ERP package selection and implementation in multi-site companies Master Thesis

(1)

ERP package selection and implementation in multi-site

companies

Master Thesis

Valerija Olsevska

Student number: S2549468

MSc Technology and Operations Management

First supervisor:

Prof. dr. ir. J.C. Wortmann

Second supervisor:

Dr. ir. S. Fazi

Groningen

(2)

2

Abstract

Purpose – Multi-site enterprises deciding to implement the ERP system face a choice between mainly three options - the multi-site ERP package, the extended single instance of ERP package or implementation of multiple ERP packages. Numerous studies provide an extensive overview of critical success factors in selection and implementation of ERP systems. However, knowledge gaps are still present in a solution selection for geographically dispersed enterprises. Thus, motivated by the gap, the purpose of this paper is to define the criteria that determine the choice between these ERP system architectures.

Design/method/approach - The paper follows the design science research approach, where the selection criteria is presented in the form of decision tables that can facilitate multi-site enterprises in the selection process of ERP. The criteria is derived basing on the available literature and in-depth interviews with experts.

Findings – The interview results showed that there are several dimensions that should be discussed prior the ERP implementation. The first dimension to look at, when choosing the ERP package, is whether the organization is located in one or more countries. The second group of criteria are internal factors, i.e. factors associated with characteristics of a company and its operations.

Originality/value – Although ERP selection and implementation has been addressed by the literature, the topic has been scarcely analyzed in detail with respect to multi-site companies. This paper provides managerial guidelines in the form of decision tables.

(3)

3

Acknowledgments

This thesis is the final work of the master studies in Technology and Operations Management at the University of Groningen. Several people have contributed to writing this thesis and deserved to be acknowledged and thanked.

Foremost, I would like to express my gratitude to my supervisor Hans Wortmann for sharing knowledge, teaching to be critical, giving feedbacks and a valuable guidance. I also appreciate the collaboration with my colleague, Marit Pool, who has worked on the same topic. Our frequent meetings with Professor Wortmann and Marit helped me to stay motivated and dedicated throughout the entire process.

I am especially grateful to each of experts who agreed to contribute and share their valuable knowledge. Without their participation and input this thesis would not have been possible in a way it is presented.

Finally, I want to sincerely thank my family and friends for being extremely supportive and providing continuous encouragements throughout all ups and downs.

June 2017,

(6)

6

List of Abbreviations

ATO – Assemble to Order CSF – Critical Success Factors DSR – Design Science Research EIS – Enterprise Information System ERP – Enterprise Resource Planning ETO – Engineer to Order

(7)

7

List of Tables

Table 3.1. Research steps, approaches and objectives ... 20

Table 3.2. The quadrants of the decision table ... 21

Table 4.1. Summary of literature review ... 24

Table 5.1. Summary of the interview results ... 26

Table 6.1. Defining the setup of the ERP system ... 38

Table 6.2. Defining the setup of procurement ... 39

Table 6.3. Defining the setup of sales and distribution ... 41

(8)

8

List of Figures

Figure 2.1. Major phases of ERP life-cycle (Stefanou 2001) ... 13

Figure 2.2. Enterprise information architecture framework (adapted from Clemmons & Simon (2001)) ... 15

Figure 2.3. Semi - autonomous units functioning in an independent fashion (Clemmons & Simon 2001) ... 15

Figure 2.4. Function centralization without high level of coordination (Clemmons & Simon 2001) ... 16

Figure 2.5. Total centralization (Clemmons & Simon 2001) ... 17

Figure 2.6. Networked organization (Clemmons & Simon 2001) ... 18

(9)

9

1. Introduction

Enterprise information systems (EIS) continue to evolve rapidly, providing new opportunities to achieve efficient management of business processes and enhance competitive advantage. Emergence of Enterprise Resource Planning (ERP) systems has significantly decreased the number of other enterprise applications developed to manage and coordinate information across functional business processes (Sherer 2010).

ERP is a commercial software package that provides solutions to an integrated business organization. ERP vendors offer a variety of software options for any type and size of business. Despite that the software becomes more affordable and less complex, ERP integration is still associated with challenges and risks, and requires careful and unique technical and managerial choices. Substantial investment in the system pre-implementation stage does not guarantee the success of ERP, and the failure rate of ERP systems is still considered as high (Dwivedi et al. 2014). This problem intensifies, when organizations are structurally complex and geographically dispersed, since the organizational structure should be resembled in the system (Ortner & Krenn 2016). However, exactly this type of organizations, which consist of several remote units or divisions, i.e. multi-site enterprises, benefit significantly from an investment in the ERP system (Ranganathan & Brown 2006).

(10)

10 Research in ERP is still a growing field due to continuous technological development, IT innovations and operational improvement. A lot of literature emphasizes benefits which the ERP system creates by integration and accommodation of multiple sites’ processes (van Fenema et al. 2007; Hitt et al. 2002). Numerous studies provide extensive overviews of critical success factors in the selection and implementation of these EIS for various sizes of organizations (Françoise et al. 2009; Muscatello et al. 2008).

However, the focus of these studies is mainly on the implementation of ERP throughout an enterprise as one system. This creates the assumption that each site of an enterprise behaves in the same way as the rest of the organization. In reality it is not entirely true and each site might require a different configuration of the enterprise system. Thus, a multi-site problem arises and enterprises need to choose the most suitable ERP system configuration that would support their business processes and objectives both on the parent and site level, minimizing the possibility of conflicts regarding requirements and implementation details (Morton & Hu 2008; Kirsch & Haney 2006).

Several efforts have been made to address these ERP multi-site issues, however, due to the scope and complexity of the problem, knowledge gaps are still present. For example, Daneva & Wieringa (2006) pose a question: “in what ways can an ERP system be arranged differently while achieving the same goals?” They also emphasize the need for research in the alignment of ERP architecture with complex organizational structures and development of guidelines, practical suggestions and defining good practices (Daneva & Wieringa 2006: 203). In addition, Lai (2001) points on the gap in research concerning establishing an information system between the parent company and its subsidiaries. According to his study, most research attempts evaluate information system integration from the parent perspective, neglecting the position of sites. As a result inappropriate system implementation strategies are developed, that create conflicts between the parent company and its sites.

