The in uence of Critical Success Factors on the deployment speed of Cloud ERP implementations

The influence of Critical Success Factors

on the deployment speed of Cloud ERP

implementations

Master Thesis - 10-08-2016

Eelco Lambers

University of Amsterdam

Business Information Systems

Supervisor/Examiner 1: Dick Heinhuis - UvA

Supervisor 2:

Sebastiaan van der Meulen - KPMG Examiner 2:

Contents

3.2 ERP Implementation and Critical Success Factors . . . 9

3.3 Cloud ERP Implementation . . . 12

3.4 Project Success . . . 14 3.5 Deployment Speed . . . 14 3.6 Framework . . . 15 4 Methodology 19 4.1 Interviews . . . 19 4.2 Survey . . . 20 5 Results 20 5.1 Interview Results . . . 20 5.1.1 Hypotheses . . . 27 5.2 Survey Results . . . 28 6 Conclusions 32 7 Discussion and Limitations 34 References 37 A Interview Protocol 40 A.1 Introduction . . . 40

A.2 Critical Success Factors . . . 40

A.3 Closing . . . 41

B Survey Protocol 41 C Survey Statistics 43 C.1 Survey Distribution . . . 43

C.2 Mean and Standard Deviation . . . 43

C.3 Chronbach’s Alpha . . . 44

C.4 Skewness and Kurtosis . . . 44

C.5 One Sample T-Test . . . 45

1

Abstract

With the upcoming use of Cloud technology, there is also the chance to use Cloud ERP applications. One of the benefits of Cloud ERP is the cost re-duction because of the outsourcing of soft- and hardware. It is also claimed that Cloud ERP implementations can be done faster than on-premise ERP im-plementations. This research will examine if this is the case and if so, how Critical Success Factors have an influence on the implementation speed. Inter-views with ERP experts showed that various Critical Success Factors have a positive influence on the deployment speed, mainly due to the outsourcing of hardware, the approach of formalizing requirements and the use of standard-ization in Cloud solutions. Statistical analysis of a survey conducted among 53 ERP/Cloud experts supported the qualitative findings. Success Factors such as IT-Infrastructure, Implementation and Top Management Support have a pos-itive influence, while factors such as Change Management, Business Process Redesign, Training and Data Migration also have a slightly positive influence. It is therefore expected that the deployment speed of Cloud ERP implemen-tations is higher than with on-premise ERP implemenimplemen-tations, although there can be some discussion whether this is purely because of Cloud technology or because of the approach used with Cloud projects.

2

Introduction

The last years more and more organizations chose the path to use cloud ap-plications. Cloud applications such as Office 365 and Google Drive are being used in organizations on a daily basis. With this shift to using cloud solutions and organizations getting familiar with “the cloud”, a new opportunity arises in using Cloud ERP systems. According to Sowan and Tahboub (2015), an ERP system is an information system with the goal to integrate all business processes and functions into a central database and provides an efficient way of management of business resources such as finance, human resource and pro-duction. An ERP system is essential for large and medium-sized enterprises in order to store data centrally and integrate departments with each other. An ERP system is installed on-premise at the organization. Cloud systems however are not installed locally at the organization, but reside at a shared server. This server is accessible via internet and makes it therefore possible to access business data and the system from everywhere. According to Schubert and Adisa (2011) services provided by cloud computing originate from two important technolo-gies: multi-tenancy, the shared use of software with individual storage for every company; and virtualization, physically sharing hardware capacity by running virtual machines on shared servers. Based on these two principles, three service models can be defined for cloud computing: Platform as a Service (PaaS), In-frastructure as a Service (IaaS) and Software as a Service (SaaS). PaaS is when an organization uses a predefined software environment. With Infrastructure as a Service however, an organization uses cloud solutions for processor capacity

Figure 1: Different Cloud types

or storage. Lastly, Software as a Service is a combination of PaaS and IaaS. With SaaS both the software and the hardware are from the same provider and are cloud-based (Raihana, 2012).

Besides the three different service models, there are also different types of clouds. A cloud can be private or public, or a combination of both named a hybrid cloud. Private clouds are specified to a certain organization and only accessible for them. With a public cloud, the cloud is shared with everyone and the application isn’t adapted to the organization. A hybrid cloud is a combination of those two types where some components are private and some are public (Elragal & Kommos, 2012), as shown in Figure 1.

There are various of benefits for an organization to choose for a Cloud ERP system or to switch from on-premise to cloud. One of the biggest benefits is cost reduction (Kiadehi & Mohammadi, 2012). The initial costs of a traditional ERP system are relatively large. Cloud systems are on a licensing basis, meaning that organizations have to buy a license and pay a monthly or yearly fee. This results in organizations not having to first make a large investment for the implementation. This means that the CAPEX, the Capital Expenses decrease, while the OPEX, the operational expenses increase.

A claimed benefit about Cloud ERP is that the implementation of the system is faster than the implementation of a traditional ERP. NetSuite (2016) claims that the deployment speed, the time it takes to deploy or implement a system, of a Cloud implementation is two times faster than a traditional ERP implemen-tation. Various other sites also claim that implementing takes less time (Lenart, 2011), (Cloudtweaks, 2012), (Castel, 2014), (Garrehy, 2015), (Muhleman, Kim, Homan, & Breese-Vitelli, 2012). This can be interesting since The ERP Report of 2015 of Solutions (2015) states that only 22% of the ERP projects were on schedule and just 3% of the projects finished under schedule. That leaves 75% of the projects overdue. An improved deployment speed of an ERP project,

the time it takes to deploy a system, can be beneficial for choosing that certain ERP system. However, the deployment speed is dependant on various of factors, which will be discussed during this research.

This research will identify which factors have an influence on the deploy-ment speed of ERP projects. After these factors have been identified, it will be researched how these factors differ between on-premise and Cloud ERP im-plementations and what their influence is on the deployment speed of Cloud ERP implementations specifically. Also, during this research the focus will be on the implementation of the ERP and not the entire ERP project. Later on, this difference will be explained.

2.1

Research Question

The research question of this paper is as follows:

What is the difference in Critical Success Factors between an on-premise ERP implementation and a Cloud ERP implementation and how do these Critical Success Factors influence the deployment speed of a Cloud ERP implementation?

In order to answer this research question, first some other questions need to be answered. The sub-questions that arise from this main research question will be pointed out briefly below:

SQ1: What are the Critical Success Factors of an ERP implementation? The term Critical Success Factor will be explained and the factors that are important in the implementation of an ERP will be discussed. Since the focus of this research is on the deployment speed of an Cloud ERP implementation specifically, only Success Factors that relate to the deployment speed will be discussed.

SQ2: What are the phases of a Cloud ERP implementation?

Which phases does a Cloud ERP implementation have and how do they relate to the phases of an on-premise ERP implementation? Based on literature, some implementation frameworks will be discussed.

SQ3: What is Project Success?

Deployment Speed (or Time) is a factor of Project Success. Other factors are Quality and Cost. Although this research focuses on the Deployment Speed and not Project Success in total, it is important to know what Project Success actually is and how the factors of Project Success influence each other.

