorganization. Our model builds on existing theory in process harmonization by bring-ing clarity to the vague terminology present in literature about process harmonization,and by providing focus to the contextual factors that have been considered relevant inthe context of harmonization in previous research.

Second, we presented a validated measurement model to assess the level of harmo-nization of business processes in an organization. This measurement model was devel-oped using a multi-method approach which is the first in our knowledge to proposethe concept of process harmonization as multidimensional. By conceptually develop-ing and empirically validating the concept of process harmonization we contribute totheory by providing a set of operational measures can be used to develop normativetheory based upon empirical investigation on process harmonization. This is criticalconsidering that recommendations provided by authors conducting empirical studiesin harmonization are made based on the specifications of their measurement modelsthat relate the latent variable representing a construct to its measures. But the scaledevelopment procedures advocated in literature of process harmonization fail to ade-quately discuss how they define appropriate conceptual definitions of harmonization.Therefore, our hierarchical measurement model can be used to guide future research.

5.3. Practical implications 93

Third, we advanced theory in process harmonization by providing a set of proposi-tions that describe the expected relationship between contextual factors and differentaspects of process harmonization. We provided grounded theory, which is the resultof an iterative process in which we observed patterns within systematically collectedempirical data and compare them with theory.

5.3 Practical implications

Our results provided important managerial guidelines for conducting harmonizationinitiatives, which are described in this section.

Our findings emphasized that it is not always optimal to impose one standardprocess, considering that there is a trade-off involved in this decision. The trade-off consists of the balance between necessary investments to reach a higher level ofstandardization and the expected gains in terms of business performance. Practitionersneed to understand the influencing factors that play a role in this trade-off. Our findingsoffer a good basis for business managers to better understand which contextual factorsare relevant when conducting harmonization projects and the specific aspects of processharmonization that they influence.

The relevance of knowing the influencing factors while conducting harmonizationinitiatives is the same as knowing the effect of turning different knobs in a cockpitfor a pilot. Following this analogy, knowledge of the influencing factors can be seenas knowing which knobs should be turned, and the type of relations (positive or neg-ative) indicates the effect of changing knobs to different sides. The factors that aremore relevant for practitioners are those under their control. The only internal factorevaluated which is out of managers’ control is the number of mergers and acquisitions.This factor showed a negative relation with respect to harmonization of IT. There-fore, it is important to consider that when new investments in IT are the ones leadingharmonization initiatives in an organization.

The way in which managers structure their organization is critical to guaranteethe expected gains of process harmonization. Two dimensions of the organizationalstructure are relevant for the analysis: formalization and centralization. Formalizationrefers to the use of rules in an organization, while centralization refers to the way inwhich power is distributed among different positions. When new investments in ITare the ones leading harmonization initiatives, then it is recommended to centralizethe decision making. This will provide a global overview of the organization to avoidsearching for local solutions. However, this centralization is only necessary at the levelin which the harmonization will take place and should be aligned with the businessgoals. In this case, the investment in IT systems is a significant aspect to be consideredin the evaluation of the trade-off. These results are based on our findings of a positiverelation between the level of organizational structure centralization and IT.

On the other hand, if the harmonization initiative is mainly process driven, thenthe most significant aspect to consider is the organizational structure formalization.In this case, the investments should focus on defining clear procedures in which theactivities to be performed in each process are clearly specified and codified. In this case,investments in resources to build and maintain such procedures have to be consideredin the evaluation of the trade-off. These results are based on our findings of a positiverelation between the level of organizational structure formalization and three aspectsof harmonization: activities, data and resources.


5.4 Limitations

This study is subject to limitations that provide avenues for future research.The first limitation is related to the completeness of the list of contextual factors

identified through literature review. The referred list is only complete with respect toour conceptualization of process harmonization. But in our definition of harmonization,we defined it as a selective standardization and therefore we mainly search for literaturein standardization. If a different conceptualization is provided, the number of sourcesmay differ and even the factors identified.

The second limitation refers to the number of cases selected to identified relationsbetween contextual factors and aspects of process harmonization. Only six cases wereselected for the case study due to restrictions in the resources available to conductthe interviews and the number of companies willing to cooperate. We did not addcases until we reached the saturation point to identify patterns in all the factors.Therefore, some of the factors found through the literature review were excluded duringthe selection of cases and others after data collection for lack of variety in the datato identify patterns. The list of factors identified combined with the limited numberof cases did not allow us to explore all the possible relations. With a more completeset of cases, an interesting research direction is to explore possible interactions amongcontextual factors in the model. It would also allow exploring the strength of theserelations to identify which type of factors (external, internal and immediate) is morerelevant when harmonizing business processes. Further studies are needed to enhanceour results and to improve external validity.

The last limitation argues about the validity of the relations found in our study. Weare cautious to generalize our findings considering that the relations found are basedon a limited sample size, particularly when subdivided into various categories to derivepatterns. We also rely on data that are obtained by interviewing a few top executivesand from documents that the organization made available to researchers.

Appendix A

Literature review selectionstrategy

96 Appendix A. Literature review selection strategy

Figure A.1: FigA1

Appendix B

Code sheet

98 Appendix B. Code sheet

Figure B.1

99

Figure B.2


Figure B.3

101

Figure B.4


Figure B.5

Appendix C

Survey

An online survey was conducted to generate indicators of process harmonization. TheEnglish version of the survey questionnaire is included bellow.

Figure C.1: Survey questionnaire introduction

104 Appendix C. Survey

Figure C.2: Survey questionnaire question 1


105



107

Figure C.8: Survey questionnaire example to introduce next questions

109


111


113


115


Figure C.17: Survey questionnaire final screen

Appendix D

Questionnaire for conductingmultiple-case studies

Description of the sample

1. Company name:

2. Industry type:

3. Company size:

Table D.1

Categories Number of employees Turnover Balanced sheet(in million euros) (in million euros)

Micro < 10 < 2 < 2Small < 50 < 10 < 10

Medium < 250 < 50 < 43Large > 250 > 50 > 43

3a) How many employees does your organization count?

a) 1-9

b) 10-49

c) 50-249

d) 250-749

e) More than 750

3b) How large was the turnover in the previous year (2012, in Euros)?

a) Less than 2 million

b) 2-10 million

c) 10-50 million

d) more than 50 million

118 Appendix D. Questionnaire for conducting multiple-case studies

3c) What is your annual balance-sheet total in the previous year (2012, in Euros)?

a) Less than 5 million

b) 5-27 million

c) More than 27 million

4. Company age. How long does your organization exist?

Contextual factors

5. Number of different locations. The organization operates in differ-ent locations, situated in: . The procurement department is locatedin different places, situated in: .

6. Number of mergers and acquisitions. How many times has your organiza-tion been merged with or bought by another company, since its foundation? When?

7. Organizational structure centralization.

7a) Index of participation in decision making: Answer the following questions usinga 5-point Likert scale: (1)Never, (2)Seldom, (3)Sometimes, (4)Often and (5)Always.

• How frequently do you usually participate in the decision to hire new staff?

• How frequently do you usually participate in decisions on the promotion of anyof the professional staff?

• How frequently do you participate in decisions on the adoption of new policies?

• How frequently do you participate in the decisions on the adoption of newprograms?

7b) Index of hierarchy of authority: Answer the following questions using a 4-pointLikert scale: (1)Definitely false, (2)False, (3)True and (4)Definitely true.

• There can be little action taken here until a supervisor approves a decision.

• A person who wants to make his own decisions would be quickly discouragedhere.

• Even small matters have to be referred to someone higher up for a finalanswer.

• I have to ask my boss before I do almost anything.

• Any decision I make has to have my boss’s approval.

119

7c) Departmental participation in decision making: Answer the following questionsusing a 5-point Likert scale: (1)Strongly disagree, (2)Disagree, (3)Neither agree ordisagree, (4)Agree and (5)Strongly agree.

To what extent do you agree with the following statements?

• Employees participate in decisions involving your work.

• Employees participate in decisions involving their work environment.

8. Organizational structure formalization. Answer the following questionsusing a 4-point Likert scale: (1)Definitely false, (2)False, (3)True and (4)Definitelytrue.

8a) Index of Job codification:

• I feel that I am my own boss in most matters.

• A person can make his own decisions without checking with anybody else.

• How things are done here is left up to the person doing the work.

• People here are allowed to do almost as they please.

• People here make their own rules on the job.

8b) Index of rule observation:

• The employees are constantly being checked on for rule violations.

• People here feel as though they are constantly being watched to see that theyobey all the rules.

8c) Index of Specificity of job:

• Whatever situation arises, we have procedures to follow in dealing with it.

• Everyone has a specific job to do.

• Going through the proper channels is constantly stressed.

• The organization keeps a written record of everyone’s job performance.

• We are to follow strict operating procedures at all times.

• Whenever we have a problem, we are supposed to go to the same person for ananswer.

8d) Written communication: Answer the following questions using a 5-point Likertscale: (1)Strongly disagree, (2)Disagree, (3)Neither agree or disagree, (4)Agree and(5)Strongly agree. To what extent do you agree with the following statements?

• The frequency of written communication in your organization is high.