(11)

11 ERP systems1_{. The multi-site ERP refers to the configuration, where some modules of the}

system are centralized, while some functions are performed by a site. In the multiple separate ERP case a site has its own independent system. The extended single instance single-site ERP assumes using one system throughout the whole organization. Basing on this, the main research questions is:

What are the criteria that define the choice between the multi-site ERP, the extended single instance of ERP or multiple separate ERP package selection?

Basing on the answer to the research question the paper develops the conceptual model that can be used for decision making process concerning the choice of the ERP system set-up.

The paper is organized as follows. First, the general theoretical background and the research conceptualization is described. Next, the description of the research methodology to collect and analyze the data is outlined. Then, the literature survey regarding the criteria is presented, followed by a summary of findings from interviews, from which the final solution models are developed. The paper concludes with conclusions and a discussion on findings, their implications to research and practice and limitations of the study.

1_{In this paper these three choices are referred as ERP system configurations, architecture or the set-up of the}

(12)

12

2. Theoretical background

To start, a general literature review of scholarly articles discussing ERP and possible ERP system configurations was conducted with the help of Google Scholar, ScienceDirect, EmeraldInsight and the search engine of University of Groningen, SmartCat. Search terms included ‘multi-site’, ‘multinationals’, ‘corporations’, ‘selecting’, ‘selection’, ‘manufacturing, ‘multiple instances’, ‘system architecture’, ‘configuration’. These keywords were used in combination with the keyword ‘ERP’. By using the abbreviation ‘ERP’ it was assured that the content relates to this type of the enterprise software. Further relevant articles were identified by following up internal references within the papers retrieved from the initial search.

The literature review begins with the general introduction to ERP software, followed by the description of the selection process and an introduction of complexities associated with ERP implementation in multi-site companies. Next, possible ERP system configurations with the respect to organizational structures are presented. The chapter ends with the conceptualization of the research.

2.1 Characteristics of ERP

In the heart of the ERP system is a shared database that creates company-wide computing environment and cross-functional integration (Davenport 1998). The ERP software provides a solution for the fragmentation of information in the enterprise by enabling “the integration of transactions-oriented data and business processes throughout an organization” (Markus, Axline, et al. 2000: 245). ERP covers all areas of business and composes from such modules as Marketing and Sales, Accounting and Finance, Human Resources and Supply Chain Management. ERP continues to expand its scope and software providers offer other business extensions such as Customer Relationship Management, Business Intelligence, Asset Management, etc. Therefore, the ERP software can be described as multifunctional in its scope, integrated in nature and modular in structure (Clemmons & Simon 2001).

(13)

13 As a commercial product, the ERP software is standardized and targeted on a wide range of industries and is highly configurable to meet diverse needs of various users (Klaus et al. 2000). The ERP package offers opportunities of a system customization as well as choosing a combination of modules that are the most suitable for an enterprise. However, providing the generic solution with series of underlying assumptions, ERP systems force enterprises to re-engineer business structures and align inconsistent operating practices in order to integrate them into seamless information flows (Davenport 1998). Thus, the implementation of ERP is a mix of business process re-engineering and software configuration (Al-Mudimigh et al. 2001). 2.2 Multi-site ERP selection and implementation

The Figure 2.1 shows that ERP package selection is one of the software lifecycle phases and involves defining business needs and requirements, as well as company’s capabilities and constraints in terms of ERP package specifications. It is important to address the gap between requirements and constraints of the company, so that the right solution can be chosen. Constraints can be technological, organizational and financial inefficiencies (Stefanou 2001).

Figure 2.1. Major phases of ERP life-cycle (Stefanou 2001)

(14)

14 framework can serve as general guidance in ERP software evaluation, the process of selecting ERP for multi-site companies is not clearly addressed in the literature.

Considerable research is done in evaluating CSFs, technical as well as organizational and cultural factors that affect ERP system selection and implementation. While these can also be referred to a multi-site ERP, additional considerations should be made in a case of a multi-site enterprise. Several publications emphasize various tradeoffs that occur in the multi-site ERP package selection and implementation (Chae & Poole 2005; Kirsch & Haney 2006; van Fenema et al. 2007).

The selection of the appropriate ERP package for multi-site enterprises is especially complicated, because multiple stakeholders are engaged in the process of determining requirements for the system. Often diverse and sometimes mutually conflicting local sites’ needs and priorities have to be addressed and compromised if necessary (Kirsch & Haney 2006). The development of criteria can simplify the process of requirements negotiation between stakeholders of the organization.

2.3 Configurations

(15)

15 among different sub-units or sites within an organization (Clemmons & Simon 2001). These two concepts help to choose an appropriate ERP system architecture.

Figure 2.2. Enterprise information architecture framework (adapted from Clemmons & Simon (2001))

Figure 2.3. Semi - autonomous units functioning in an independent fashion (Clemmons & Simon 2001)2

Semi - autonomous units functioning in an independent fashion – in this type of a configuration headquarters control sub-units, i.e. sites, only at the financial level. ERP systems are configured, implemented and maintained independently by each site (Markus, Tanis, et al. 2000). The Figure 2.3 shows that sites are connected to headquarters via a controlling module which is

2_{Clemmons & Simon (2001) use the example of the IBM Personal System Group; SID is a database (Storage}

(16)

16 used for financial accounting and reporting. Thus, a minimal level of control and coordination over local processes is required (Clemmons & Simon 2001). This strategy is mostly used by diversified manufacturers.

Function centralization without high level of coordination – these companies seek a high degree of control but use minimal coordination to achieve their objectives. A single site ERP package is normally selected (Markus, et al. 2000). All operations and functions are centralized at headquarters and enterprise units are connected to the single instance of the ERP package with one database through the user interface (Clemmons & Simon 2001). A single site software is mostly used when entities, e.g. plants or distribution centers, have common business processes.

Figure 2.4. Function centralization without high level of coordination (Clemmons & Simon 2001)

(17)

17 the overall supply chain of the organization with a degree of customization to the local market (Clemmons & Simon 2001). Thus, this strategy better suits companies of a bigger size and operating in several countries.