SQ4: Which factors influence the deployment speed of a Cloud ERP im-plementation?

Next, the question of which factors influence the deployment speed of a Cloud ERP implementation specifically needs to be answered. The relation of the Critical Success Factors, that are discussed earlier for an on-premise ERP, with Cloud ERP will be discussed.

SQ5: What is the influence of those factors on the deployment speed of a Cloud ERP implementation?

At last, the influence of the Critical Success Factors will be examined. Both qualitative and quantitative research will be used to answer this question and with it also the main research question.

2.2

Relevance

There is a lot written about ERP implementations but only little is known about Cloud ERP implementations, mainly because Cloud ERP is relatively new. Although, as mentioned before, various sources claim that implementing a Cloud ERP takes less time than implementing an on-premise ERP, this has never scientifically been proven yet. Therefore, this research will examine whether there is a difference in deployment time between an on-premise and Cloud ERP and if so, an explanation for the difference will be given. Also, the current literature lacks a well-founded comparison between Cloud and traditional ERP implementations and their frameworks.

Besides that there is a gap in the scientific literature, there is also a practical relevance. It is assumed that various Critical Success Factors positively influence the deployment speed, but how, why and if this is really the case is not sure. This research can highlight the main differences between on-premise and Cloud ERP implementations (related to the deployment speed) and could give organizations a list of factors to consider when implementing a Cloud ERP system.

3

Literature Review

During the literature review, the framework presented in Figure 2 will be used to answer the main research question and it’s corresponding sub-questions.

Figure 2: Literature Review Framework

The differences between the implementation phases, and the time those phases take, of a Cloud ERP and a traditional ERP will be examined dur-ing this literature review. First, main concepts and terms will be explained using the literature. Via a funnel, with every step the research will get more concrete and the scope narrows down. First, the ERP Life Cycle, presented in the literature, will be discussed. Since this research will focus on the actual implementation, the focus then will be on the implementation related phases of the Life Cycle.

Essential for a successful ERP implementation are the Critical Success Fac-tors. The factors that have influence on the project success and especially the deployment speed will be pointed out. In the end, these CSF’s will be used to review how they relate to Cloud ERP implementations specifically. The dis-cussed CSF’s will be held against the Cloud ERP implementation steps and the relation to the deployment speed will be pointed out.

The articles used in the literature review where found based on various key search-terms: ERP implementation, Cloud ERP, ERP, Critical Success Factors, Critical Success Factors of ERP, ERP implementation speed. Some variations on these search-terms were used. In a limited amount of cases white-papers of organizations were used, but those were supported with scientific literature. The white-papers were used to show practical relevance or to emphasize assumptions made by organizations.

At the end of the literature review a new framework will be presented. This framework will then be explained and discussed. Also the propositions made based on this framework will be explained. Further in the research these propo-sitions will be tested.

3.1

ERP Life Cycle

The deployment of an ERP system usually follows a phase-based framework. Markus and Tanis (2000) propose a life cycle of an ERP implementation which

consists of four phases: Project Chartering, Project, Shakedown and Onward and Upward, which is shown in Figure 3.

Figure 3: ERP Life Cycle

In the Project Chartering phase the requirements are set and the decisions are made on what system to choose or build. During the Project phase the sys-tem is being built and tested. In the third phase, the Shakedown, the syssys-tem is up and running and the last bugs are being dealt with. Lastly, the Onward and Upward phase, is where maintenance and upgrading is being done. Shaul and Tauber (2012) present somewhat similar phases named: planning, implemen-tation, stabilization and enhancement. These phases are based on the article of Maheshwari, Kumar, and Kumar (2010). In the planning phase the require-ments are being defined and an outline of the project time will be made. A time-frame of what needs to be done at which moment will be constructed. Af-ter this phase, the actual implementation takes place with configuring the ERP system. The next phase, the stabilization phase, is where the system actually goes live, minor bugs will be eliminated and at the end the system is stable. The enhancement phase is where continuous improvements are being done and updates are performed.

Another, slightly different ERP Life Cycle is presented by Bancroft, Seip, and Sprengel (1998). According to Bancroft et al. (1998), the process consists of 5 phases: focus, as-is, to-be, construction and testing, and implementation. The first phase, the focus phase, is similar to the project planning phase of Markus and Tanis (2000). During this phase, the planning of the project and the selection of the project team is being done. In the as-is phase, the organi-zation’s current business processes are being analyzed and modeled. The to-be phase then shows how the business processes will look like after the implemen-tation of the ERP systems. The last two phases, construction and testing and implementation are similar again to the models of Markus and Tanis (2000) and

Maheshwari et al. (2010). During these phases a stable, live version of the ERP system is implemented and user support is given. The difference between the model of Markus and Tanis (2000) and Bancroft et al. (1998) is that the latter focuses on the current and future business processes and gives more attention to the fact that some business process redesigning is necessary to adapt the organization to the system.

Although there are also many other life cycles (Dantes & Hasibuan, 2011), the main concept of all of them remains the same. There is a planning phase, an actual implementation phase and lastly a maintaining/stabilizing phase where the ERP system is solid and runs without errors. Motwani, Subramanian, and Gopalakrishna (2005) simplified the implementation framework to a three stage framework: Pre-Implementation, Implementation and Post-Implementation. This resembles what is stated before, that there are three major phases that are sim-ilar in every different framework or can be found in every framework. Dantes and Hasibuan (2011) presents an overview of all the different life cycles in the literature, ordered in one table and categorized on the three stages presented by Motwani et al. (2005). The resulting table of Dantes and Hasibuan (2011) is shown in Figure 4. During this research, the model of Markus and Tanis (2000) will generally be used. Furthermore, since this research wants to discover which CSF’s have an influence on the deployment speed of Cloud ERP implementa-tions, the focus of this literature review will be on the actual implementation and the phases related to the implementation.

3.2

ERP Implementation and Critical Success Factors

The project chartering and the project phase are the phases within the ERP Life Cycle of Markus and Tanis (2000) where the implementation of an ERP system takes place. The project chartering phase is where the requirements for the actual implementation are being set and in the project phase the user gets acquainted with the software and the ERP system is rolled out. Generally also an implementation framework is used for the actual implementation. In this way, implementing the ERP system is done in a structured way and the schedule and tasks of the organization are clear. Implementing an ERP system isn’t simply rolling out the system. It also requires some data migration for example from the current used system. In order to do this in a structured way, implementing frameworks are being used. An example of an implementation framework is shown in Figure 5 which shows ASAP, a framework for SAP implementations.

Figure 5: ASAP, an implementation framework for SAP

The success of an implementation not only depends on the chosen frame-work or method to implement the ERP system, it also depends on various Critical Success Factors (CSF). Based on the literature reviewed, the most im-portant success factors related to the implementation of an ERP system will be discussed. Of course there are also other CSF’s for an ERP project, but the factors that not directly relate purely to the implementation of the ERP, fall out of scope for this research and will not be mentioned. In Table 1, the Critical Success Factors that will be discussed are shown. In the second column, all the articles that discuss this success factor are mentioned. The CSF’s that will be researched have been chosen since they were repeatedly discussed in most of the literature and they are relevant to the actual implementation of an ERP system. Below, all the CSF’s mentioned in Table 1 will be explained shortly.