9. Maturity level.9a) Level 1 (Initial):

• Formal procedures for the execution of processes do not exist in our organization.

• If procedures are defined, they are rarely followed.

• Everybody executes tasks in its own way, in other words: everybody has its ownmethods.

9b) Level 2 (Managed):

• At the beginning of a project, we make agreements about which methods andtechnology we will use.

• If we make agreements about work methods, they will be documented such thatthey can be executed in the same way at another time.

• We use planning and management procedures to control our individual projects.

9c) Level 3 (Standardized):

• Procedures are standardized for the whole organization.

• Work procedures and objectives are well documented in our whole organization.

• Processes are defined such that they will be in the same way by different workgroups.

9d) Level 4 (Predictable):

• Performance is managed statistically (e.g. by measuring KPIs) to understandperformance and to control variation.

• Processes/tasks are managed in such a way that they meet agreed-upon perfor-mance and quality goals.

• If processes do not perform according to predefined standards, they are correctedto meet the quantitative goals.

9e) Level 5 (Innovating):

• Our organization understands its critical business issues and areas of concern byusing feedback from performance measurements.

• Our organization sets quantitative improvement goals to constantly reorganizeprocesses when perceived necessary.

• We constantly pilot with new ideas and new technologies to improve ourprocesses.

10. IT governance centralization. How is the management of the IT governancein your department (Procurement)?

• Centralized

• Decentralized

121

11. Product type.11a) Are there different roles responsible for the procurement of production-related

(direct) and non-production related (indirect) products?

• Yes. Explain why is important to have different roles? What are the im-plications of processing both type of products (production and non-productionrelated) in the same way?

• No

11b) Do you have different roles responsible for the procurement of products andservices?

• Yes , Explain why is important to have different roles? What are the implica-tions of processing both, products and services in the same way?

• No

Questions to be answered for each process (ordering process, supplier’sselection and determining specifications)

12. How many variants of this process exit?

13. Which are the main variants of this process?

14. For each process variant, answer the following question:14.a)Input criteria.

• What are the inputs and outputs of this process?

• Are the type of inputs/ outputs always known and specified?

14.b) Assessment. What can be the assessment of the output of your process?

• i.e. approved or disapproved (binary)

• an expected variety of outputs (can be classified based on experience)

• or even unexpected variety of outputs (apply tacit knowledge)

14.c) Level of structuredness. How is the set of activities that you need toperform to complete this process?

• always follow the same order

• varies within a predefined set of options. The primary source of bad perfor-mance lies in an inappropriate selection of alternatives

• sometimes needs to develop new knowledge and heuristics to accomplish theprocess

14.d) Personal differences. How much experience is required to perform the activi-ties in this process (on average)? How long is the training for this role (the responsibleof this process i.e. indirect procurement manager for supplier selection for indirect?


15. Business process harmonization. Answer the following questions to evalu-ate the level of process harmonization. What is ...

Data

• (IA1)...the percentage of common activities in the process?

• (ID1)...the number of different documents used as input for the same process?

• (ID2)...the number of different output reports?

Resources

• (IR1)...the percentage of common roles in the process?

• (IR2)...the number of different roles executing the same activity?

IT

• (IT1)...the number of different software applications in used in the process?

• (IT2)...the number of different supplier’s paid for the software applications?

• (IT3)...the amount of money paid for the software applications?

Appendix E

Results of pattern matching inmultiple-case studies

The results of the multiple-case studies are depicted in a tabular format organizedper contextual factor. To understand how we derived the different relations betweenfactors and process harmonization,a set of rules are summarized in Table E.1. The firstcolumn of Table E.1 includes the list of contextual factor. Two possible relations aredefined for each factor (positive or negative) as shown in the last column of the table.The type of relation depends on the combination between the different categories ofthe factors and the level of harmonization shown in the bottom of the table. If a lowlevel of the factor is associated with a low level of harmonization, then the relation ispositive, else it is negative. The indicators are divided in two groups: (1) IA1 and(2)ID1-ID2-IR2-IT1. The reason to split them is because a low value of IA1 meansthat the percentage of common activities in the process is low and this is associatedto a low level of harmonization. While a low value of the others, for instance IT1means a low number of different IT systems in used which is related to a high level ofharmonization. Therefore, a distinction is made between these two categories.

An example was provided in subsection 4.2.7 to explain how to interpret the resultsshown in the following tables E.2 until E.14.

124 Appendix E. Results of pattern matching in multiple-case studies

Table E.1: Rules for the classification of the types of relations and the level of harmonization

Contextual factorsIndicators values

Type ofrelations

IA1 ID1-ID2-IR2-IT1low values high values low values high values

Number of differentlocations

low high high low positivehigh low low high negative

IT governancecentralization

decentralized centralized centralized decentralized positivecentralized decentralized decentralized centralized negative

Product typedivided combined combined divided positive

combined divided divided combined negative

Maturity levellow high high low positivehigh low low high negative



Organizational structurecentralization:- Participation in decisionmaking


- Hierarchy of authoritylow high high low positivehigh low low high negative

- Departmentalparticipation


Organizational structureformalization:

- Job codificationlow high high low positivehigh low low high negative

- Rule observationlow high high low positivehigh low low high negative

- Specificity of joblow high high low positivehigh low low high negative

- Written communicationlow high high low positivehigh low low high negative

Level of structurednesslow high high low positivehigh low low high negative

Level of harmonization low high high low —

125

E.0.1 Patterns in number of locations

IndicatorsNumber of locations

RelationsC6 C3 C1 C2 C5 C41 2 3 3 7 50

IA1 75.00 87.50 85.00 68.33 60.00 88.75 —ID1 2.50 4.00 4.50 7.00 11.00 8.00 —ID2 1.00 3.00 2.25 1.00 1.50 2.00 —IR2 6.00 3.50 4.25 4.00 4.00 3.50 —IT1 1.50 1.00 1.50 3.25 1.50 1.25 —



RelationsC6 C1 C2 C5 C41 3 3 7 50

IA1 75.00 75.00 70.00 60.00 75.00 —ID1 4.00 10.00 4.33 3.00 6.00 —ID2 1.00 5.00 1.00 1.00 2.00 —IR2 4.00 6.00 3.00 4.00 3.00 —IT1 2.00 1.00 3.00 1.00 1.00 —




IA1 75.00 87.50 97.50 65.00 60.00 90.00 —ID1 1.00 4.00 1.50 15.00 19.00 10.00 —ID2 1.00 3.00 1.00 1.00 2.00 2.00 —IR2 8.00 3.50 4.50 6.00 4.00 4.00 —IT1 1.00 1.00 2.00 4.00 2.00 1.50 —


Table E.2: Relations identified with pattern matching using the number of locations as a criteria forclustering of cases.


E.0.2 Patterns in IT governance centralization

Indicators

IT governance

RelationsDecentralized Centralized

C4 C1 C2 C3 C5 C60 1 1 1 1 1

IA1 88.75 85.00 68.33 87.50 60.00 75.00 NegativeID1 8.00 4.50 7.00 4.00 11.00 2.50 —ID2 2.00 2.25 1.00 3.00 1.50 1.00 —IR2 3.50 4.25 4.00 3.50 4.00 6.00 —IT1 1.25 1.50 3.25 1.00 1.50 1.50 —


Indicators

IT governance


C4 C1 C2 C5 C60 1 1 1 1

IA1 75.00 75.00 70.00 60.00 75.00 —ID1 6.00 10.00 4.33 3.00 4.00 —ID2 2.00 5.00 1.00 1.00 1.00 —IR2 3.00 6.00 3.00 4.00 4.00 —IT1 1.00 1.00 3.00 1.00 2.00 —


Indicators

IT governance


C4 C1 C2 C3 C5 C60 1 1 1 1 1

IA1 90.00 97.50 65.00 87.50 60.00 75.00 —ID1 10.00 1.50 15.00 4.00 19.00 1.00 —ID2 2.00 1.00 1.00 3.00 2.00 1.00 —IR2 4.00 4.50 6.00 3.50 4.00 8.00 —IT1 1.50 2.00 4.00 1.00 2.00 1.00 —


Table E.3: Relations identified with pattern matching using the IT governance as a criteria forclustering of cases.

127

E.0.3 Patterns in product type

Indicators

Product type

RelationsCombined

C1 C2 C3 C4 C6 C5

IA1 75.00 70.00 87.50 75.00 75.00 60.00 —ID1 10.00 4.33 4.00 6.00 4.00 3.00 —ID2 5.00 3.00 1.00 2.00 1.00 1.00 —IR2 6.00 3.00 3.50 3.00 4.00 4.00 —IT1 1.00 3.00 1.00 1.00 2.00 1.00 —


Indicators

Product type

RelationsCombined

C1 C2 C4 C6 C5

IA1 75.00 70.00 75.00 75.00 60.00 —ID1 10.00 4.33 6.00 4.00 3.00 —ID2 5.00 1.00 2.00 1.00 1.00 —IR2 6.00 3.00 3.00 4.00 4.00 —IT1 1.00 3.00 1.00 2.00 1.00 —


Indicators

Product type

RelationsCombined

C1 C2 C3 C4 C6 C5

IA1 75.00 70.00 87.50 75.00 75.00 60.00 —ID1 10.00 4.33 4.00 6.00 4.00 3.00 —ID2 5.00 3.00 1.00 2.00 1.00 1.00 —IR2 6.00 3.00 3.50 3.00 4.00 4.00 —IT1 1.00 3.00 1.00 1.00 2.00 1.00 —


Table E.4: Relations identified with pattern matching using the product type (direct versus indirects)as a criteria for clustering of cases.