Figure 2.5. Total centralization (Clemmons & Simon 2001)

(18)

18

Figure 2.6. Networked organization (Clemmons & Simon 2001)

Some hybrid configuration arrangements also exist. Markus, Tanis, et al. (2000) distinct a configuration, where headquarters have high coordination of transactions between sites. This type of configuration provides a high degree of a site’s autonomy in operations, but headquarters remain its control over the overall supply chain. Headquarters have access to information regarding purchasing, inventories and production. Such configuration is used when advantages can be created from common purchasing or serving a common customer (Markus, Tanis, et al. 2000).

2.4 Research conceptualization

(19)

19 The Figure 2.7 shows the research conceptualization. The aim is to identify the influencing factors, i.e. criteria that define the fit between the ERP configuration and the required level of control or coordination.

Figure 2.7. Research conceptualization

(20)

20

3. Methodology

This chapter discusses the research approach applied and steps performed that leaded to final results of the paper. It starts with the overall outline of the research design, followed by the description of data collection and analysis methods. At the end the research validity is discussed.

3.1 Research design

Despite having an explorative nature, this research develops a conceptualized model that can serve as guidance in the decision making process for the ERP package selection. Therefore, as a research strategy, that aims to develop instrumental knowledge and addresses practical problems, the design science research (DSR) methodology is applied. Contrary to the pure-knowledge research strategy, DSR is conceptualized as a research approach, that focuses on “knowledge that can be used in an instrumental way to design and implement actions, processes or systems to achieve desired outcomes in practice” (Aken et al. 2016:1). The design science is a problem-solving paradigm and seeks to create an artifact, i.e. a generic design, that solves a practical problem (Hevner et al. 2004).

In this paper the choice between different multi-site ERP system architectures is viewed as a practical problem that enterprises face. Thus, the problem of ERP selection is driven both by business needs and a scientific knowledge gap. In order to solve this problem the generic solution was designed in the form of decision tables that can support the management in the implementation of ERP.

In order to achieve the research goal two major steps were conducted, presented in the Table 3.1.

Table 3.1. Research steps, approaches and objectives

Research process Research approach Research objectives Exploring the

problem and

defining the criteria

Literature review Semi-structured interviews

- To identify criteria that influence the setup of ERP system in the multi-site company.

- To establish relations between criteria and a particular ERP system setup. Developing the tool

for decision making

(21)

21 Exploring the problem and defining the criteria. First, concepts and criteria derived from previously conducted studies were confirmed in current settings with richer data available. The qualitative method was utilized to achieve these objectives, because it provides an opportunity to analyze the problem in settings, where little is known and is hard to define and quantify variables (Perry et al. 2000). As a method, which can provide with rich qualitative data, i.e. factors and criteria that determine the choice of ERP, in-depth interviews with experts in the field was chosen.

The main advantage of the interview method is its holistic approach. Environmental complexity is preserved and is taken into account, instead of attempting to control it (Lutters & Seaman 2007). As a result, a variety of problem perspectives can be captured, and contexts and causes of a particular ERP system setup are explored and analyzed. Moreover, practitioners have valuable knowledge of ERP systems that is not presented extensively in the literature. Thus, currently the main source of knowledge in the ERP field are people whose work is closely related to ERP systems. Translating the experience of experts and deriving the criteria from the practice helps to capture and systematize this knowledge.

Developing the artifact for decision making. Basing on the results acquired from the literature review and in-depth interviews with experts, interrelations between factors and the ERP system setups were identified. That formed a ground for developing decision tables as a conceptual solution for the decision making process. The decision table, consisting from four quadrants (see Table 3.2), is a tool to model a complex conditional logic (Huysmans et al. 2011). The left part of the table refers to conditions and actions, while the two right quadrants represent the entries. The column of the entry part forms a rule with the corresponding action to take.

Table 3.2. The quadrants of the decision table

Conditions Condition alternatives Actions Actions entries

(22)

22 potential solution design for ERP selection problem. Nevertheless, the experts had an opportunity to provide feedback on the developed decision tables.

3.2 Data collection

To capture the knowledge of experts, data collection was performed by semi-structured interviews with open questions. Lutters and Seaman (2007) describe open interviews as a conversational style of data collection, where concepts are co-constructed by the interviewer and participants. Open questions provided an opportunity to identify new criteria that have not been mentioned in the literature yet.

Overall, four in-depth interviews were conducted with experts who have significant experience in multi-site ERP systems. All individuals interviewed were located in the Netherlands and mainly have worked in the consultancy sector, additionally one of the respondents has been involved in the ERP software development. Al-Mudimigh et al. (2001) emphasize the importance of consultants, due to their critical capabilities and an in-depth knowledge of the software. Furthermore, consultants were chosen as a source of information because during their work practice they encounter numerous ERP implementations, and can draw parallels between different ERP projects and identify common patterns that lead to a certain result.

Interviews lasted approximately 1 hour, and were conducted by one interviewer and recorded with the permission of respondents. Interviews were based on the protocol in order to assure reliability and validity of the study (Karlsson, 2016). Despite that a semi-structured interview assumes some freedom and a conversation usually goes beyond the bounds of the protocol, the protocol served as a prompt and set the direction of the interview (Lutters & Seaman 2007; Karlsson 2016). In addition, questions were modified, excluded or new were added during the data collection phase, as the author was gaining deeper understanding of the topic, thus enhancing the dialogue with experts. Since the solution development started during the data collection that also influenced the structure of subsequent interviews. The concepts identified in the theoretical background formed a base for questions. The sample questions set is presented in the Appendix A.

3.3 Data coding and analysis

(23)

23 interview transcript and data reduction, followed by data coding. Coding is assigning tags or labels to pieces of data, e.g. phrases, sentences or paragraphs (Miles & Huberman, 1994). Coding helps to identify and establish concepts, in this case ERP selection criteria and influencing factors, and arrange data into categories by spotting common properties or dimensions within and across analyzed interview transcripts ( Karlsson, 2016, Rog & Bickman 2009).