Critical Success Factor Literature

Top Management Support (Abu-Shanab et al., 2015), (Prasad et al., 2006), (Finney & Corbett, 2007), (Holland & Light, 1999), (Law & Ngai, 2013), (Motwani et al., 2005), (Murray & Coffin, 2001), (Shaul & Tauber, 2012), (Soliman, 2013), (Somers & Nelson, 2001), (Sowan & Tahboub, 2015)

Implementation Strategy (Estevas & Pastor, 2006), (Finney & Corbett, 2007), (Holland & Light, 1999), (Shaul & Tauber, 2012) Change Management (Abu-Shanab et al., 2015), (Prasad et al., 2006),

(Estevas & Pastor, 2006), (Finney & Corbett, 2007), (Motwani et al., 2005), (Shaul & Tauber, 2012), (Somers & Nelson, 2001), (Sowan & Tahboub, 2015) Business Process Redesign and

Software Configuration

(Abu-Shanab et al., 2015), (Prasad et al., 2006), (Estevas & Pastor, 2006), (Finney & Corbett, 2007), (Holland & Light, 1999), (Hong & Kim, 2002), (Shaul & Tauber, 2012), (Somers & Nelson, 2001)

IT-Infrastructure (Estevas & Pastor, 2006), (Finney & Corbett, 2007), (Holland & Light, 1999), (Kumar et al., 2004), (Shaul & Tauber, 2012), (Soliman, 2013)

Training (Abu-Shanab et al., 2015), (Estevas & Pastor, 2006), (Finney & Corbett, 2007), (Holland & Light, 1999), (Murray & Coffin, 2001), (Shaul & Tauber, 2012), (Somers & Nelson, 2001)

Data Migration Plan (Estevas & Pastor, 2006), (Finney & Corbett, 2007), (Holland & Light, 1999), (Hong & Kim, 2002), (Shaul & Tauber, 2012)

System Testing (Prasad et al., 2006), (Estevas & Pastor, 2006), (Finney & Corbett, 2007), (Shaul & Tauber, 2012)

Table 1: Critical Success Factors in the literature

Top management commitment and support

One of the most important CSF is top management commitment and support. It is essential for an implementation project to have committed leadership from the top management level. If top management explicitly identifies the project as a top priority, this will be followed by the rest of the organization. (Finney & Corbett, 2007), (Nah, Lau, & Kuang, 2001)

Implementation strategy

Another important aspect of the implementation is planning the implementation strategy and the time-frame (Estevas & Pastor, 2006). The implementation strategy includes the phasing in which an ERP system is being implemented, which approach is chosen and which model is being used (Holland & Light, 1999).

Change management

With implementing a new ERP system, changes will come. In order to cope with this correctly, a change management program needs to be prepared. One aspect of the needed change management is to create user acceptance for the

project and the needed changes. The desired result is that employees have a positive attitude towards the new ERP system (Parr & Shanks, 2000), (Nah et al., 2001), (Aladwani, 2001). According to Somers and Nelson (2001), it is estimated that half of ERP implementations fail because the efforts that are involved in change management are underestimated. Therefore it is vital for the project’s success to make change priority number one by employees.

Business Process Redesign and Software Configuration

In order to implement the new ERP system, it is common that several busi-ness processes need to change. To cope with these changes, Busibusi-ness Process Redesign (BPR) needs to be conducted (Al-Mashari & Zairi, 2000). The result of BPR is a complete overview and description of how the to-be processes are after the implementation. This overview will then be used to inform and train users and will be the guideline for the configuration of the ERP.

IT infrastructure

The IT infrastructure needs to be prepared for the ERP implementation. The IT infrastructure needs to handle with the ERP system and be aligned or fit with the current systems in place (Holland & Light, 1999). Besides that, system administrators need to be trained on how to maintain the new ERP hardware and software or new staff members need to be hired. This step is often forgotten. It is an utopia that no maintenance is needed after the implementation of an ERP.

Training

An important part of the ERP implementation is teaching employees how to work with the new system (Prasad et al., 2006). Employees need to be trained in using the ERP and new business processes need to be explained (Somers & Nelson, 2001). It can be the case that workflows are changed. Besides that, it is also useful for the employees to not only know all the possibilities but also the limitations of the software. (Murray & Coffin, 2001)

Data migration plan

The data from the old ERP system or others systems needs to be migrated to the new ERP system. In order to do so, there is a need for a solid migration plan. Not only the question how to migrate the data needs to be answered, a planning of when what data will be migrated will be made. This is necessary because then everyone knows when the data will be migrated and which data needs to be available at what moment. On top of that, when data is migrated, users must verify that the data is migrated successfully and that they accept the results (Estevas & Pastor, 2006).

System testing

One of the final stages of the actual implementation is testing. Before the ERP can definitely go-live and will be used, the system needs to be thoroughly tested

in order to be sure that it is stable. (Sengupta, 2015)

Next, Cloud ERP implementations will be discussed in order to understand more clearly which steps such an implementation follows and what the distinc-tion is with on-premise ERP implementadistinc-tions.

3.3

Cloud ERP Implementation

In this section both the Cloud framework ByDesign from SAP and the Cloud framework from Oracle will be discussed, since these are the two biggest ERP vendors (Columbus, 2014), (Solutions, 2015). For ERP implementations, the ASAP framework shown in Figure 5 is used by SAP. They have also developed a modified framework for Cloud ERP implementations, called ByD Go-Live, which is being used for implementations of their Cloud solution ByDesign. Fig-ure 6 shows the ByD Go-Live framework. The framework of Oracle will be discussed later.

Figure 6: ByD Go-Live framework from SAP for Cloud solution ByDesign During the Prepare phase, a common understanding of the project scope and goals are defined and also the way to achieve those goals. Other key activities during this phase are: creating a project planning, the review of business sce-narios and key business and process-related decisions. If needed, some business process redesign will be done.

The Fine-tune and Integrate & Extend phase consists of two parts. One is the fine-tune aspects, which focuses on system-related activities. Here the last needed configurations of the Cloud ERP system will be defined and done. The integrate & extend part of this phase is more concentrated on the migration of data and the alignment with other systems. Especially the migration of data during this phase is essential. The data not only needs to be migrated, but also needs to be tested and validated. At the end of this phase an agreement is made on the result of the data migration, in order to check if everyone is satisfied with the result of the migration.

The third phase, the Test phase, is where the entire system is tested, includ-ing the integration with other systems. Durinclud-ing this phase, test scenarios will run in order to see if the system can handle all these scenarios. At the end of this phase there is a so called Go-Live Readiness Acceptance Checkpoint where will be decided whether all tests were successful and the system is ready to start with the final phase, the Go-Live phase.