Indicators

Product type

RelationsDivided Combined

C3 C4 C1 C2 C5 C61 1 0 0 0 0

IA1 87.50 88.75 85.00 68.33 60.00 75.00 NegativeID1 4.00 8.00 4.50 7.00 11.00 2.50 —ID2 3.00 2.00 2.25 1.00 1.50 1.00 —IR2 3.50 3.50 4.25 4.00 4.00 6.00 NegativeIT1 1.00 1.25 1.50 3.25 1.50 1.50 Negative


Indicators

Product type


C4 C1 C2 C5 C60 1 1 1 1

IA1 75.00 75.00 70.00 60.00 75.00 —ID1 6.00 10.00 4.33 3.00 4.00 —ID2 2.00 5.00 1.00 1.00 1.00 —IR2 3.00 6.00 3.00 4.00 4.00 —IT1 1.00 1.00 3.00 1.00 2.00 —


Indicators

Product type


C3 C4 C1 C2 C5 C61 1 0 0 0 0

IA1 87.50 90.00 97.50 65.00 60.00 75.00 —ID1 4.00 10.00 1.50 15.00 19.00 1.00 —ID2 3.00 2.00 1.00 1.00 2.00 1.00 PositiveIR2 3.50 4.00 4.50 6.00 4.00 8.00 —IT1 1.00 1.50 2.00 4.00 2.00 1.00 —


Table E.5: Relations identified with pattern matching using the product type (products versus ser-vices) as a criteria for clustering of cases.

129

E.0.4 Patterns in maturity level

IndicatorsMaturity level

RelationsLow HighC6 C2 C4 C1 C3 C53 4 4 5 5 5

IA1 75.00 68.33 88.75 85.00 87.50 60.00 —ID1 2.50 7.00 8.00 4.50 4.00 11.00 NegativeID2 1.00 1.00 2.00 2.25 3.00 1.50 —IR2 6.00 4.00 3.50 4.25 3.50 4.00 PositiveIT1 1.50 3.25 1.25 1.50 1.00 1.50 —



RelationsLow HighC6 C2 C4 C1 C53 4 4 5 5

IA1 75.00 70.00 75.00 75.00 60.00 —ID1 4.00 4.33 6.00 10.00 3.00 —ID2 1.00 1.00 2.00 5.00 1.00 —IR2 4.00 3.00 3.00 6.00 4.00 —IT1 2.00 3.00 1.00 1.00 1.00 —



RelationsLow HighC6 C2 C4 C1 C3 C53 4 4 5 5 5

IA1 75.00 65.00 90.00 97.50 87.50 60.00 —ID1 1.00 15.00 10.00 1.50 4.00 19.00 NegativeID2 1.00 1.00 2.00 1.00 3.00 2.00 —IR2 8.00 6.00 4.00 4.50 3.50 4.00 PositiveIT1 1.00 4.00 1.50 2.00 1.00 2.00 —


Table E.6: Relations identified with pattern matching using the maturity level as a criteria forclustering of cases.


E.0.5 Patterns in number of mergers and acquisitions

IndicatorsNumber of mergers


IA1 87.50 60.00 75.00 85.00 88.75 68.33 —ID1 4.00 11.00 2.50 4.50 8.00 7.00 —ID2 3.00 1.50 1.00 2.25 2.00 1.00 —IR2 3.50 4.00 6.00 4.25 3.50 4.00 —IT1 1.00 1.50 1.50 1.50 1.25 3.25 Negative



RelationsC5 C6 C1 C4 C21 1 2 2 3

IA1 60.00 75.00 75.00 75.00 70.00 —ID1 3.00 4.00 10.00 6.00 4.33 —ID2 1.00 1.00 5.00 2.00 1.00 —IR2 4.00 4.00 6.00 3.00 3.00 —IT1 1.00 2.00 1.00 1.00 3.00 Negative




IA1 87.50 60.00 75.00 90.00 97.50 65.00 —ID1 4.00 19.00 1.00 10.00 1.50 15.00 —ID2 3.00 2.00 1.00 2.00 1.00 1.00 —IR2 3.50 4.00 8.00 4.00 4.50 6.00 —IT1 1.00 2.00 1.00 1.50 2.00 4.00 Negative


Table E.7: Relations identified with pattern matching using the number of mergers and acquisitionsas a criteria for clustering of cases.

131

E.0.6 Patterns in organizational structure centralization

IndicatorsParticipation in decision making


2.50 3.17 3.75 4.50 4.63 5.00

IA1 68.33 85.00 75.00 87.50 60.00 88.75 —ID1 7.00 4.50 2.50 4.00 11.00 8.00 —ID2 1.00 2.25 1.00 3.00 1.50 2.00 —IR2 4.00 4.25 6.00 3.50 4.00 3.50 —IT1 3.25 1.50 1.50 1.00 1.50 1.25 Positive


Indicators

Participation in decision making


2.50 3.17 3.75 4.63 5.00

IA1 70.00 75.00 75.00 60.00 75.00 —ID1 4.33 10.00 4.00 3.00 6.00 —ID2 1.00 5.00 1.00 1.00 2.00 —IR2 3.00 6.00 4.00 4.00 3.00 —IT1 3.00 1.00 2.00 1.00 1.00 Positive


IndicatorsParticipation in decision making


2.50 3.17 3.75 4.50 4.63 5.00

IA1 65.00 97.50 75.00 87.50 60.00 90.00 —ID1 15.00 1.50 1.00 4.00 19.00 10.00 —ID2 1.00 1.00 1.00 3.00 2.00 2.00 —IR2 6.00 4.50 8.00 3.50 4.00 4.00 —IT1 4.00 2.00 1.00 1.00 2.00 1.50 Positive


Table E.8: Relations identified with pattern matching using the participation in decision making asa criteria for clustering of cases.


IndicatorsHierarchy of authority

RelationsHighC1 C2 C4 C6 C3 C53.59 3.50 3.63 3.75 4.00 4.63

IA1 85.00 68.33 88.75 75.00 87.50 60.00 —ID1 4.50 7.00 8.00 2.50 4.00 11.00 —ID2 2.25 1.00 2.00 1.00 3.00 1.50 —IR2 4.25 4.00 3.50 6.00 3.50 4.00 —IT1 1.50 3.25 1.25 1.50 1.00 1.50 —


Indicators

Hierarchy of authority

RelationsHigh

C1 C2 C4 C6 C52.50 3.17 3.75 4.63 5.00

IA1 75.00 70.00 75.00 75.00 60.00 —ID1 10.00 4.33 6.00 4.00 3.00 —ID2 5.00 1.00 2.00 1.00 1.00 —IR2 6.00 3.00 3.00 4.00 4.00 —IT1 1.00 3.00 1.00 2.00 1.00 —


IndicatorsHierarchy of authority

RelationsHighC1 C2 C4 C6 C3 C53.59 3.50 3.63 3.75 4.00 4.63

IA1 97.50 65.00 90.00 75.00 87.50 60.00 —ID1 1.50 15.00 10.00 1.00 4.00 19.00 —ID2 1.00 1.00 2.00 1.00 3.00 2.00 —IR2 4.50 6.00 4.00 8.00 3.50 4.00 —IT1 2.00 4.00 1.50 1.00 1.00 2.00 —


Table E.9: Relations identified with pattern matching using the hierarchy of authority as a criteriafor clustering of cases.

133

IndicatorsDepartmental participation

RelationsHighC4 C1 C2 C3 C5 C6

4.00 4.00 4.00 4.00 4.00 4.50

IA1 88.75 85.00 68.33 87.50 60.00 75.00 —ID1 8.00 4.50 7.00 4.00 11.00 2.50 —ID2 2.00 2.25 1.00 3.00 1.50 1.00 —IR2 3.50 4.25 4.00 3.50 4.00 6.00 —IT1 1.25 1.50 3.25 1.00 1.50 1.50 —


Indicators

Departmental participation

RelationsHigh

C4 C1 C2 C5 C64.00 4.00 4.00 4.00 4.50

IA1 75.00 75.00 70.00 60.00 75.00 —ID1 6.00 10.00 4.33 3.00 4.00 —ID2 2.00 5.00 1.00 1.00 1.00 —IR2 3.00 6.00 3.00 4.00 4.00 —IT1 1.00 1.00 3.00 1.00 2.00 —


IndicatorsDepartmental participation

RelationsHighC4 C1 C2 C3 C5 C6

4.00 4.00 4.00 4.00 4.00 4.50

IA1 90.00 97.50 65.00 87.50 60.00 75.00 —ID1 10.00 1.50 15.00 4.00 19.00 1.00 —ID2 2.00 1.00 1.00 3.00 2.00 1.00 —IR2 4.00 4.50 6.00 3.50 4.00 8.00 —IT1 1.50 2.00 4.00 1.00 2.00 1.00 —


Table E.10: Relations identified with pattern matching using the departmental participation as acriteria for clustering of cases.