Two approaches of data coding exist – deductive and inductive. In this study the inductive technique was applied, and in-vivo codes were assigned to passages while reading. That type of codes are 1st order concepts and codes’ names are close to the initial information expressed by a respondent. Since interviews were quite extensive and to avoid a high number of codes, 1st order concepts were reviewed, compared and merged if similarities were identified. After that the 1st order codes were grouped into 2nd order concepts, in Atlas.ti that was represented by assigning codes to code families. While 1st order codes represented ERP selection criteria and influencing factors, 2nd order codes mainly referred to ERP system modules, e.g. purchasing, sales, inventory management etc. As the final step the code tree or the network was built. That enabled to see the interconnections between different codes and formed a ground for further development of decision tables. Examples of networks can be found in the Appendix B. 3.4 Research validity

(24)

24

4. Deriving criteria from the literature review

Keywords used in the literature search regarding factors leading to a particular ERP system setup can be divided into two groups. First, articles mentioning multinationals, corporations, international enterprises, big companies with regards to ERP implementation were viewed. The second search strategy was focused on finding relevant factors forcing companies to avoid “vanilla” implementation of the single instance of ERP. Here such keywords as ‘customization’, ‘best-of-breed’, ‘configuration’, ‘differentiation’ were used. Reasons why companies, for example, customize their systems are relevant in defining selection criteria, because they can explain why a certain site sometimes requires a different solution from another site. The summary of the literature review is presented in the Table 4.1.

Table 4.1. Summary of literature review

Criteria/ Factors Findings relevant to a multi-site enterprise Authors

N at ion al di ff er en ce

s Culture and _language  Cultural and language differences affect selection, adaption and

centralization of ERP, and can force a site to implement its own solution due to package’s limitations to support several single bite and double bite languages, communication barriers and the cultural effect on operating processes.

Sheu et al. (2004)

Government regulations

 Trade restrictions between countries, incl. import/export regulations, affect trans-border data flows and transactions, forcing companies to modify ERP systems and can form challenges for implementation of the single instance ERP.

Yen & Sheu (2004) C om pan y cha rac te ri st ics Company’s and site’s culture

 Site autonomy, differences in goals, culture and local expertise can cause organizational conflict;

 The tradeoff of corporate standardization versus local optimization;

 The tension between the global integration and local adaptation of a site. Umble et al. (2003), van Fenema et al. (2007), Company size  Large and medium-sized firms implement multiple systems due

to the complexity of the organization;

 Majority of small firms implement a single ERP package.

Mabert et al. (2003) Organizational

structure

 Some multi-site organizational structures have a good fit with ERP systems, while others not;

 Companies with the low formalization level and high structural differentiation and decentralization have lower likelihood of the ERP success;

 The single instance of ERP does not support all organizational structures. Ifinedo & Nahar (2009), Markus et al. (2000) Morton & Hu (2008) Technical personnel and labor skills

 The internal technical personnel and labor skills influence centralization of ERP implementation decisions. The lack of qualified human resources at a site can force to outsource some of functions such as site management and support to another site.

(25)

25

Competitive advantage

 Companies focused on customization or volume flexibility, implement the ERP to support their competitive advantage with more information sharing, higher local autonomy, more software packages adaptation, and easier accessibility to the ERP database;

 Strategy focused on the quality consistency and MTS production approach requires standardized manufacturing and operating processes, centralization and control, more processes adaptation, limited information sharing between sites, and low information accessibility;

 No clear findings regarding competitive strategy based on fast and on-time delivery.

Yen & Sheu (2004) O per at ion al and s it e c h ar ac ter is ti cs

Site autonomy  The degree of a site autonomy depends on the degree of process and product consistency across sites and the need or desire for centralized control of information, system setup and usage.

Umble et al. (2003) Inter-dependencies and differentiation among sites

 The greater the interdependence of one plant with other plants in an organization, the greater the ERP related coordination improvements accrued;

 The greater the differentiation of a plant from other plants, the lower the ERP related coordination and efficiency improvements accrued by that plant.

Gattiker & Goodhue (2005) Production strategy

 The production strategy affects the difficulty of selecting an ERP system, and MTO companies find selection of the ERP more difficult than MTS companies;

 The lack of alignment between MTO company needs and ERP functionality (need for add-ons).

Aslan et al. (2012, 2015)

Operational requirements

 The alignment within and between operational requirements and ERP system structures: the flexibility of the system is related to the system decentralization, and benefits arise when the variety of transaction routines is aligned with the number of ERP system structures, e.g. the level of order processing should match the level of the ERP decentralization.

Bendoly & Jacobs (2006)

(26)

26

5. Interview results

This chapter discusses the main findings from interviews. First, a general summary of results is presented, and the following sections are organized according to the criteria division into external and internal factors.

5.1 Summary of findings

Basing on the interview results the main step in the selection and implementation of the ERP package for a multi-site company is defining the scope of implementation. That includes analyzing the company’s structure and its operations, and deciding which operations are integrated into the single solution and what is required to be isolated.

Aligning the ERP system structure with the organizational structure means defining legal entities and operating units and interrelations between them. Accordingly, the following conclusion can be already identified: in order to choose the right ERP system and its configuration, first, a firm needs to be analyzed from the legal point of view by looking at the legal entities and, secondly, from the supply chain perspective, by analyzing operating units3. As a result the organizational structure emerges that is incorporated into the ERP system and is ultimately reflected in the ERP multi-site data structure.

Assignment of operating entities to legal entities is important due to financial and legal rules. Thus, the first dimension to look at, when choosing the ERP package, is whether the organization resides in one or more countries. Criteria associated with the country’s characteristics and legal regulations can be grouped into external criteria. The second group of criteria are internal factors, i.e. factors associated with the company’s characteristics. The Table 5.1 summarizes criteria identified from interviews.