The final phase is the Go-Live phase. As the name already says, this is the phase where the system actually goes live and will be used in real-time. Test will now be production and all data will be migrated to this production environment. A final check will be made whether both the system and all the users are ready to use the Cloud ERP. If this is the case, the system will become live and all users will then only use the new Cloud ERP system.

Oracle (2015) proposes a 5 phases implementations plan as shown in Figure 7. During the Build a Plan phase, the requirements are being defined. On top of that the time-frame will be developed and the project team will be formed. Important is to have amongst the team, members of the finance organization who can assist in critical aspects of configuring the system, such as workflows and chart of accounts. Furthermore, other considerations during this phase are which data needs to be migrated and what training do users need to have and how will these users be trained within the time-frame.

Figure 7: Cloud ERP implementation framework from Oracle

The second phase, the System Design, the requirements from the Build a Plan phase will be translated into a design of the system. Instead of designing the ERP to the needs of the organization and adapting the system to the cur-rent business processes, Cloud ERP focuses on redesigning the business process in order to align with the Cloud ERP pre-defined processes. These business processes are best practices, which then leaves only 20% needing to be tailored to the organizations needs.

The migration of data from the current ERP system or other systems needs to be migrated and converted to the Cloud ERP solution. A data conversion strategy should include on defining which data is critical for reporting and com-pliance and thus needs to be migrated. Data conversion includes cleansing the data before they can be imported into the new Cloud ERP. After the data im-portation, some agreement will be made on if the data is migrated successfully. It is common that other systems of an organization need to be integrated with the new Cloud ERP system. Implementing a Cloud ERP doesn’t entirely remove complexity, but it does make the IT infrastructure less complex. In order to integrate successfully and take into account all the other current systems, process flows can be made for better understanding of the incoming and outgoing data flows of the new Cloud ERP to and from other systems.

Critical to the success of a (Cloud) ERP implementation project is the adop-tion of users of the ERP and training those users. In order to work efficiently, users need to be aware of all the system’s capabilities and benefits. User in-volvement during the implementation is therefore crucial. Including knowledge workers in the project team, helps validating that the system meets the client’s

business objectives and gives them a sense of system ownership. These users can then become experts en people other users can ask questions.

3.4

Project Success

The Critical Success Factors mentioned before have an influence on the Project Success. Project success is generally measured and defined as a triangle of criteria: time, cost and quality (Ebbesen & Hope, 2013), (Atkinson, 2005). The general concept of this so called Iron Triangle (Figure 8), or Golden Triangle, is that focusing on one criteria has effect on the others.

Figure 8: The Iron Triangle

For example, if the time a project takes increases this will also increase the total project costs, but could result in better quality of the project result. The connection with Critical Success Factors is that these factors have an influence on the criteria for project success (Ika, 2009). For example, the previously discussed CSF training has an influence on all the criteria for project success. Training users will increase the quality of the project, but could mean that it takes more time and that it will increase the project costs.

Although this research focuses on the deployment speed, i.e. the time criteria in the Iron Triangle, the other criteria of the triangle will also be influenced by this. However, for this research the influence of the discussed CSF’s on the other cost and quality dimension is out of scope. If needed, they will be mentioned shortly, but they will not be researched extensively.

3.5

Deployment Speed

The main question of this research is how CSF’s influence the deployment speed of Cloud ERP implementations. Generally, most ERP implementation projects are overdue (Parr & Shanks, 2000). As mentioned before, in 2015 over 75% of the projects were overdue (Solutions, 2015). In the paper of Elragal and Kommos

(2012) an on-premise ERP is compared with a Cloud ERP implementation. In this case, Elragal and Kommos (2012) compare SAP ECC 6.0 (an on-premise ERP system of SAP) and SAP ByDesign (a Cloud ERP system of SAP) and point out the differences between both systems regarding costs and time. In this sector the time factor will be discussed.

According to Elragal and Kommos (2012) the time reduction already starts before the implementation. Because there is no hardware needed with a Cloud ERP, the HR department of an organization does not need to hire additional IT administrators that will help with the technical implementation and the maintenance of the hardware after the implementation.

In the ASAP methodology, explained before, there is a Business Blueprint phase. This is the phase where the future business processes are being modeled based on the current processes. With a Cloud solution however, most of these processes are already defined, modeled and embedded in the system. KPMG for example claims that from experience, 80% of the process related configuration of the ERP is standard across companies (KPMG, 2014). What is left in this phase is users reviewing this template (Elragal & Kommos, 2012) and configuring the last 20% (KPMG, 2014). Elragal and Kommos (2012) also states that with the pre-configured templates, only some fine-tuning of the configuration is needed.

Also the data migration aspect takes less time with Cloud ERP. While the data migration at the on-premise ERP project generally starts just before the product goes live, with Cloud ERP data migration can start immediately.

Another factor that reduces the implementation time is the training time. With an on-premise ERP, two types of people need to be trained: users that will use the ERP daily and system administrators, that will monitor and maintain the ERP. The last one doesn’t have to be trained anymore since there is no hardware installed at the organization itself. Monitoring the performance of the servers will now be done by the supplier of the Cloud ERP. The daily users of the system still need to be trained in using the Cloud ERP. However, training can be started earlier since data migration is being done in an earlier stage. Most common is to migrate a dataset, so a part of all the data to be migrated, so tests can be based on that migration and training can already start. In this way, migrating all the data and training the users runs more simultaneously.

3.6

Framework

Based on the literature review above, a new framework will be represented. The resulting framework is shown in Figure 9

Figure 9: Lambers Framework of the relation between CSF’s of Cloud imple-mentation and Deployment Speed

The presented framework shows the relations between the 8 Critical Success Factors discussed earlier and the deployment speed. For every relation, there is a proposition that will be tested via qualitative research. Below we will shortly point out the propositions that will be tested. For all propositions counts that with influence on the deployment speed, the deployment speed of a Cloud ERP implementation is meant. Also, a positive influence on the deployment speed means that the deployment is faster and the implementation can be done faster. P1: Ensuring Top Management Support will not have a significant positive or negative influence on the deployment speed

Although Top Management Support is also vital for the success of a Cloud ERP implementation, no time can be won with this factor. Getting Top Management Support will be the same for a Cloud ERP implementation as for an on-premise ERP and thus have no influence on the deployment speed.

P2: Formalizing a good Implementation Strategy will not have a significant positive or negative influence on the deployment speed

There is no significant difference found in the literature for the Implementation Strategy between on-premise and Cloud ERP implementations. Some literature suggest a more agile methodology, but this doesn’t necessarily mean there is a positive (or negative) influence on the deployment speed.

P3: Change management will have a positive influence on the deployment speed

During a Cloud implementation a lot of aspects change for the users, which need to be dealt with correctly. Although you want to do this process as fast as possible, with almost everything in such a project, it is important to take the time with change management. Cloud implementations do follow a more agile methodology, during which an organization incrementally releases features over time. Using this method, users can be trained simultaneously and get famil-iarized with the software, while the system still is being implemented. Because of the agile methodology, business change will be introduced more gradually (Krigsman, 2012). This gives users more time to adapt to business and process changes. This will lead to more efficient change and thus a positive effect on the deployment speed.