E.0.7 Patterns in organizational structure formalization

IndicatorsJob codification

RelationsHigh LowC1 C6 C2 C3 C4 C5

2.13 2.50 2.75 3.00 3.00 3.00

IA1 85.00 75.00 68.33 87.50 88.75 60.00 —ID1 4.50 2.50 7.00 4.00 8.00 11.00 —ID2 2.25 1.00 1.00 3.00 2.00 1.50 —IR2 4.25 6.00 4.00 3.50 3.50 4.00 PositiveIT1 1.50 1.50 3.25 1.00 1.25 1.50 —


Indicators

Job codification

RelationsHigh Low

C1 C6 C2 C4 C52.13 2.50 2.75 3.00 3.00

IA1 75.00 75.00 70.00 75.00 60.00 —ID1 10.00 4.00 4.33 6.00 3.00 —ID2 5.00 1.00 1.00 2.00 1.00 —IR2 6.00 4.00 3.00 3.00 4.00 —IT1 1.00 2.00 3.00 1.00 1.00 —


IndicatorsJob codification

RelationsHigh LowC1 C6 C2 C3 C4 C5

2.13 2.50 2.75 3.00 3.00 3.00

IA1 97.50 75.00 65.00 87.50 90.00 60.00 —ID1 1.50 1.00 15.00 4.00 10.00 19.00 —ID2 1.00 1.00 1.00 3.00 2.00 2.00 NegativeIR2 4.50 8.00 6.00 3.50 4.00 4.00 PositiveIT1 2.00 1.00 4.00 1.00 1.50 2.00 —


Table E.11: Relations identified with pattern matching using the job codification as a criteria forclustering of cases.

135

IndicatorsRule observation


2.50 2.50 2.50 2.81 3.44 4.38

IA1 87.50 88.75 75.00 60.00 85.00 68.33 —ID1 4.00 8.00 2.50 11.00 4.50 7.00 —ID2 3.00 2.00 1.00 1.50 2.25 1.00 —IR2 3.50 3.50 6.00 4.00 4.25 4.00 —IT1 1.00 1.25 1.50 1.50 1.50 3.25 —


Indicators

Rule observation

RelationsLow High

C4 C6 C5 C1 C22.50 2.50 2.81 3.44 4.38

IA1 75.00 75.00 60.00 75.00 70.00 —ID1 6.00 4.00 3.00 10.00 4.33 —ID2 2.00 1.00 1.00 5.00 1.00 —IR2 3.00 4.00 4.00 6.00 3.00 —IT1 1.00 2.00 1.00 1.00 3.00 —


IndicatorsRule observation


2.50 2.50 2.50 2.81 3.44 4.38

IA1 87.50 90.00 75.00 60.00 97.50 65.00 —ID1 4.00 10.00 1.00 19.00 1.50 15.00 —ID2 3.00 2.00 1.00 2.00 1.00 1.00 —IR2 3.50 4.00 8.00 4.00 4.50 6.00 —IT1 1.00 1.50 1.00 2.00 2.00 4.00 —


Table E.12: Relations identified with pattern matching using the rule observation as a criteria forclustering of cases.


IndicatorsSpecificity of job


2.50 2.92 3.54 3.75 3.96 4.06

IA1 68.33 60.00 87.50 88.75 75.00 85.00 PositiveID1 7.00 11.00 4.00 8.00 2.50 4.50 —ID2 1.00 1.50 3.00 2.00 1.00 2.25 —IR2 4.00 4.00 3.50 3.50 6.00 4.25 —IT1 3.25 1.50 1.00 1.25 1.50 1.50 —


Indicators

Specificity of job

RelationsLow High

C2 C5 C4 C6 C12.50 2.92 3.75 3.96 4.06

IA1 70.00 60.00 75.00 75.00 75.00 PositiveID1 4.33 3.00 6.00 4.00 10.00 —ID2 1.00 1.00 2.00 1.00 5.00 —IR2 3.00 4.00 3.00 4.00 6.00 —IT1 3.00 1.00 1.00 2.00 1.00 —


IndicatorsSpecificity of job


2.50 2.92 3.54 3.75 3.96 4.06



Table E.13: Relations identified with pattern matching using the specificity of job as a criteria forclustering of cases.

137



2.50 4.00 4.00 4.50 5.00 5.00



Indicators



2.50 4.00 4.00 4.50 5.00

IA1 60.00 70.00 75.00 75.00 75.00 PositiveID1 3.00 4.33 4.00 10.00 6.00 NegativeID2 1.00 1.00 1.00 5.00 2.00 —IR2 4.00 3.00 4.00 6.00 3.00 —IT1 1.00 3.00 2.00 1.00 1.00 —




2.50 4.00 4.00 4.50 5.00 5.00



Table E.14: Relations identified with pattern matching using the written communication as a criteriafor clustering of cases.


E.0.8 Patterns in level of structuredness

IndicatorsDifference in percentage

Average RelationsC1 C2 C3 C4 C5 C6

IA1 23.08 -7.69 — 16.67 0.00 0.00 9.49 —ID1 -566.67 71.11 — 40.00 84.21 -300.00 212.40 —ID2 -400.00 0.00 — 0.00 50.00 0.00 90.00 —IR2 -33.33 50.00 — 25.00 0.00 50.00 31.67 —IT1 50.00 25.00 — 33.33 50.00 -100.00 51.67 —

(a) Differences in percentages between indicators of ordering and supplier’s selection

IndicatorsValues per case

RelationsC1 C2 C4 C5 C6

IA1 75.00 70.00 75.00 60.00 75.00 —ID1 10.00 4.33 6.00 3.00 4.00 —ID2 5.00 1.00 2.00 1.00 1.00 —IR2 6.00 3.00 3.00 4.00 4.00 —IT1 1.00 3.00 1.00 1.00 2.00 —

(b) Indicators per case for supplier’s selection process

IndicatorsValues per case

RelationsC1 C2 C3 C4 C5 C6

IA1 97.50 65.00 87.50 90.00 60.00 75.00 —ID1 1.50 15.00 4.00 10.00 19.00 1.00 —ID2 1.00 1.00 3.00 2.00 2.00 1.00 —IR2 4.50 6.00 3.50 4.00 4.00 8.00 —IT1 2.00 4.00 1.00 1.50 2.00 1.00 —

(c) Indicators per case for Ordering process

Table E.15: Relations identified with pattern matching using the level of structuredness as a criteriafor clustering of cases.

Bibliography

[1] Michael Aiken and Jerald Hage. Organizational interdependence and intra-organizational structure. American sociological review, pages 912–930, 1968. (citedon pp. 66 and 79)

[2] Zhiyi Ang and Peter Massingham. National culture and the standardization versusadaptation of knowledge management. Journal of Knowledge Management, 11(2):5–21,2007. (cited on pp. 17, 18, and 22)

[3] JUST Armstrong and Terry Overton. Estimating nonresponse bias in mail surveys.Journal of marketing research, 14:396–402, 1977. (cited on p. 41)

[4] JC aWetherbe. The mis conundrumcentralization or decentralization. Chief FinancialOfficer USA, 1988. (cited on p. 75)

[5] Bjorn Axenath, Ekkart Kindler, and Vladimir Rubin. The aspects of business processes:An open and formalism independent ontology. Fachberichte Informatik tr-ri-05-256,Universitat Paderborn, 2005. (cited on p. 88)

[6] Richard P Bagozzi and Claes Fornell. Theoretical concepts, measurements, and meaning.A second generation of multivariate analysis, 2(2):5–23, 1982. (cited on p. 35)

[7] Jan-Michael Becker, Kristina Klein, and Martin Wetzels. Hierarchical latent variablemodels in pls-sem: guidelines for using reflective-formative type models. Long RangePlanning, 2012. (cited on pp. 35 and 43)

[8] Daniel Beimborn, Fabian Gleisner, Nils Joachim, and Andreas Hackethal. The roleof process standardization in achieving it business value. In System Sciences, 2009.HICSS’09. 42nd Hawaii International Conference on, pages 1–10. IEEE, 2009. (citedon pp. 1, 2, 20, 22, 23, 25, 27, 28, 34, and 84)

[9] Peter M Blau and Richard A Schoenherr. The structure of organisations. Basic Book,New York, NY, 1971. (cited on p. 61)

[10] Kees Boersma and Sytze Kingma. Developing a cultural perspective on erp. BusinessProcess Management Journal, 11(2):123–136, 2005. (cited on pp. 18, 22, 75, 83, and 91)

[11] Kenneth A Bollen. Structural equation models. Wiley Online Library, 1998. (cited onp. 43)

[12] Andrew C Boynton, Gerry C Jacobs, and Robert W Zmud. Whose responsibility isit (information technology) management? Sloan Management Review, 33(4):32, 1992.(cited on p. 65)

[13] John Brewer. Flow of communications, expert qualifications and organizational author-ity structures. American Sociological Review, pages 475–484, 1971. (cited on p. 61)