Table 5.1. Summary of the interview results

Category Factors

Ex

ternal

Geographical factors Number of countries Cultural differences

Location/proximity of different business units Legal factors Software support of localizations

Government regulations/ legal requirements

(27)

27

Inte

rnal

Strategic factors (Non-operational)

Company acquiring/ selling policy Organizational structure

Organizational culture

Authority of the headquarters

Operational factors Mutual flows between sites – informational and material

Commonality between sites Presence of joint processes

Production approach – ETO vs ATO / MTS Shared resources and capacities between sites Cost optimization

Process optimization 5.2 External criteria and multiple ERP instances

Experts emphasized that companies should strive to implement the single instance of ERP, because the single instance means that the data is available across all sites, and the company achieves such benefits as a lower inventory level, lower safety stocks, quicker reporting, lower ownership costs, easier upgrades, etc. Nevertheless, in some occasions a company is forced to implement multiple ERP instances despite losing advantages of the company-wide integrated data.

Factors that mainly influence splitting the system into several instances are external factors, and refer to the country characteristics, where a site resides. These are national differences and associated legal regulations. However, factors mentioned by experts with regards to national differences are associated more with a technical side of the ERP implementation rather than with cultural differences, which were emphasized by studies of Sheu et al. (2003; 2004). The reason for this might be that cultural issues associated with process and operations organization can be resolved through change management, thus splitting the system can be avoided, while technical problems are harder to overcome. Therefore, if a company operates in different regions or continents it can create technical limitations for the ERP system:

(28)

28 would you do it again’, they say no. I think the regional one is making our life a lot easier.” 6:1 (1:1)

Another issue, where challenges can arise and lead to implementation of multiple ERP instances, are legal requirements concerning financials:

“Another thing is localizations. When we implement an ERP, you have to take care of all local rules. So of course the Netherlands, and Germany and the States, these kind of countries are supported by ERP. If I now go to a larger country which is more difficult, Brazil, India, or go to a smaller country wherever in the world, they all have some local rules we have to fulfill. Then it's a question if I can do that in the single instance of my ERP… at this moment for some countries you still have to go for a separate instance to support all local regulations...” 4:31 (26:26)

Nevertheless, despite being a quite complex matter, finance and bookkeeping organization is perceived by consultants and experts as the easiest in the ERP implementation in terms of having a clear set of requirements and needs. It was noted by experts, that solutions for bookkeeping exist for different ERP system configurations, thus implementation of several ERP instances does not create obstacles for finance consolidation:

“Multi-site can be good. And let's say the easiest is always the financial thing. You say: ‘well, at least I want to have a financial systems in one software system that can be consolidated easily.’ That is very well overseen.” 2:138 (85:85)

Finance is one of the main performance indicators and having constant insights is essential for a company and stakeholders, and especially for stock listed companies, who require uniformity and high speed of reporting. Nevertheless, some warning is relevant, since the driver for the ERP implementation can be a company’s movement from privately-held to a public company, thus an organization can be much more focused on the financal side, neglecting importance of non-financial factors:

(29)

29 However, legal requirements might refer not only to finance, tax and trade regulation, but can be also regarding the data storage:

“Our other customers are in the defense industry. And there you have a lot of regulations, especially in the US, but also in other countries, about the data and where it should be. That means, that some data should be in the country where you are on your own computer, on your server. That means, when you have an Israeli company and an American company, both need to have their own server, and both their own instance. So that also limits the possibilities.” 4:30 (24:24)

In addition, during the selection process the further maintenance of the system should be considered. For example, upgrading a single global ERP instance is almost impossible without causing disruptions of operations at some location, while an upgrade of multiple instances also has its challenges:

“...the risk is there again that Europe and the US will deviate and have different implementations. It also means that if you do an upgrade you have to do it twice or three times. So that's the balance to be found. Because I think if you have more than 4 instances, you should be very careful. Because then you can get out of control. If you have 3 or 2, then you have a fair chance of keeping them aligned… that the balance I normally found, 2 or 3 maybe 4 different instances.” 6:2 (5:5)

Therefore, in some cases, when the company’s operations are widely dispersed across the globe, multiple ERP instances are unavoidable. Even though technology advances rapidly and overcomes geographical challenges, technical issues are still present and should not be neglected, since those directly influence the further use and maintenance of the system.

5.3 Internal criteria and configuration of modules

(30)

30 despite experts’ strong arguments against multiple ERP instances, some cases were mentioned that can lead to installation of a separate instance at the site.

As it was mentioned before, the ERP package selection should be viewed from two perspectives – finance and a supply chain. The impact of finance on the ERP set-up was discussed in the previous section, since it is highly influenced by legal regulations. From the supply chain and operations perspective, a structure of sites and a type of operations needs to be considered. Three ERP functional areas, which are usually implemented by manufacturing companies, are chosen for this analysis – Sales and Distribution, Procurement, and Production control and Planning. The centralization of all these modules would lead to implementation of the extended single instance of ERP, while decentralization of one of modules would mean implementation of the multi-site ERP.

Sales and Distribution

Operations in sales offices, distribution centers and warehouses are quite limited. Usually a sales office serves as a representing agent of a company in a certain area. Thus, there is a limited flow between a sales office and main operations consisting from a periodic data stream for financial results. The sales office serves a local market, develops its own customer relationships and pricing policies and can have its local warehouse or distribution center with minor operations such as packaging and labeling. Consequently, a sales office together with warehouses can act quite independently and even have its own system:

“If you only talk about production facility and multiple sale offices worldwide, then sales offices itself they only bare financial data and sales related data. So it might be easier to have an independent software running for your bookkeeping and have only very tiny integration with your central ERP system.” 1:4 (12:12)

In addition, other arguments supporting the implementation of a local system at foreign sales offices mentioned by respondents are better adaptability and flexibility, as well as easier system implementation.