P4: Business Process Redesign and Software Configuration will have a positive influence on the deployment speed

The Business Process Redesign and Software Configuration factor consists of two parts, which will be discussed separately. The first part, the Business Process Redesign, focuses on the fact that processes within the organization need to be adapted to the Cloud ERP. With on-premise solutions, the ERP was adapted to the processes of the organizations, but with Cloud ERP there is less customization possible, which leads to the fact that business processes need to change. This can take more time than with an on-premise implementation. The new processes need to be designed and implemented using change management. However, because Cloud ERP systems can be seen as more ’standard’, due to the lack of customization, business processes then follow a standard template. The result is that it doesn’t have to take a lot of time to build new business processes.

The second part is the Software Configuration. Because there is less cus-tomization possible and due to the fact that already 80% of the configuration is embedded in the Cloud solution (KPMG, 2014), time is saved with configuring the ERP system. Now only the last 20% needs to be adapted to fine-tuned to the specific IT-infrastructure and the business processes of the specific organization. This results in a significant time reduction.

P5: The factor IT-Infrastructure has a positive influence on the deploy-ment speed

With an on-premise ERP implementation the IT-infrastructure needs to be pre-pared for the upcoming new system and hardware needs to be installed. How-ever, with Cloud ERP, the hardware is stored at the vendor and no additional hardware needs to be bought and installed at the organization (Muhleman et al., 2012). This eliminates the time spent on installing and configuring hard-ware (Johansson, Alajbegovic, Alexopoulos, & Desalermos, 2014). What also reduces the implementation time is that, with all the hardware installed at the vendor, the maintenance of the software will also be done at the vendor’s side. Therefore, training for system administrators on how to maintain the ERP is not needed anymore.

P6: Training will have a positive influence on the deployment speed Also with a Cloud ERP systems users need to be trained. It is a new system and users need to know how to work with the new system and need to know the possible new work-flows. The time taken for training doesn’t necessarily will be reduced, however it can be done more efficiently. Because some imple-mentation steps can be done faster or in an earlier stage, there is more time available for training. Also, training can take place during other steps, such as the configuration of the system. Since Cloud implementations usually follow an agile methodology, with incrementally delivering working parts of the system, trainings can start from the moment a part of the system is done. In this way, there is no need to wait for the entire system to be configured, before starting with the training of users.

P7: Data migration plan will have a positive influence on the deployment speed

Still one of the most vital steps is the migration of data. Although the process itself can’t easily be shortened, it can be done more efficiently. The advantage of Cloud ERP is that data migration can be started from the beginning, instead of when the entire system is built. The requirements for how the data should be implemented are set before the project and the process of data migration can run simultaneously with other processes.

P8: System testing will not have a significant influence on the deployment speed

Testing the system is just as important for Cloud ERP as for on-premise ERP. However, there is no clear difference in the method of testing and the deployment speed found in the literature. Therefore it can be concluded that this factor doesn’t have a positive (or negative) influence on the deployment speed.

4

Methodology

The main research question of this thesis, what is the difference for Critical Success Factors between an on-premise ERP implementation and a Cloud ERP implementation and how do these Critical Success Factors influence the deploy-ment speed of a Cloud ERP impledeploy-mentation, consists of two parts. The first part, what is the difference for Critical Success Factors between an on-premise ERP implementation and a Cloud ERP implementation, will be answered using qualitative research. The second part of the question will also be researched us-ing qualitative research, but will be validated and supported with quantitative research.

The qualitative research will consist of interviews with ERP experts. During this interview, the presented framework will be evaluated and all the proposi-tions will be tested. The goal of the interview is to not only get yes or no answers to the propositions on whether the specific CSF has an influence on the deployment speed, but also is aimed to answer the question how a CSF influences the deployment speed.

The quantitative research will be in the form of a survey, set out to (Cloud) ERP experts. The results of the survey will be used to validate the interview results and support the qualitative results with a larger population.

4.1

Interviews

Qualitative research will be done by conducting semi-structured interviews with experts that have experience with Cloud ERP implementations and on-premise implementations. A semi-structured interview is chosen in order to get optimal results and give the flexibility to ask ad-hoc questions and additional questions related to the answers given.

The interview will follow a protocol, presented in Appendix A. The interview consists of three phases: Introduction, Critical Success Factors and Closure. During the introduction a short introduction of the research and the purpose of the interview will be given to the interviewee. After this, the interviewer will ask the interviewee to shortly introduce himself.

During the Critical Success Factors part of the interview, the focus will be on discussing CSF’s and answering the question what the influence is of those CSF on the deployment speed. First, to start with an open mind, the interviewee is asked to give Critical Success Factors that, according to him/her, influence the deployment speed of a Cloud ERP implementation compared to an on-premise ERP. Then the interviewer will match the discussed success factors with those of the framework presented in Figure 9. The factors that are not yet discussed will be asked.

The last part of the interview is the Closing, where is asked if the interviewee has questions for the interviewer. Also, the interviewee will be thanked for his/her co-operation.

Every interview is summarized and the answer to the main questions are given in Appendix D. All the interviews were recorded and are available on

request via e-mail.

The interviews will be held amongst Cloud/ERP experts and organizations that have implemented a Cloud ERP. In this way, both the implementation experts can be asked for their findings and the organizations can elaborate from their experience of their implementation project.

4.2

Survey

As discussed earlier, the qualitative interviews will be supported by a quan-titative research. The goal of this survey is to quantify the results from the interviews, to support the theory and qualitative results and test these on a bigger population. The survey consists of 18 questions. The first two questions are about the age of the respondent and the years of Cloud/ERP experience. For every Critical Success Factor there are two statements: There is a difference in the specific factor in Cloud ERP implementations compared to on-premise ERP and the specific factor in Cloud ERP implementations has a ... influence on the deployment speed. Both questions can be answered with a 7-point Lik-ert scale. For the full survey see Appendix B. Because of this 7-point LikLik-ert scale, survey results will not only show if a factor has a positive or negative influence on the deployment speed, but also how much of an influence (slightly positive/negative, positive/negative, strongly positive/negative).

The survey will be sent to ERP implementation experts. The network of KPMG will be used to find these experts. Besides that, communities and Linkedin groups are used. The targeted group is to find Cloud and ERP experts from various roles or organizations, for example: organizations that have done a Cloud ERP implementation, experts that have implemented a Cloud ERP system and vendors that offer Cloud ERP systems.

Survey’s will be sent via email to known experts and they will be asked to share the survey amongst their network. In this way a wide-spread network can be reached. Also, as mentioned before, the survey will be posted on the following Linkedin groups: ERP Community (90,000+ members) and ERP Im-plementation (571 members).

5

Results

5.1

Interview Results

In total 8 interviews were conducted over a period of 4 weeks. All interviewees were either ERP experts or involved in ERP implementation projects. The in-terviewees all preferred to conduct the interview anonymous. The ERP experts had a minimum of 5 years experience in the ERP field. All of the interviewees also had experience with Cloud implementations, but since this is a relatively new technology not a minimum of 5 years but 1 year. The length of the inter-views varied from 35 minutes up to an hour, but on average took 45 minutes.