[14] A Brown and G Grant. Framing the frameworks: A review of it governance research.Communications of the Association for Information Systems (Volume 15, 2005), 696(712):712, 2005. (cited on pp. 65, 66, and 75)

140 BIBLIOGRAPHY

[15] Carol V Brown. Examining the emergence of hybrid is governance solutions: Evidencefrom a single case site. Information systems research, 8(1):69–94, 1997. (cited on pp. 75and 91)

[16] Carol V Brown and Sharon L Magill. Alignment of the is functions with the enterprise:toward a model of antecedents. Mis Quarterly, pages 371–403, 1994. (cited on p. 75)

[17] Dirk Buchta, Marcus Eul, and Helmut Schulte-Croonenberg. It optimizationreducingcosts without diminishing returns. Strategic IT Management: Increase value, controlperformance, reduce costs, pages 133–153, 2007. (cited on pp. 20, 22, 26, and 27)

[18] Carl F Cargill. Why standardization efforts fail. Journal of Electronic Publishing, 14(1), 2011. (cited on p. 2)

[19] Donna Carmichael. Ibm’s journey towards a market-driven process-managed businessmodel. Journal of Market-Focused Management, 2(1):99–103, 1997. (cited on p. 1)

[20] Susan Cartwright and Richard Schoenberg. Thirty years of mergers and acquisitionsresearch: Recent advances and future opportunities. British Journal of Management,17(S1):S1–S5, 2006. (cited on p. 2)

[21] Alfred Dupont Chandler. Strategy and structure, volume 4. MIT press Cambridge, MA,1962. (cited on p. 3)

[22] Chuansheng Chen, Shin-ying Lee, and Harold W Stevenson. Response style and cross-cultural comparisons of rating scales among east asian and north american students.Psychological Science, pages 170–175, 1995. (cited on p. 70)

[23] H Kurt Christensen and Cynthia A Montgomery. Corporate economic performance:Diversification strategy versus market structure. Strategic Management Journal, 2(4):327–343, 1981. (cited on p. 4)

[24] Gilbert A Churchill Jr. A paradigm for developing better measures of marketing con-structs. Journal of marketing research, pages 64–73, 1979. (cited on pp. 34 and 36)

[25] L. Clark. Mittal scraps global standardisation plan. Computer Weekly, 2007. (cited onp. 2)

[26] Debbie Collins. Pretesting survey instruments: an overview of cognitive methods. Qual-ity of Life Research, 12(3):229–238, 2003. (cited on p. 40)

[27] European Commission et al. Commission recommendation of 6 may 2003 concerningthe definition of micro, small and medium-sized enterprises. Official Journal of theEuropean Union, L, 124(36):2003, 2003. (cited on p. 65)

[28] James M Conway and Allen I Huffcutt. A review and evaluation of exploratory factoranalysis practices in organizational research. Organizational research methods, 6(2):147–168, 2003. (cited on p. 42)

[29] Harris M Cooper. Scientific guidelines for conducting integrative research reviews. Re-view of educational research, 52(2):291–302, 1982. (cited on pp. 12 and 13)

[30] Harris M Cooper. Organizing knowledge syntheses: A taxonomy of literature reviews.Knowledge in Society, 1(1):104–126, 1988. (cited on p. 12)

[31] Mick P Couper, Michael W Traugott, and Mark J Lamias. Web survey design andadministration. Public opinion quarterly, 65(2):230–253, 2001. (cited on pp. 39 and 40)

[32] Scott D Crawford, Mick P Couper, and Mark J Lamias. Web surveys perceptions ofburden. Social science computer review, 19(2):146–162, 2001. (cited on p. 39)

[33] John W Creswell. Research design: Qualitative, quantitative, and mixed methods ap-proaches. Sage Publications, Incorporated, 2013. (cited on p. 67)

[34] Tomi Dahlberg, Niina Mallat, Jan Ondrus, and Agnieszka Zmijewska. Past, present andfuture of mobile payments research: A literature review. Electronic Commerce Researchand Applications, 7(2):165–181, 2008. (cited on p. 3)

BIBLIOGRAPHY 141

[35] Qizhi Dai, Robert J Kauffman, and Bin Wang. The value of it-enabled business processstandardization from the real options perspective. In Exploring the Grand Challengesfor Next Generation E-Business, pages 160–165. Springer, 2011. (cited on pp. 19, 22,26, and 28)

[36] Fariborz Damanpour. The adoption of technological, administrative, and ancillary in-novations: Impact of organizational factors. Journal of management, 13(4):675–688,1987. (cited on p. 4)

[37] Thomas H Davenport. The coming commoditization of processes. Harvard BusinessReview, 83(6):100–108, 2005. (cited on pp. 1 and 65)

[38] Gregory G Dess and Richard B Robinson. Measuring organizational performance in theabsence of objective measures: The case of the privately-held firm and conglomeratebusiness unit. Strategic management journal, 5(3):265–273, 1984. (cited on p. 33)

[39] Don A Dillman. Mail and Internet surveys: The tailored design method–2007 Updatewith new Internet, visual, and mixed-mode guide. Wiley. com, 2011. (cited on pp. 38,39, and 46)

[40] Robert B Duncan. Characteristics of organizational environments and perceived envi-ronmental uncertainty. Administrative science quarterly, pages 313–327, 1972. (cited onpp. 3 and 84)

[41] Christof Ebert. Best Practices in Software Measurement: How to Use Metrics to ImproveProject and Process Performance; 37 Tables. Springer, 2005. (cited on p. 33)

[42] Jeffrey R Edwards. Multidimensional constructs in organizational behavior research:An integrative analytical framework. Organizational Research Methods, 4(2):144–192,2001. (cited on p. 43)

[43] Kathleen M Eisenhardt. Building theories from case study research. Academy of man-agement review, 14(4):532–550, 1989. (cited on pp. 63, 64, 70, 72, 73, and 82)

[44] Kathleen M Eisenhardt. Better stories and better constructs: the case for rigor andcomparative logic. Academy of Management review, 16(3):620–627, 1991. (cited onp. 62)

[45] Kathleen M Eisenhardt and Melissa E Graebner. Theory building from cases: oppor-tunities and challenges. Academy of management journal, 50(1):25–32, 2007. (cited onpp. 62, 63, and 69)

[46] Jean-Francois Etter and Thomas V Perneger. Analysis of non-response bias in a mailedhealth survey. Journal of clinical epidemiology, 50(10):1123–1128, 1997. (cited on p. 42)

[47] Jude Fernandez and Jyoti Bhat. Addressing the complexities of global process harmo-nization. Handbook of Research on Complex Dynamic Process Management: Techniquesfor Adaptability in Turbulent Environments, IGI Global, pages 368–385, 2010. (cited onpp. 1, 4, and 5)

[48] Claes Fornell and David F Larcker. Evaluating structural equation models with unob-servable variables and measurement error. Journal of marketing research, pages 39–50,1981. (cited on p. 43)

[49] L Annette Foster and Billy T Herndon. Implementation of a reengineered procurementprocess at duke university, using technology and work process design. Proceedings ofthe Information Profession and the Information Profession, 1997. (cited on p. 33)

[50] Lawrence E Fouraker and John M Stopford. Organizational structure and the multina-tional strategy. Administrative Science Quarterly, pages 47–64, 1968. (cited on p. 3)

[51] Frances X Frei, Ravi Kalakota, Andrew J Leone, and Leslie M Marx. Process vari-ation as a determinant of bank performance: evidence from the retail banking study.Management Science, 45(9):1210–1220, 1999. (cited on p. 1)

142 BIBLIOGRAPHY

[52] Ron Garland. The mid-point on a rating scale: Is it desirable. Marketing Bulletin, 2(1):66–70, 1991. (cited on p. 70)

[53] Cees J Gelderman and Janjaap Semeijn. Managing the global supply base throughpurchasing portfolio management. Journal of Purchasing and Supply Management, 12(4):209–217, 2006. (cited on p. 64)

[54] Ari Ginsberg and Nenkat Venkatraman. Contingency perspectives of organizationalstrategy: a critical review of the empirical research. Academy of Management Review,10(3):421–434, 1985. (cited on p. 4)

[55] Stephane JG Girod and Joshua B Bellin. Revisiting the modern multinational enterprisetheory: An emerging-market multinational perspective. Research in Global StrategicManagement, 15:167–210, 2011. (cited on pp. 5, 18, 19, 20, 22, 26, 27, 64, 75, 78, 83,and 91)

[56] David E Gray. Doing research in the real world. Sage, 2009. (cited on pp. 64, 69, 70,and 73)

[57] Peter H Grinyer and Masoud Yasai-Ardekani. Strategy, structure, size and bureaucracy.Academy of Management journal, 24(3):471–486, 1981. (cited on p. 4)

[58] Anil K Gupta and Vijay Govindarajan. Business unit strategy, managerial characteris-tics, and business unit effectiveness at strategy implementation. Academy of Manage-ment journal, 27(1):25–41, 1984. (cited on p. 4)