(31)

31 “If you have a customer asking for a product which your company makes, but your production is not happening until next week and there is nothing in stock, but a customer really wants it now. If you can make an order from a different country, where they have it in stock, and it can reach your customer in two days, you are going to sell it from a different country, because it makes your customer much happier. So then you allow your sales person to define their own discounts, their own sets of rules, but they have special cases when they actually can go for co-selling from other countries… Or I want to sell something to my customer, but my production facility doesn't do that, but German production facility has capabilities next week, and I can send a request to them to produce this for me for the next week. So then you can say I am selling locally, but I can actually trigger the plant production across.” 6:18 (70:71)

This competitive strategy based on a quick delivery is also mentioned by Yen & Sheu (2004), however, in their study they observe low information sharing in companies focusing on a fast delivery, what limits the realization of this competitive priority. As it can be seen, cases, when a sales person can trigger the production or check stocks in warehouses in other regions, require an access to the data across other sites. That is hardly achievable with the separate system. Integrating warehouses into one system also not only contributes to better satisfaction of customer’s needs but lowers the inventory holding cost as well:

“… if we choose two completely separated warehouses with no overall planning function, and, for example, we optimize Western Europe and Eastern Europe, then average stocks and safety stocks will be higher than when you see that as one logistic system… if we have more demand in Eastern Europe and less demand in Western Europe we can ship some goods from west to east and vice versa. Then you have savings in stocks and inventory.” 2:90 (16:16)

Besides the inventory holding cost, the central planning of distribution also means the optimization of inner logistics:

(32)

32 Taking these considerations into account, the setup of sales and distribution is influenced by the company’s competitive strategy, i.e. on which competitive priorities the company focuses to serve its customer. Second perspective is optimization of inventories and logistics, and it concerns the management of warehouses and distribution centers.

Procurement

In the procurement several factors play a role. First, establishing the centralized procurement can be motivated by gaining the bargaining power. Thus, if resources for majority of sites are similar or can be purchased from the same supplier, the ordering quantity might result in discounts for a company:

“But with the system move, they move it [procurement] centrally. Because even within a country you might have different plants ordering the same stuff without knowing it. And volume discounts are very important in this business... So also negotiations, the way to order stuff and understanding how ordering works, it makes a lot of sense to put it as a shared service and then the procurement goes to a central location, serving more of your factories and locations. So the trend is definitely there to centralize that.” 6:9 (18:20)

That is particularly a case for process industries, where bulk materials are purchased, as well as for companies with assembling facilities with standardized components, and having one central purchasing department results in mutual benefits for both a site and a whole organization. However, other factors can offset these benefits, and the setup of procurement also depends on the type of required resources and knowledge of purchasers:

“… it [centralization of procurement] has to do with in which country you are or from which country we have to purchase. Do we have local knowledge, is it really needed, yes or no? What kind of parts or resources do you have to purchase?” 1:54 (120:120)

For example, specific features of the resource can create an obstacle for central procurement: “If you are in the process industry, every part which you purchase, you have to use it at that day, because otherwise it expires, so you are not able to use it anymore. If you have certain characteristics of that purchasing part, then maybe local presence is very good.” 1:56 (124:124)

(33)

33 be shared with other sites, but it is hard to align the needs of the ETO site with other sites in terms of ordering and delivering requirements:

“…you have less volume [in ETO production], but you still might be ordering huge quantities. Because ETO would be a ship building industry. They order huge amount of steel. Now, if I am taking a certain supplier of steel, I might be making at this location stuff for water pumps also made from certain amount of steel, which could be taken from the same supplier. So from the production it is totally different, but from the supplier perspective there might be some overlap, not very strong, but there might be some overlap. But it doesn’t hurt to have a separate system, because you have different suppliers and you are allowed to order differently, you have different negotiations, different contracts with that customer or vendor, because, of course, ordering some steel, or ordering steel for a ship is totally different domain. In ETO you want to have one large order from many different components delivered at certain dates. Otherwise you end up with a whole bunch of steel at your shipyard that you cannot use next couple of months. So you want your vendor to be selling it to you and shipping it to you couple of days before you want to use it.” 6:10 (22:24)

In overall, acquiring resources on the central level can require much more coordination efforts, while purchasing at the site might reduce the associated risk and contingencies. However, some experts remarked that if the company can afford an expensive ERP package and purchasers have necessary competences regarding the resource, procurement can be arranged effectively on the central level, despite other factors.

Production Control and Planning

Manufacturing plants encompass much more complex operations than other sites. However, the main driver for the complex integration of all company’s manufacturing operations is process and cost optimization:

(34)

34 The setup of the ERP system and production planning heavily depends on the manufacturing approach of a site and the intensity of mutual flows between sites. These findings are in consistency with conclusions of Gattiker & Goodhue (2005) and Aslan et al. (2015). For companies with sites, that have relatively high commonality in terms of the end-product manufactured that is usually observed in MTS, there is no incentive to implement multiple ERPs:

“I cannot think of any other reasons from efficiency perspective to separate those [sites], because modern ERP systems allow you to take into account specifics of production plants. If you have a much more modern production plant, and it's turned on much more quickly, the ERP system knows this, and will schedule production accordingly. Because if you change your production to a different product it's much cheaper for you than if I will do it. And you will change much quicker than I would. So that means that you still optimize per production location… Because I think centralized production planning, centralized procurement in the long run will be the optimal solution. Centralized from the system perspective. You might still have a production planner that works for you only locally, but he works in the central system and the system will say ‘yes, you can produce it locally, but look at this - if you produce in other country it can be done much quicker or cheaper’… But if you don't have system links or not a single system, you don't know how you can optimize, and you do it locally and by definition you lose money.” 6:20 (66:66)

In the ATO production approach coordination of flows between sites on the higher level balances the production, and consequently contributes to shorter throughput time and lead time. This can be achieved by the central planning function:

“If, for example, there is one production facility producing end products and all other ones are producing parts in order to be assembled for end products, then you have a supply chain consisting from all different companies. It makes a lot of sense to make sure you have one overall planning.” 1:48 (100:100)

(35)

35 “If it's ETO, then the chances that you have more or less an independent setup are much bigger than it is in MTS. Because in MTS there is an easier production process and it is easier to plan. And on another side, in ETO normally the number of end products is much lower than in MTS.” 1:46 (92:92)

On another side, if there are limited mutual deliveries, but there is commonality between ETO sites in terms of production processes and equipment, there might be a reason for implementing a complex multi-site ERP with regards to planning:

“For example, a construction industry or shipyards, and they have a yard in Finland and a yard in France. Sometimes there are shared resources, critical capacities that can be used at both places… if you can claim these critical resources over companies and over sites, and if you want to have it in consistence, you need a really complex software system that can support these kinds of things.” 2:100 (19:19)