The interviews were summarized and translated. Per question, the key-answers were then added to the Appendix D table. This table is then used to form the conclusions on the propositions made earlier. Conclusions were drawn (as much as possible) on the amount of interviewees that had the same answer. On top of that, strong and solid arguments of experts were also taken into consideration when making the conclusions of the propositions.

The conclusions of the propositions will lead to hypotheses that will be quantitatively tested using a survey, as explained before. The results of the survey will be discussed later.

P1: Ensuring Top Management Support will not have a significant positive or negative influence on the deployment speed

During the interviews most interviewees believed ensuring Top Management Support is easier. This is mainly because of two reasons: Cloud solutions are popular for businesses and since the implementation time is assumed to be shorter, Top Management is more interested in the project and there is more momentum. Also, one interviewee stated:

“Cloud could have more Top Management Support because there is more prestige to gain for them”.

On top of that, the buy-in of Top Management Support can also be realized in an earlier stage. Deciding to implement a Cloud ERP system, is automatically a decision for sticking to standard as much as possible, according to one of the interviewees. Since the power of Cloud ERP lies in the fact that it is more standardized and therefore easier and faster to implement, only sticking to that standard as an organization will result in benefiting of the advantages of Cloud ERP. So the buy-in from Top Management Support to stick to standard is when they decide to choose for a Cloud solution.

During the interviews some interviewees also believed ensuring Top Manage-ment Support will become more important. If there is no sufficient Top Man-agement Support, there will come change requests during the project which will move the solution further away from Cloud and which will result in an increase of the deployment time. One interviewee stated:

“If you stay close to the standard, you could say that the du-ration of the implementation will become shorter, which results in prolonged interest of top management.”

And another interviewee mentioned the importance of strong Top Manage-ment Support:

”Weak top management will eventually give multiple change re-quests needed to be filled in later. With strong project management or top management you can restrain this by just declining those re-quests. And the earlier you do that in the process, the better it is. And with Cloud, top management has already chosen to do it this way.”

It can be concluded that with Cloud implementations, Top Management Support can be ensured easier, but this doesn’t necessarily mean that the im-plementation time is less. Although strong Top Management will prevent the project not to be delayed, this doesn’t implicate that with Cloud ERP Top Management Support will have a more positive influence on the deployment speed, compared to on-premise.

P2: Formalizing a good implementation strategy and time-frame of the implementation will not have a significant positive or negative influence on the deployment speed

The part that changes the most for the implementation strategy is the formal-izing of the requirements. With on-premise implementations, requirements are being made based on what the organization wants to have in the new ERP system. With Cloud solutions however, since you want to stick to the standard, it comes more to validating. Validating means that the standard processes of the Cloud solution will be taken and will be held against the processes of the organization. A fit/gap-analysis is done to see where the Cloud ERP meets with the organization’s processes and where there is a gap. One interviewee stated:

”On the front-end you will get a different type of requirements definition. You will go less to users with the question how they want to have their system, but you will say: this is our package, it works in this way and where are the differences compared to your standard process and how can we incorporate them or work around them. I think it will be a more compelling story and a more compelling implementation strategy.”

As stated, the gap can then be resolved in different ways. For instance, the organization can change the business process to align with the Cloud solution or a workaround can be chosen.

”You make a fit/gap analysis and you say: these requirements match in the system and these are the gaps. Then you will consider if there is a business case for every gap to change the system or that you can still fill the gap with minor changes in methods.”

The Implementation Strategy has a positive influence on the deployment speed because of the different approach of defining the requirements.

”I think that the implementation time will become shorter and that is in particular the validating instead of designing.”

However, one interviewee also mentioned that this approach can also be done for on-premise implementations, but with Cloud implementations an organiza-tion is more forced to do so.

”As long as you stick to standard it will go faster, regardless of whether it’s Cloud or on-premise. But with Cloud you know that you have less options and therefore it cannot lead to delays.”

In conclusion, there is a big difference in the Implementation Strategy since there is more validation of whether the new solution fits with the current pro-cesses of the organization and how to fill in those gaps, instead of sitting to-gether with the organization and formalizing all the requirements according to the desires of the organization. This methodology has a positive influence on the deployment speed. Although this method could also be used for on-premise implementations, with Cloud an organization is more forced to do it this way, since there isn’t the possibility of customizing according to requirements of or-ganizations.

P3: Change management will have a positive influence on the deployment speed

The main difference between Cloud and on-premise is that Cloud systems most of the time have a newer, better looking and more intuitive interface. This helps with employees accepting the new system more easily. One interviewee acknowledged:

”I don’t think it really matters. Cloud does have the advantage compared to on-premise that the interfaces are better. More func-tionalities, better integration with outlook and that sort of things. Therefore your user experience is better. That makes an implemen-tation easier.”

Most interviewees don’t see a big difference between change management for on-premise implementations and Cloud implementations. There is a change and this still needs to be managed correctly.

”I think that the adoption of a system, taking people through the acceptance test, that will take 6 months on average and it’s best to do this in an earlier stage. There is a sort of minimum, even if you don’t have to configure anything or migrate any data, still you need to teach people the new process, and that will take 4-5 months.”

What interviewees do see is that, because of the more user-friendly interface, the buy-in for the changes can be reached more easily, which speeds up the process. Also, because of the approach of validating, as described earlier, change management also can be done more easily and faster. One interviewee showed the efficiency of Cloud:

”With on-premise, you then get 450 gaps and you want to reduce that. Then you will say no 200 times to the client and you still have 250 gaps. So you still have a lot of customization. With Cloud you

say: This is it, then the client will complain and you have 150 gaps. Then you say no to the client 100 times and keep 50 gaps. In the end you have said no to the client fewer times and you keep less customization. With Cloud this is easier to do because you start from less customization.”

Change management has a positive influence on the deployment speed. Al-though there is no big difference between Change Management on-premise and Cloud, with Cloud it can be done more easily. This is mainly because Cloud is hot and has an intuitive, good looking interface. Users will be more willing to adopt the new Cloud system and are more eager to learn the new system, since it has much user-friendly advantages.

P4: Business Process Redesign and Software Configuration will have a positive influence on the deployment speed

Cloud has a few advantages when you look at the Business Process Redesign factor. First of all, Cloud is more standard so there is a reference to work to. On top of that, the processes of Cloud ERP systems are based on best practices. This means that these processes are based on what is, according to various of other different organizations, the best way to design a process. Therefore, an organization could also become more efficient, since they will now work according to a best practice.

”An advantage of SaaS is that processes are best practices. So it can also be the case that you get some things out of it, which you never thought of yourself.”

Other interviewees also stated that some processes are standard for every organization and there is no competitive advantage. Therefore, it is better to move to the standard and choose for a Cloud system:

”All the standard processes, HR, Finance, Procurement, you can put in the Cloud. Their is no competitive advantage. I think that a lot of organizations are adapting their processes to what is possible in a SaaS solution.”