[59] Jerald Hage and Michael Aiken. Relationship of centralization to other structural prop-erties. Administrative Science Quarterly, pages 72–92, 1967. (cited on pp. 61, 66, 78,79, and 89)

[60] JF Hair, WC Black, BJ Babin, RE Anderson, and RL Tatham. Multivariate dataanalysis sixth edition pearson education. New Jersey, 2006. (cited on pp. 42, 43,and 50)

[61] Alena Hallerbach, Thomas Bauer, and Manfred Reichert. Capturing variability in busi-ness process models: the provop approach. Journal of Software Maintenance and Evo-lution: Research and Practice, 22(6-7):519–546, 2010. (cited on p. 1)

[62] Donald C Hambrick, Marta A Geletkanycz, and James W Fredrickson. Top executivecommitment to the status quo: Some tests of its determinants. Strategic ManagementJournal, 14(6):401–418, 1993. (cited on p. 46)

[63] Michael Hammer. Reengineering work: don’t automate, obliterate. Harvard businessreview, 68(4):104–112, 1990. (cited on p. 33)

[64] Michael Hammer and Steven Stanton. How process enterprises really work. Harvardbusiness review, 77:108–120, 1999. (cited on p. 1)

[65] Paul Harmon. Business process change: a guide for business managers and BPM andsix sigma professionals. Morgan Kaufmann, 2010. (cited on p. 1)

[66] Evi Hartmann, Gerhard Trautmann, and Christopher Jahns. Organisational designimplications of global sourcing: A multiple case study analysis on the application ofcontrol mechanisms. Journal of Purchasing and Supply Management, 14(1):28–42, 2008.(cited on p. 63)

[67] Markus Helfert. Challenges of business processes management in healthcare: Experiencein the irish healthcare sector. Business Process Management Journal, 15(6):937–952,2009. (cited on pp. 18 and 22)

[68] Michael A Hitt, R Duane Ireland, and KA Palia. Industrial firms’ grand strategy andfunctional importance: Moderating effects of technology and uncertainty. Academy ofManagement Journal, 25(2):265–298, 1982. (cited on p. 4)

BIBLIOGRAPHY 143

[69] Christian Homburg, John P Workman Jr, and Harley Krohmer. Marketing’s influencewithin the firm. The Journal of Marketing, pages 1–17, 1999. (cited on p. 3)

[70] Andreas Hufgard and Eduard Gerhardt. Consolidating business processes as exemplifiedin sap erp systems. In S-BPM ONE-Learning by Doing-Doing by Learning, pages 155–171. Springer, 2011. (cited on pp. 1, 5, and 22)

[71] Watts S Humphrey. Characterizing the software process: a maturity framework. Soft-ware, IEEE, 5(2):73–79, 1988. (cited on pp. 65 and 66)

[72] Lawrence R James. Aggregation bias in estimates of perceptual agreement. Journal ofapplied psychology, 67(2):219, 1982. (cited on p. 78)

[73] Lawrence R Jauch, Richard N Osborn, and William F Glueck. Short term financialsuccess in large business organizations: The environment-strategy connection. StrategicManagement Journal, 1(1):49–63, 1980. (cited on p. 4)

[74] Jayanth Jayaram, Shawnee K Vickery, and Cornelia Droge. The effects of informationsystem infrastructure and process improvements on supply-chain time performance. In-ternational Journal of Physical Distribution & Logistics Management, 30(3/4):314–330,2000. (cited on p. 1)

[75] Burke Johnson and Lisa A Turner. Data collection strategies in mixed methods research.Handbook of mixed methods in social and behavioral research, pages 297–319, 2003. (citedon p. 64)

[76] Takashi Kobayashi, Sen’ichi Onoda, and Norihisa Komoda. Workflow business templatefor application processes in administration department. Information Technology andManagement, 3(1-2):43–66, 2002. (cited on pp. 22, 26, and 27)

[77] R. Kumar. Research Methodology: A Step-by-Step Guide for Beginners. SAGE Publi-cations, 2005. ISBN 9781412911948. URL http://books.google.nl/books?id=x_kp_

_WmFzoC. (cited on p. 15)

[78] Sameer Kumar and Ralph Harms. Improving business processes for increased opera-tional efficiency: a case study. Journal of Manufacturing Technology Management, 15(7):662–674, 2004. (cited on pp. 22, 26, and 27)

[79] Lance B Kurke and Howard E Aldrich. Notemintzberg was right!: A replication andextension of the nature of managerial work. Management Science, 29(8):975–984, 1983.(cited on p. 65)

[80] Marcello La Rosa, Marlon Dumas, Arthur HM Ter Hofstede, and Jan Mendling. Config-urable multi-perspective business process models. Information Systems, 36(2):313–340,2011. (cited on p. 1)

[81] Paul R Lawrence and Jay W Lorsch. Differentiation and integration in complex orga-nizations. Administrative science quarterly, pages 1–47, 1967. (cited on p. 61)

[82] Paul R Lawrence, Jay W Lorsch, and James S Garrison. Organization and environment:Managing differentiation and integration. Division of Research, Graduate School ofBusiness Administration, Harvard University Boston, MA, 1967. (cited on p. 3)

[83] Jerry W Lee, Patricia S Jones, Yoshimitsu Mineyama, and Xinwei Esther Zhang. Cul-tural differences in responses to a likert scale. Research in nursing & health, 25(4):295–306, 2002. (cited on p. 70)

[84] Paul Lillrank. The quality of standard, routine and nonroutine processes. OrganizationStudies, 24(2):215–233, 2003. (cited on pp. 19, 20, 22, 23, 64, 65, and 66)

[85] Chengfei Liu, Qing Li, and Xiaohui Zhao. Challenges and opportunities in collaborativebusiness process management: Overview of recent advances and introduction to thespecial issue. Information Systems Frontiers, 11(3):201–209, 2009. (cited on p. 2)

http://books.google.nl/books?id=x_kp__WmFzoC

http://books.google.nl/books?id=x_kp__WmFzoC

144 BIBLIOGRAPHY

[86] Jan-Bernd Lohmoller. Latent variable path modeling with partial least squares. Physica-Verlag Heidelberg, 1989. (cited on p. 43)

[87] Scott B MacKenzie, Philip M Podsakoff, and Nathan P Podsakoff. Construct measure-ment and validation procedures in mis and behavioral research: integrating new andexisting techniques. MIS quarterly, 35(2):293–334, 2011. (cited on pp. 34 and 43)

[88] Karl B Manrodt and Kate Vitasek. Global process standardization: a case study.Journal of Business Logistics, 25(1):1–23, 2004. (cited on pp. 1, 22, and 26)

[89] Paula Ventura Martins and Alberto Rodrigues da Silva. A case study applying processand project alignment methodology. Journal of the Brazilian Computer Society, 12(3):65–82, 2006. (cited on p. 22)

[90] Tim McLaren, Milena Head, and Yufei Yuan. Supply chain collaboration alternatives:understanding the expected costs and benefits. Internet Research, 12(4):348–364, 2002.(cited on pp. 22, 26, and 27)

[91] Nigel Melville, Kenneth Kraemer, and Vijay Gurbaxani. Review: Information technol-ogy and organizational performance: An integrative model of it business value. MISquarterly, 28(2):283–322, 2004. (cited on p. 5)

[92] Danny Miller. Environmental fit versus internal fit. Organization science, 3(2):159–178,1992. (cited on pp. 3 and 84)

[93] Danny Miller, Manfred FR Kets De Vries, and Jean-Marie Toulouse. Top executivelocus of control and its relationship to strategy-making, structure, and environment.Academy of Management journal, 25(2):237–253, 1982. (cited on p. 4)

[94] Henry Mintzberg. The structuring of organizations: A synthesis of the research. Uni-versity of Illinois at Urbana-Champaign’s Academy for Entrepreneurial Leadership His-torical Research Reference in Entrepreneurship, 1979. (cited on p. 61)

[95] Linda Moffat and Norm Archer. Knowledge management in production alliances. In-formation Systems and e-Business Management, 2(2-3):241–267, 2004. (cited on pp. 18,20, 22, and 26)

[96] Ole Mortensen and Olga W Lemoine. Integration between manufacturers and third partylogistics providers? International Journal of Operations & Production Management, 28(4):331–359, 2008. (cited on pp. 22, 26, and 27)

[97] Bjoern Munstermann, Andreas Eckhardt, and Tim Weitzel. Join the standard forces-examining the combined impact of process and data standards on business process per-formance. In System Sciences, 2009. HICSS’09. 42nd Hawaii International Conferenceon, pages 1–10. IEEE, 2009. (cited on pp. 22, 23, 24, 25, 26, 27, 34, and 84)

[98] Bjoern Munstermann, Peter Moederer, and Tim Weitzel. Setting up and managingbusiness process standardization: Insights from a case study with a multinational e-commerce firm. In System Sciences (HICSS), 2010 43rd Hawaii International Confer-ence on, pages 1–11. IEEE, 2010. (cited on pp. 1, 16, 22, 23, 26, 27, and 61)