Nevertheless, if a company has both ETO and MTS or ATO facilities, it is extremely hard to configure an ERP package that would support all these production approaches. If the commonality between sites is extremely low the success of integrating all sites depends on capabilities of the software. Additionally, another expert noted that even if sites are at the same location that does not necessary mean that one system should be implemented, although it also a question of commonality between sites:

(36)

36 Gattiker & Goodhue (2005) note that high differentiation between sites limits the ERP related coordination and efficiency improvements. Thus, a company can go for multiple ERP package implementation, and that would also simplify the overall system integration, because attempts to integrate two completely different sites into one solution would require much more efforts. 5.4 Internal non-operational criteria

The decision regarding the system choice is also complicated by other non – operational criteria, such as the general strategic vision of the company and the organizational culture. For instance, some organizations employ an aggressive acquiring policy, and the organization should take into account the likelihood of changing the company’s portfolio in the long-term, since it also affects an ERP system configuration, and raises the question of how adaptive and agile the system should be. Having a completely integrated central ERP or a multi-site ERP with a high level of function centralization forms an obstacle for an effective acquisition of new or selling of existing business units or sites, but at the same time it also reduces possibilities of getting more gains from the acquisition:

“If you want to buy and sell your business units often, then it is much easier to do, when processes are not that much integrated and when you have implementations more isolated. So if you now have a single data base, a single instance, it is much more difficult to shell apart it, than there are different instances and different ERPs. So that is also a way of how you look at your business. If you want to buy another company, you would like it to be integrated as soon as possible into your own processes and want to benefit from commonalities and have one sales organization, who is going to sell from both companies, and in that case it is better to have a single instance. So what companies need to do is that they have to look at all the commonalities they have in the processes, look at the joint processes. And based on that you have to decide if it's possible for you to go for multiple instances or if you want to combine it into a single instance.” 4:24 (22:22)

Another issue that should not be underestimated is the organizational conflict. The parent – site conflict is mentioned a lot in the literature (Morton & Hu 2008; Kirsch & Haney 2006), and implementation of a single instance ERP is associated with cultural and ‘political’ factors that emerge from the parent and sites clashing cultures:

(37)

37 instance, but I would like to have the server in the room just besides me. Then I can control it. Then I know, if something goes wrong, I have my own engineer who takes care of that. And that's possible. But that limits the possibilities for implementation, then you have to go for multiple instances.” 4:27 (23:23)

“Or if you have multiple bold companies, who also have different company cultures you really need to think about that as well, because the more cultural differences there are, the less easy it is to have central ERP implementation.” 1:51 (104:104)

Finding the right balance of local specifics and standardization is crucial for the implementation, what is also emphasized in the literature (Umble et al. 2003; van Fenema et al. 2007). While the single-instance single-site ERP implementation can fail due to the low system acceptance of a site, implementation of the complex multi-site ERP might require extremely high efforts in the project management and the company’s capabilities to integrate the system successfully can be overestimated:

“The less costly ERP systems very often have a lot of choices on the top level and you become a bit frustrated at a lower level, on the operational level, because you have to deal with conditions, which are not optimal for your local site. And that's also a reason why most of multi-site implementations fail completely, if you look at the level of satisfaction. At the end of multi-site implementation it is extremely low, because during package selection ... they make a really long list of requirements what they want. And at the end it means, that they have to go for a really complex system, and the implementation is also complex. And somewhere during the journey you will lose control. And very often there are points of intervention of the top management after one year or a half year after the implementation, saying, for example, ‘now we stop local specifics, and we go for one thing, and we all do this.’ So we do ‘materials used’ and the way they are booked everywhere the same... and then you have frustration again in many sites. So the level of satisfaction is extremely low in multi-site implementation. Because it is so complex to manage.” 2:130 (68:68)

While criteria concerning operations of a company directly influence the ERP system configuration, the non-operational factors, such as cultural differences, possible expansions of the organization and managerial capabilities in implementation, have an indirect impact on the choice of a system. Therefore, those are harder to anticipate during the ERP selection process and accordingly outcomes can be easily overlooked.

(38)

38

6. Developing the decision tables

The decision tables 6.1-6.4 summarize the findings from interviews and the literature review. The first table in this section presents the choice of the system, while others three help to define the setup of the previously discussed ERP system modules – Procurement, Sales and Distribution, and Production Control and Planning. The conditions refer to criteria that are either true (T) or false (F). A dash symbol (-) represents that the condition is irrelevant, while T/F means that the condition does not have a strong influence in particular settings and can take either of values. The combination of certain criteria forms a decision expressed as an action. Since this research is the first step in developing the solution design, the actions form recommendations rather than absolute rules.

Defining the setup of the ERP system

Table 6.1. Defining the setup of the ERP system

Suggestions

Conditions: 1 2 3 4 5 6 7 8 9

Organization is highly distributed in different world regions/continents

T T T T T T - - -

Organization resides within one country - - - T T T/F Legal requirements of all countries are

supported by software

T F T/F T F T/F - - -

Absence of technical limitations (stable infrastructure, language support, ease of upgrades, etc.)

T T/F F T T/F F - - -

Commonality between sites is high (processes/inputs/outputs)

T T T F F F T F F

Mutual flows between sites (informational/material)

T/F T/F T/F T T T T/F T F Actions:

Multiple ERP instances x x x x x

Single instance single-site ERP x x x x

Multi-site ERP x x x x

(39)

39 Suggestion 2: If the system cannot support legal requirements, multiple regional ERP packages or a separate package in a specific country is unavoidable. Within the region, if commonality between sites is high, a single instance single-site ERP package is chosen.

Suggestion 3: Similar to the suggestion 2, but with the respect to technical limitations.

Suggestion 4: Similar to the suggestion 1, but differentiation between sites is high. The presence of mutual flows requires coordination and the multi-site ERP package is recommended. Suggestion 5: The legal requirements cause the system splitting, and one site or the region has its own solution. If the regional ERP is implemented, within the region the multi-site ERP is recommended due to the presence of mutual flows and differentiation among sites.