Because of the advantage of standardization and best practices, Business Process Redesign has a positive influence on the deployment speed. Because there is a standard to work to, redesigning of processes doesn’t take extra time. And when processes are aligned with the Cloud system, the implementation time will be reduced. In theory, you can also do this with on-premise by creating a standard, but with Cloud it is more accepted when an implementation party says: this is not possible. Whilst with on-premise there is always the solution to do customization.

P5: The factor IT-Infrastructure has a positive influence on the deploy-ment speed

A big advantage of Cloud concerning IT-Infrastructure is that no hardware needs to be installed. With a Cloud solution, the hardware is at the vendors side. So no hardware needs to be bought and configured. An interviewee answered the following on the question what the influence is of the factor IT-Infrastructure on the deployment speed:

”Absolute a positive influence. Absolutely. Of course there is a huge reduction of time. Where you first had to do a request for the hardware and after 2 months it was delivered and then configured. Now everyone with a credit-card can buy a Cloud solution. That is a huge time reduction.”

A new problem that arises is the integration of the Cloud ERP system with other (on-premise) systems of the organization. As one interviewee stated there is not simply ”One ERP” and organizations always have some other systems that need to be integrated. However, one interviewee also mentioned the advantage of standardization for integration of systems:

”I don’t know if it takes the same amount of time. I think that integration is somewhat easier, because that Cloud party has done it 500 times, so he can help you with it.”

Although there is some time needed for integration, all interviewees acknowl-edged that the factor IT-infrastructure has a strong positive influence on the deployment speed. Instead of ordering hardware, installing and configuring it, an organization can simply buy a license within 5 minutes. On top of that, scal-ing and sizscal-ing of the hardware needed is not an issue anymore. Cloud solutions are fully scalable to perfectly fit the needs of an organization, also in the future.

P6: Training will have a positive influence on the deployment speed There are two parts of the factor Training. First, there is the training of end-users. Most interviewees believed that the training of end-users is the same. Again, standardization however is an advantage and could have a positive in-fluence on the deployment speed. Since the solution is more standardized and there is less customization, exceptions and custom work don’t need to be in-cluded in the training. Trainings (and materials) can be more standardized and easier to learn.

”I think that the factor training will not change that much. You just need to learn. An advantage may be that, because it’s stan-dard, it is easier to learn and you don’t have all the exceptions and variants.”

The second part of the factor Training is the training of IT-personnel. Var-ious interviewees mentioned that there is a big difference compared to on-premise.

”There is a big difference. You don’t have to educate the IT ad-ministrators, because you don’t have technical administration. You only have functional administrators.”

Since the hardware is at the vendor’s side, technical support also will be done by the vendor. This means that system administrators don’t need to be educated to maintain the Cloud ERP. With not having to train and educate your IT-personnel, there will be a time reduction. One interviewee stated:

”It definitely saves time and it could also reduce costs.”

It can be concluded that Training has a positive influence on the deployment speed because there is no need to train IT-personnel on how to maintain the new ERP system. Also, because of the standardization, trainings can be done more easily and no exceptions or custom work need to be included in the training program, which saves time in preparation and makes the trainings simplified.

P7: Data migration plan will have a positive influence on the deployment speed

Most of the interviewees stated that there is practically no difference in Data Migration between on-premise implementations and Cloud. Data migration is always needed and it doesn’t matter whether the organization is migrating to an on-premise solution or a cloud solution.

”It is about the cleaning of the source-data, not the target-data. What you see is that old data isn’t migrated, you just leave it in the current system. For the rest it doesn’t matter where it goes to, you always have the same problems.”

An advantage could be that since Cloud is standardized, this reduces com-plexity.

”Cleaning of data are always needed, for on-premise and Cloud. But if I look at the amount of complexity and the amount of technical support, than Cloud is way more easier that on-premise.”

However, when asked to the influence on the deployment speed, most inter-viewees stated that the factor Data Migration doesn’t have a positive or negative influence.

”It is not the expectation that there will be time won. We don’t see an advantage on this factor in the deployment speed.”

Everything is dependant on the starting situation of the organization and although Cloud data structures are more simplified, the process that takes the most time remains the cleaning of data, which is the same as for on-premise implementations.

P8: System testing will not have a significant influence on the deployment speed

There is no big difference between testing the system with on-premise im-plementations and Cloud imim-plementations. Because of the standardization and less customization the test scripts can be standard and less exceptions need to be tested. However, what does need to be tested is the integration with other systems.

”None, if I look at financials and supply-chain. Maybe you could say that it will go a little bit faster, because it’s more standard. So you don’t have to test all kind of exceptions or custom work. The test-scripts are pre-defined. On the other hand, you do need to test the interfaces.”

There is no influence on the deployment speed because tests still need to done and there aren’t big differences between on-premise and Cloud. One interviewee stated:

”I don’t see that there will be much time won in system testing.” While another elaborated more:

”I don’t really believe that you can go to the customer and you can say, you have a standard process and these are your test scripts. You can do that for regression tests. But I think it’s related to user acceptance that the customer still will make it’s own test scripts.” The difference for System Testing is that it is somewhat more standardized, but you also get new tests that test the integration of the Cloud system. There is no influence on the deployment speed, because testing is the same for Cloud implementations as for on-premise implementations.

5.1.1 Hypotheses

As explained earlier, a survey is also conducted. The survey will examine the influence of the factor on the deployment speed. For every factor there is a hypothesis that states there is no influence on the deployment speed. This is done to conduct the same statistical analysis with the same test value for every factor.

H1: Top Management Support will have no influence on the deployment speed

H2: Implementation Strategy will have no influence on the deployment speed

Figure 10: The age and years of experience distribution

H4: Business Process Redesign and Software Configuration will have no influence on the deployment speed

H5: IT-Infrastructure will have no influence on the deployment speed H6: Training will have no influence on the deployment speed

H7: Data Migration Plan will have no influence on the deployment speed H8: System Testing will have no influence on the deployment speed These hypotheses will be tested via a survey, as elaborated earlier in the methodology.

5.2

Survey Results

The survey was conducted over a total of 53 ERP experts. Figure 10 shows the age and amount of experience distribution amongst the respondents. Interesting to see is that 58,5 percent of the respondents has 5 or more years of experience in ERP implementations, which indicates a high level of expertise amongst the respondents.

First, all the factors were checked for normality using the Skewness and Kur-tosis and by checking the distribution graphs. The results for the Skewness and Kurtosis can be found at Appendix C.4. After this, for the normally distributed factors an One-Sample T-Test will be conducted to check significant difference. This parametric test is chosen for all the factors that are normally distributed because the data also is on an interval scale, namely a 7-point Likert scale. Since the Likert scale is symmetric (because of the 7-points Likert scale and equidistant presentation), the scale can be considered to be on a interval scale (Brown, 1983), (Labovitz, 1967).