[99] Bjorn Munstermann, Andreas Eckhardt, and Tim Weitzel. The performance impact ofbusiness process standardization: An empirical evaluation of the recruitment process.Business Process Management Journal, 16(1):29–56, 2010. (cited on pp. 2, 16, 22, 23,25, 26, 27, 34, 40, 43, and 84)

[100] David Nadler and Michael Tushman. Competing by design: The power of organizationalarchitecture. Oxford University Press, 1997. (cited on p. 64)

[101] Andy Neely, Mike Gregory, and Ken Platts. Performance measurement system design: aliterature review and research agenda. International Journal of Operations & ProductionManagement, 25(12):1228–1263, 2005. (cited on p. 5)

[102] Alex Norta and Rik Eshuis. Specification and verification of harmonized business-processcollaborations. Information Systems Frontiers, 12(4):457–479, 2010. (cited on p. 22)

BIBLIOGRAPHY 145

[103] Johannes Pennings. Measures of organizational structure: A methodological note. Amer-ican Journal of Sociology, pages 686–704, 1973. (cited on pp. 61, 65, 66, 79, and 89)

[104] Alessandro Perego and Alessandro Salgaro. Assessing the benefits of b2b trade cycleintegration: a model in the home appliances industry. Benchmarking: An InternationalJournal, 17(4):616–631, 2010. (cited on pp. 22, 26, 27, and 61)

[105] Charles Perrow. A framework for the comparative analysis of organizations. Americansociological review, pages 194–208, 1967. (cited on pp. 61, 66, and 89)

[106] Lieven Quintens, Pieter Pauwels, and Paul Matthyssens. A globalizing perspectiveon purchasing strategies. In Proceeding of the Annual IMP Conference, Rotterdam,volume 21, 2005. (cited on pp. 18, 22, 26, 27, and 61)

[107] Justus J Randolph. A guide to writing the dissertation literature review. PracticalAssessment, Research & Evaluation, 14(13):2, 2009. (cited on pp. 12, 14, and 16)

[108] Albrecht Richen and Ansgar Steinhorst. Standardization or harmonization? you needboth. European Health Informatics (November 2005), 2005. (cited on pp. 5 and 22)

[109] Christian Ringle, Marko Sarstedt, and Detmar Straub. A critical look at the use ofpls-sem in mis quarterly. MIS Quarterly (MISQ), 36(1), 2012. (cited on pp. 35 and 43)

[110] Michael Rosemann, Jan Recker, and Christian Flender. Contextualisation of businessprocesses. International Journal of Business Process Integration and Management, 3(1):47–60, 2008. (cited on pp. 18 and 22)

[111] Christoph Rosenkranz, Stefan Seidel, Jan Mendling, Markus Schaefermeyer, and JanRecker. Towards a framework for business process standardization. In Business processmanagement workshops, pages 53–63. Springer, 2010. (cited on pp. 18, 19, 21, 22, 76,and 83)

[112] Jeanne W Ross, Peter Weill, and David C Robertson. Enterprise architecture as strategy:Creating a foundation for business execution. Harvard Business Press, 2006. (cited onpp. 18, 22, 23, 25, 34, 75, 77, 80, 83, and 91)

[113] Denise M Rousseau and Yitzhak Fried. Location, location, location: contextualizingorganizational research*. Journal of organizational Behavior, 22(1):1–13, 2001. (citedon p. 4)

[114] Jennifer Rowley. Using case studies in research. Management research news, 25(1):16–27, 2002. (cited on pp. 62, 63, and 82)

[115] Herbert J Rubin and Irene S Rubin. Qualitative interviewing: The art of hearing data.Sage, 2011. (cited on p. 69)

[116] Richard P Rumelt. Strategy, structure, and economic performance. Division of Research,Graduate School of Business Administration, Harvard University Boston, MA, 1974.(cited on p. 3)

[117] Alzira Salama, Wayne Holland, and Gerald Vinten. Challenges and opportunities inmergers and acquisitions: three international case studies–deutsche bank-bankers trust;british petroleum-amoco; ford-volvo. Journal of European industrial training, 27(6):313–321, 2003. (cited on pp. 2, 77, and 89)

[118] Vallabhajosyula Sambamurthy and Robert W Zmud. Arrangements for informationtechnology governance: a theory of multiple contingencies. MIS quarterly, pages 261–290, 1999. (cited on p. 75)

[119] Dipl-Kfm Markus Schafermeyer, Christoph Rosenkranz, and Roland Holten. The im-pact of business process complexity on business process standardization. Business &Information Systems Engineering, 4(5):261–270, 2012. (cited on pp. 16, 22, 23, 25, 34,40, 43, and 84)

146 BIBLIOGRAPHY

[120] Markus Schafermeyer, Daniel Grgecic, and Christoph Rosenkranz. Factors influencingbusiness process standardization: A multiple case study. In System Sciences (HICSS),2010 43rd Hawaii International Conference on, pages 1–10. IEEE, 2010. (cited on pp. 4,5, 19, 20, 21, 22, and 26)

[121] Donald P Schwab. Construct validity in organizational behavior. Research in organiza-tional behavior, 2(1):3–43, 1980. (cited on p. 35)

[122] Shari Shang and Peter B Seddon. Managing process deficiencies with enterprise systems.Business Process Management Journal, 13(3):405–416, 2007. (cited on p. 22)

[123] K Sheikh and S Mattingly. Investigating non-response bias in mail surveys. Journal ofepidemiology and community health, 35(4):293–296, 1981. (cited on p. 42)

[124] Andrew K Shenton. Strategies for ensuring trustworthiness in qualitative researchprojects. Education for information, 22(2):63–75, 2004. (cited on p. 82)

[125] Ismail Sila. Examining the effects of contextual factors on tqm and performance throughthe lens of organizational theories: an empirical study. Journal of Operations manage-ment, 25(1):83–109, 2007. (cited on p. 4)

[126] Olav Sorenson and Jesper B Sørensen. Finding the right mix: Franchising, organiza-tional learning, and chain performance. Strategic Management Journal, 22(6-7):713–724,2001. (cited on pp. 18, 22, and 83)

[127] Gray Southon, Chris Sauer, and Kit Dampney. Lessons from a failed information systemsinitiative: issues for complex organisations. International journal of medical informatics,55(1):33–46, 1999. (cited on p. 2)

[128] William A Stock. Systematic coding for research synthesis. The handbook of researchsynthesis, 236:125e138, 1994. (cited on p. 16)

[129] Anslem Straus and Juliet Corbin. Basics of qualitative research: Techniques and pro-cedures for developing grounded theory. 2 nd Ed, 1998. (cited on p. 36)

[130] James Thomas and Angela Harden. Methods for the thematic synthesis of qualitativeresearch in systematic reviews. BMC medical research methodology, 8(1):45, 2008. (citedon p. 16)

[131] Hans B Thorelli and Stephen C Burnett. The nature of product life cycles for industrialgoods businesses. The Journal of Marketing, pages 97–108, 1981. (cited on p. 4)

[132] Richard J Torraco. Writing integrative literature reviews: Guidelines and examples.Human Resource Development Review, 4(3):356–367, 2005. (cited on p. 12)

[133] Roger Tourangeau. The psychology of survey response. Cambridge University Press,2000. (cited on p. 37)

[134] Roger Tregear. Business process standardization. In Handbook on Business ProcessManagement 2, pages 307–327. Springer, 2010. (cited on pp. 1, 2, 17, 18, 19, 20, 21, 22,23, 25, 26, 27, 34, 73, 74, 83, and 91)

[135] Hans van Leijen. The role of contextuality in process standardization. In KnowledgeManagement and Management Learning, pages 251–286. Springer, 2005. (cited on p. 22)

[136] Diederik W van Liere, Lorike Hagdorn, Martijn R Hoogeweegen, and Peter HM Vervest.Embedded coordination in a business network. Journal of Information Technology, 19(4):261–269, 2004. (cited on pp. 22 and 27)

[137] Arij Jan van Weele and Arjan J Weele. Purchasing and supply chain management:Analysis, strategy, planning and practice. Cengage Learning EMEA, 2005. (cited onpp. 2, 64, 65, and 66)

[138] Natarjan Venkatraman and Vasudevan Ramanujam. Measurement of business perfor-mance in strategy research: a comparison of approaches. Academy of managementreview, 11(4):801–814, 1986. (cited on p. 5)

BIBLIOGRAPHY 147

[139] EM Von Simson. The recentralization of it. Computerworld, 29(51):1–5, 1995. (citedon p. 75)

[140] Chris Voss, Nikos Tsikriktsis, and Mark Frohlich. Case research in operations man-agement. International journal of operations & production management, 22(2):195–219,2002. (cited on pp. 64, 69, 71, and 82)

[141] Cordula Wagner, Peter P Groenewegen, Dinny H de Bakker, and Gerrit van der Wal.Environmental and organizational determinants of quality management. Quality Man-agement in Healthcare, 9(4):63–76, 2001. (cited on p. 4)

[142] Jane Webster and Richard T Watson. Analyzing the past to prepare for the future:Writing a. MIS quarterly, 26(2), 2002. (cited on pp. 16 and 36)

[143] Leslie Willcocks. The next step for the ceo: moving it-enabled services outsourcingto the strategic agenda. Strategic Outsourcing: An International Journal, 3(1):62–66,2010. (cited on p. 2)