Suggestion 6: Similar to the suggestion 5, but with the respect to technical limitations.

Suggestion 7: High commonality between sites and a location of sites within one country leads to a single instance of ERP implementation throughout the organization.

Suggestion 8: If an organization is located in one country and its sites have low commonality, but deliver to each other a multi-site package can be chosen.

Suggestion 9: If there is no commonality between sites, as well as no mutual flows, sites can implement independent ERP packages, with a separate software for financial consolidation.

Defining the setup of procurement

If a company implements the multi-site ERP, the next step is to define the configuration of modules. The Table 6.2 summarizes the main criteria that influence the setup of the procurement module.

Table 6.2. Defining the setup of procurement

Suggestions

Conditions: 1 2 3 4 5

Supplier overlap between sites T/F T F - -

Common resources among sites T F F - -

Highly critical resources F F - T -

Need for local knowledge F F - - T

Actions:

Central procurement x x

(40)

40 Suggestion 1: If sites are purchasing similar resources that can be acquired from the same supplier, the central procurement leads to better purchasing performance.

Suggestion 2: If sites require different input, but these resources can be purchased from same suppliers, organizing the central procurement can help to gain power over a supplier.

Suggestion 3: If there is no overlap in suppliers or input materials between sites, then centralizing procurement in the system does not create any additional benefits for the company. Suggestion 4: If there are certain characteristics of resources, such as short on-shelf life, it is harder to manage the inventory of such materials and tighter control is required to minimize the associated risk of inappropriate material quality, thus decentralization and a local presence of purchasers can limit these risks.

Suggestion 5: If the specific knowledge is needed to purchase right materials, setting the procurement centrally can create a risk of not meeting needs and demands of a plant, thus procurement decentralization might be required.

Defining the setup of sales and distribution

The Table 6.3 presents possible decision for the organization of sales and distribution. It was noted by experts that sales offices usually are relatively independent from the rest of the organization, thus the high integration of sales means that a sales office is able to access data across the company’s facilitates, and a multi-site ERP is required. Meanwhile the minimal integration of sales offices applies that the office is connected to the rest of organization only by financials and sales related data, thus an independent software can be implemented. The central sales organization means establishing one sales service within the single instance single-site or multi-single-site ERP, depending on the influence of other factors.

Suggestion 1: If the product is designed for a certain market there is no need for higher integration of sales offices.

(41)

41 Suggestion 3: Here a competitive priority of an organization is a possibility to customize a product and produce according to a customer order, thus to achieve such flexibility, a sales office has to be integrated into a multi-site system.

Suggestion 4: If an organization provides the global support and services that is usually connected with sales offices it requires an effective information sharing system and standardization of the service, thus a sales office needs to be part of a multi-site system. Suggestion 5: If the company produces in a low volume, centralizing sales and distribution is the optimal solution.

Table 6.3. Defining the setup of sales and distribution

Suggestions

Conditions: 1 2 3 4 5

Each sale office serves a specific market T - - - -

Product designed for a specific market T - - F -

Global product support and service F T/F T/F T T/F

Customization and volume flexibility - T/F T T/F -

Fast delivery - T T/F T/F T/F

Low volume production - T/F T/F T/F T

Actions:

High integration x x x

Minimal integration x

Central sales organization x

Defining the setup of production control and planning

The main decision criteria regarding the production control and planning is depicted in the Table 6.4.

Table 6.4. Defining the setup of production control and planning

Suggestions

Conditions: 1 2 3 4

High commonality between sites T F F F

Mutual flows between facilities T/F T F T/F

Low volume production with highly customized products F F - T Actions:

Central planning x x

(42)

42 Suggestion 1: If production facilities are similar, establishing central planning creates such benefits as production optimization with more effective machinery set-ups and consistency in processes among sites.

Suggestion 2: If there is high differentiation between sites, but significant mutual flows are present, the establishment of central planning depends on capabilities of the software to support different facilities. If the variance between sites’ processes is not so extreme and the ERP software can be configured to support needs of each site, the central production control and planning can help to balance production and improve coordination of mutual flows.

Suggestion 3: If there are no mutual dependencies between sites as well as produced products or processes differ among sites, each site can perform its production planning and control independently.

Suggestion 4: If there is no commonality between sites and production involves high cooperation with a customer to manufacture customer specific products, similar to ETO manufacturing environment, or in some cases ATO, the production planning can be performed at a site.

(43)

43

7. Discussion

This chapter discusses the findings regarding the initial research conceptualization. Next, theoretical and practical contributions of this paper are reviewed, followed by reflection on limitations and suggestions for further research.

7.1 Discussion

As the background of the research the work of Clemmons & Simon (2001) was used. However, it was noticed that the concept of control did not appear much during interviews, and was not mentioned as a reason of choosing a certain system configuration. Hence, this construct cannot be an absolute determinant for the system choice. Interviewees much more often were mentioning the need for coordination and discussing it from the logistical perspective, while control was mostly associated with the financials. The explanation of this can be that, the control itself is not the driver or the aim towards centralizing certain ERP system modules. Incentives for centralizing process management can be leveraging volume, process and cost optimization, etc. In the multi-site enterprises the question is rather who should have a control over a particular process, thus the decentralization of the system does not necessary mean a low level of control, as it can be distributed among sites. For example, Bendoly & Jacobs (2006) mention the concentration of centralized versus decentralized control, while Clemmons & Simon (2001) distinguish between low and high levels of control. Perhaps, this contradiction appears due to different views on the mechanisms of control and better conceptualization of this construct in the context of ERP is needed. This study appeared to be more in line with the control conceptualization of Bendoly & Jacobs (2006).

In contrast, the construct of coordination can define the setup of the ERP system, and is associated with such criteria as the intensity of mutual flows between sites, the level of commonality of sites in terms of products and/or processes, and a production approach. Thus, these factors define the required level of coordination, and subsequently the level of integration of a site into one system. Clemmons & Simon (2001) associate the need of coordination mechanisms with the level of interdependence and differentiation between sites, what was also observed in this study.

ERP package selection and implementation in multi-site companies Master Thesis