On top of that, Chronbach’s alpha was used to see if all the different factors, the independent variables, measure the same dependant variable: the deploy-ment speed. As displayed in Appendix C.3 the Chronbach alpha for the question on the difference in the factors between on-premise and Cloud implementations is 0,736. For the influence of the factors on the deployment speed the Chron-bach’s alpha is 0,700. Although slightly, both values are 0.7 or higher which indicates that the independent variables all measure the same dependant vari-able (Bland & Altman, 1997).

As mentioned before, the normal distribution was analyzed using the Skew-ness and Kurtosis. In Appendix C.4 the SkewSkew-ness and Kurtosis for both the questions about the difference as the questions about the influence are shown. A variable is normal distributed when the Skewness and Kurtosis is between -2.00 and +2.00 (George & Mallery, 2010). As shown in the Appendix, all fac-tors have a normal distribution, except for the question about the difference between on-premise and Cloud for IT-Infrastructure. Therefore, this question is analyzed using the Sign Test.

The other factors were analyzed using the One-Sample T-Test. This tests shows if there is a significant difference of the variable with the test value. In this case the test value = 4. The value 4 is the average on a 7-point Likert scale that is used during the survey. For the questions about the difference 1 corresponds to strongly disagree and 7 to strongly agree (to the statement that there is a difference between on-premise and Cloud for the specific factor). For the questions about the influence of the factor on the deployment speed, 1 corresponds to strongly negative influence and 7 to strongly positive influence. There is a significant difference with the test value when sig. <0.05 (of the T-Test).

Below the 8 different hypotheses will be discussed according to the statistical analysis of the survey data. All the statistical analysis can be found in Appendix C.

H1: Top Management Support will have no influence on the deploy-ment speed

According to the One-Sample T-Test of the questions regarding the difference between on-premise and Cloud for the factor Top Managament Support, there is no significant difference (sig. = 0,109). The test states that only values of lower than 0,05 are significantly different from the test value which in this case was 4 on a Likert scale from 1 to 7. The mean also indicates that most of the respondents agree there is no significant difference (M = 4,43, SD = 1,937).

The T-Test for the influence of Top Management Support on the deployment speed does indicate a significant difference. With a 0,000 value, it indicates that there is an influence on the deployment speed. Looking at the mean, than it can be concluded that the influence is positive (M = 5,96, SD = 1,055). The hypothesis however is that Top Management Support will have no influence on the deployment speed. This hypothesis will be rejected based on the results from the survey and a conclusion will be drawn that Top Management has a positive influence on the deployment speed.

Top Management Support has a positive influence on the deployment speed H2: Implementation Strategy will have no influence on the deploy-ment speed

For the difference of the factor Implementation Strategy between on-premise and Cloud the M = 5,32 (SD = 1,638). This is a difference of 1,32 from the

average 4. This indicates that there is a difference between on-premise and Cloud Implementation Strategy. The sig of 0,000 indicates that this difference of 1,32 is significant.

Also the influence indicates a significant difference with a sig. of 0,000. This indicates that Implementation Strategy has an influence on the deployment speed. The mean can indicate if this is a positive or negative influence and how strong this influence is. With M = 5,62 (SD = 1,078), it can be concluded there is a positive influence. Therefore the hypothesis can be rejected and it can be concluded that Implementation Strategy will have a positive influence on the deployment speed.

Implementation Strategy has a positive influence on the deployment speed H3: Change management will have no influence on the deployment speed

The interviews indicated that there are some differences in Change Management between on-premise and Cloud implementations. However, it was also stated that these differences were small. With M = 4,81 (SD = 1,744), this suggest there is a slight difference according to the 53 respondents. The sig. of 0,001 indicates that this difference is significant and it can be concluded that there is a difference for the factor Change Management between on-premise and Cloud. The sig. of the T-Test indicates that there is a significant difference with the average value of the influence on the deployment speed. This indicates that Change Management has a significant influence on the deployment speed. With M = 4,98 (SD = 1,366), it can be concluded that this influence is positive. This means that the hypothesis can be rejected and the conclusion is that Change Management has a positive influence on the deployment speed.

Change Management has a positive influence on the deployment speed H4: Business Process Redesign and Software Configuration will have no influence on the deployment speed

Most respondents acknowledged that there is a difference in Business Process Redesign and Software Configuration. With M = 4,85 (SD = 2,013) and a sig. of 0,003, the values significantly differ from the test value (4, which means no difference for the factor). Therefore it can be concluded that there is a difference (Mean difference = 0,811) for the factor between on-premise and Cloud ERP implementations.

Looking at the hypothesis that BPR and Software Configuration will have a positive influence on the deployment speed, a closer look is taken at the T-Test results of the survey for the influence related question. This indicates that there is a positive influence on the deployment speed, because M = 4,49 (SD = 1,705). It suggests that there is almost no significant difference, however the T-Test shows that the sig. is 0,041, which indicates that there is a significant

difference. It can be concluded that Business Process Redesign and Software Configuration has a positive influence on the deployment speed.

Business Process Redesign and Software Configuration has a positive in-fluence on the deployment speed

H5: IT-Infrastructure will have no influence on the deployment speed As mentioned before, the difference of the IT-Infrastructure between on-premise and Cloud is analyzed with the non-parametric Sign Test, since the variable is not normally distributed. In the interviews there was a strong indication that there is a big difference between on-premise and Cloud for the IT-infrastructure of an organization. The respondents acknowledged the findings of the inter-view by stating that there is a difference between on-premise and Cloud for IT-infrastructure, looking at M = 6,40 (SD = 0,906). Using the Sign Test (Ap-pendix C.6) results, because of the Asymp. Sig. value of 0.000, there is a strong indication that this difference is significant.

To test the hypothesis, the results of the T-test for the influence question needs to be examined. With M = 5,60 (SD = 1,276) and a sig. of 0,000, it can be concluded that there is a significant difference and that the influence is positive. Therefore the hypothesis will be rejected.

IT-Infrastructure has a positive influence on the deployment speed H6: Training will have no influence on the deployment speed

Looking at the results of the survey, there is no difference between on-premise and Cloud for Training, according to the respondents. This is indicated by M = 4,26 (SD = 1,831 and Mean difference is 0,26). On top of that, since sig. ¿ 0,05 (sig. = 0,298), it can be concluded that there is no significant difference.

For the influence of Training on the deployment speed of an implementation there is a bigger mean (M = 4,89, SD = 1,155). Together with the 0,000 sig. there is an indication that there is an influence on the deployment speed. Since the mean is 4,89, this indicates that the influence is positive, which corresponds with the results from the qualitative research. The hypothesis that Training will have no influence on the deployment speeds will be rejected.

Training has a positive influence on the deployment speed

H7: Data Migration Plan will have no influence on the deployment speed

Qualitative research concluded that there is no difference in Data Migration for Cloud, compared to on-premise ERP implementations. The results from the survey support this conclusion. With a mean of 4,08 (SD = 0,250) there is no difference between the average of 4. Also, because of the sig of 0,764, there is a strong indication that there is no significant difference and therefore no difference in the factor between on-premise and Cloud.