[144] Christopher Winsten and Margaret Hall. The measurement of economies of scale. TheJournal of Industrial Economics, pages 255–264, 1961. (cited on p. 35)

[145] Robert M Worcester and Timothy R Burns. A statistical examination of the relativeprecision of verbal scales. Journal of the Market Research Society, 17(3):181–197, 1975.(cited on p. 70)

[146] Kim Wullenweber, Daniel Beimborn, Tim Weitzel, and Wolfgang Konig. The impact ofprocess standardization on business process outsourcing success. Information SystemsFrontiers, 10(2):211–224, 2008. (cited on pp. 2, 4, 5, 22, 23, 25, 26, 27, 34, 61, and 84)

[147] Robert K Yin. Case study research: Design and methods, volume 5. Sage, 2009. (citedon pp. 82 and 84)

[148] Valarie A Zeithaml, Carl P Zeithaml, et al. The contingency approach: its founda-tions and relevance to theory building and research in marketing. European Journal ofMarketing, 22(7):37–64, 1988. (cited on p. 3)

[149] Fang Zhao. Management of information technology and business process re-engineering:a case study. Industrial Management & Data Systems, 104(8):674–680, 2004. (cited onpp. 22, 26, 27, and 61)

Summary


In the attempt to enhance competitiveness mergers and acquisitions have emerged asvery popular and widely adopted business strategies of this globalized era. As a result,challenges arise when organizations engage in joint activities, considering that differentorganizations implement their business processes differently and interact differentlywith the partners and supplier in their value chains. Significant challenges arise for thecoordination of activities and information exchange that need to take place within theirvalue chain, considering the process variants existent among companies and businessunits.

To overcome these challenges, business process harmonization has become a neces-sity for global operating companies. Managers have been interested in the definitionof standard processes, considering that it suggests operational economies of scale andthe development of uniform best practices within a wide range of advantages. Theadvantages of process harmonization are well acknowledged in the literature. However,several attempts of business processes harmonization have failed in practice, whichhave been attributable to factors such as their lack of flexibility to attend to the re-quirements of the different process variants. This situation shows the trade-off faced bycompanies for the definition of a standard process. The trade-off consists of making achoice between the extent to which a process should be harmonized and the necessaryinvestments to achieve the expected harmonization level.

Despite indications of high levels of failure in harmonization initiatives and thehigh economic investments required to conduct these efforts, study of the factors thatinfluence this trade-off is still in its infancy. Therefore, the aim of this research is toidentify the factors that determine the extent to which a business process can be har-monized and the type of influence exerted by each factor on different aspects of processharmonization. Knowledge of these influencing factors is a prerequisite to formulatestrategies for ensuring the expected benefits of process harmonization. Without a com-plete understanding of the big picture, there are very significant risks that the benefitsof any harmonization initiative will not be exploited.

Based on contingency theory, we propose that the extent to which a business pro-cess can be harmonized is dependent on contextual factors in the organization and itsenvironment. Three studies were conducted to achieve the goal of this research. Thefirst study identifies contextual factors that determine the extent to which a process canbe harmonized and the performance increase that the organization achieves through

150 Summary

process harmonization. These factors are identified through a structured analysis ofexisting literature and summarized in a conceptual model. In total, eleven contextualfactors were identified and classified in three groups (immediate, internal and external)based on their level of influence. Immediate factors are contextual factors that describea process that is being harmonized. Two immediate factors were identified: the levelof process structuredness and personal differences in terms of experience and knowl-edge of resources. Internal factors are those that describe the internal environmentof an organization. Six internal factors identified include: the number of locations,level of centralization of IT governance, differences in product type, maturity level,organizational structure and number of mergers and acquisitions. External factor arethose beyond the control of an individual organization. Three external factors wereidentified relevant to our study: cultural differences, different regulations and powerdistance with other organizations in the business network in which the organizationoperates.

The second study examines the operationalization of process harmonization. Theresearch approach consisted on two phases: an exploratory phase to derive a mea-surement model using literature review and interviews; and a confirmatory phase toempirically evaluate the model, using survey-based research. Factors analysis was usedto assess the dimensionality of the construct under study, while Partial least square(PLS) was chosen to analyze the measurement model. A valid and reliable second-order hierarchical measurement model was developed, to evaluate the level of processharmonization in an organization. As part of the conceptualization, four aspects of pro-cess harmonization (data, activities, resources and information technology) and sevenmeasures were identified.

Finally, the third study examines the type of effect of contextual factors on thedifferent aspects of process harmonization. Six comparative case studies are used toexplore the relationship between the contextual factors identified in the first studyand the level of harmonization measured using the model derived in the second study.The results of this last study show that the role of the contextual factors differ perprocess type, being more significant for more standard processes than less standardones. Relations were found between four contextual factors (maturity level, processtype, number of mergers and organizational structure) and all the different aspects ofprocess harmonization, which can be summarized in the following statements:

• The harmonization of resources is high in organizations with a high level of pro-cess maturity.

• The harmonization of IT is high in organizations with a high number of mergersand acquisitions.

• The harmonization of activities, resources and IT is high when different roles areassigned to manage products with different strategic focus in the organization.

• The harmonization of IT is high when the organizational structure is centralized.

• The harmonization of data, resources and activities is high when the formalizationof the organizational structure is high.

This study contributes to the exiting theory in process harmonization by sheddinglight on the specific aspects of a process that can be influenced by contextual factors.Our results provided important managerial guidelines for conducting harmonizationinitiatives. The first managerial implication is that it is not always optimal to imposeone standard process, considering that there is a trade-off involved in this decision. Our

151

findings offer a good basis for business managers to better understand which contextualfactors are relevant when conducting harmonization projects and the specific aspectsof process harmonization that they influence. The way in which managers structuretheir organization is critical to guarantee the expected gains of process harmonization.When new investments in IT are the ones leading harmonization initiatives, then itis recommended to centralize the decision making. The recommended centralizationin this case provides a global overview of the organization to avoid searching for localsolutions and investments in IT systems have to be included in the analysis of the trade-off. However, when a harmonization initiative is mainly process driven, then the mostsignificant aspect to consider is the organizational structure formalization. In this case,the investments should focus on defining clear procedures in which the activities to beperformed in each process are clearly specified and codified. Investments in resourcesto build and maintain such procedures have to be considered in the evaluation of thetrade-off.

About the Author

Heidi Romero was born on 22 November 1978 in Santo Domingo, Dominican Republic.She graduated from secondary school at Colegio Serafın de Asıs in 1995. In that yearshe enrolled in the studies of Industrial Engineering at Instituto Tecnologico de SantoDomingo (Intec), from which she obtained a Bachelor degree with a summa cum laudedistinction in 1999.

In 1998 she started an internship followed by a position as leader for the implemen-tation of quality (ISO 9000) and environmental (ISO 14000) standards at the Packagingdepartment of Empresas Leon Jimenes. In 2001, Heidi moved to Puerto Rico to carryout a master’s program in Management Science at University of Puerto Rico-Mayaguezcampus (RUM), receiving her Master in Science in 2003. Her master’s thesis entitled“Genetic algorithm approach for reorder cycle time determination in multi-stage sys-tem” was supervised by dr. Viviana Cesanı and prof.dr. Sonia Bartolomei. Duringthe program she also worked as a student assistant in courses related to simulationmodeling, queueing theory and applied industry problems.

After her graduation, she returned to Santo Domingo to get exposure to practicalexperience in different fields of Industrial Engineering. She worked as a ManagementAnalyst in Ambev (InBev Brewery Division for Dominican Republic) Logistic Coor-dinator in Puerto Rico Storage and Distribution (PRSD), Customer Service Analystand Buyer in Colgate International. In 2004 Heidi also started lecturing undergraduatestudents in Industrial Engineering at Intec, which she continues until today throughvirtual courses.

In 2007 Heidi moved to Eindhoven to enroll in a master’s program in OperationsManagement and Logistics at Eindhoven University of Technology (TU/e) receivingher degree in January 2010. The title of her master’s thesis supervised by dr.ir. NicoDellaert and dr. Monique Jansen-Vullers is “Improving the admission and capacityplanning in a dermatology oncology outpatient clinic”.

In October 2009 Heidi started her PhD research at the Industrial Engineering andInnovation Sciences department at Eindhoven University of Technology. Her researchproject was co-funded by the European Supply Chain Forum (ESCF), Information Sys-tems (IS) group and Innovation Technology Entrepreneurship and Marketing (ITEM)group. Besides her PhD research, Heidi was involved in activities of the Beta PhDCouncil and the Student Supply Chain Forum. Her research project was supervised byprof.dr.ir. Paul Grefen, prof.dr. Arjan van Weele and dr.ir. Remco Dijkman. In 2014,she completed her doctoral thesis entitled “The role of contextual factors in processharmonization”. Results of this research are presented in this dissertation.

The role of contextual factors in process harmonization · Javier, Francisco Javier, Manuel Jos e,...

Documents

Transcript of The role of contextual factors in process harmonization · Javier, Francisco Javier, Manuel Jos e,...