Capturing Value from Platform-as-a-Service Technology

Platform-as-a-Service Adoption Model for Large Enterprises

Image source: www.microsoft.com

Aryo Primagati

Management of Technology Faculty of Technology, Policy and Management Delft University of Technology, The Netherlands

Delft, July 2011

Suggestion for citation:

Primagati, A. (2011). Capturing Value from Platform-as-a-Service Technology: Platform-as-a- Service Adoption model for Large Enterprises. Unpublished master’s thesis, Delft University of technology, Delft.

This manuscript is designed for double-sided printing

Master Thesis Project

Thesis submitted in partial fulfillment of the requirements for the degree of Master of Science in Management of Technology

Faculty of Technology, Policy and Management Section Information and Communication Technology Delft University of Technology

Graduation Committee: Chairman : Dr. Yao-Hua Tan (Delft University of Technology) First Supervisor : Dr. ir. Mark de Reuver (Delft University of Technology) Second Supervisor : Dr. Martijn Warnier (Delft University of Technology) Third Supervisor : Fatima Nikayin, MSc. (Delft University of Technology) External Supervisor : Ir. Peter Wegbrands (Accenture)

Author: Aryo Primagati Student ID 4035992

Management of Technology Faculty of Technology, Policy and Management Delft University of Technology, The Netherlands

This page intentionally left blank

Executive Summary

Cloud computing market is expected to grow rapidly in the next five years. Even though Software-as- a-Service and Infrastructure-as-a-Service solutions will dominate most of the market, Platform-as-a- Service solution is forecasted to have the fastest-growing segment, especially in Western Europe. On the other hand, cloud computing is also argued to be the new playing ground for Telecommunication industry. As the industry facing a threat for just being the network “bit piper”, telecom operators might find a new revenue source within the cloud computing domain.

The issues explained above have led us to conduct a research that aimed at two objectives. First is to give guidance to larger enterprise companies in the Netherlands on how to capture values from PaaS offering. Second, this research aimed to identify whether telecom companies could create particular new business value and opportunities from the PaaS service provision. Therefore this study raised a challenging main questions:

“What factors are important for large enterprise clients to adopt Platform-as-a-Service in order to capture the business value that is offered from Platform-as-a-Service offering, and what are the strategic implications of the Platform-as-a-Service adoption for Telecom industry?”

To answer the main questions, five research sub-questions were formulated. A theoretical approach from business model literature is used to guide answering the main research questions. A conceptual model of PaaS adoption factors was developed, which is based on the STOF business model framework. Analysis of PaaS market in the Netherlands was conducted among 9 client companies and 3 PaaS provider companies, in order to validate the conceptual model.

The results showed that the PaaS market in the Netherlands are still immature, thus making the validation of the conceptual model for PaaS adoption hard to be generalized. Modification of the analysis approach was taken that enable us to analyze the conceptual model for SaaS adoption as well.

Our study found that there are several factors within the Service domain that are important for clients when they consider which applications to be moved to PaaS or SaaS environment. The most important factor is criticality of application that wants to be moved to the cloud. These factors from the Service domain influence client’s technical requirements on Security, Quality of Service, and System Integration issues. Furthermore, our research found that there are several factors that client take into account when they about to choose PaaS and SaaS providers. These factors serve as assessment framework whether the providers’ profile can meet clients’ requirements. Our finding shows that Branding is the most important factor of PaaS and SaaS providers. Branding is important because it is closely associated with trustworthiness. Moreover, trust also becomes really important, because in PaaS and SaaS implementation clients are losing control of their application data. In addition, our finding also shows that the factor of data location is really important in PaaS and SaaS adoption. A key stakeholder, namely the regulators, drives this data location factor. We also found that clients perceived system integrators and network providers as the other important key stakeholder in the provision of PaaS and SaaS services.

Page | i

Furthermore, our study also found that telecom companies could play two roles in cloud computing domain in order to create new business value, and hence making new revenue source. The first role is to be Cloud service provider. This role means that telecom companies could exploit their main resources of Internet network infrastructure to deliver a guaranteed and high level end-to-end cloud services (both PaaS or SaaS services). The second role is to be Cloud service broker. This role exploits telecom companies’ resource of broad customer base and customer relationship in order to mediate cloud providers and the clients. Being cloud service broker enables telecom companies to gain monetary incentives by means of revenue sharing.

Our study concludes that clients need to take into account factors in the four domains of Service, Technology, Organization, and Finance in order to capture the value of PaaS and SaaS offering. On top of that, clients need to take into consideration the regulation applied in order to legally implement cloud computing solutions. Clients also need to acknowledge the importance of system integrators and network providers in utilizing cloud computing services.

This study adds several main contributions to the literature. Firstly, this study has made the first attempt to construct an adoption model for PaaS and SaaS specifically. Secondly, this study has contributed to business model theory, particularly by extending the applicability of STOF model. Thirdly, this study confirms the existing literature that describes key stakeholders on cloud computing. On top of that, we contribute by adding Network Provider as another key stakeholder.

Nevertheless, this study bears some limitations. Due to small sample size, this research serves as means of first validation of the constructed model. We are aware that further research can be conducted to do more extensive validation by means of survey research. However due to the immaturity of cloud computing market, especially PaaS, we suggest that further research on PaaS is better to be conducted later to wait for the market to be developed. Besides the sampling issue, our research was not able to validate our findings on telecom companies role in cloud computing. Further research could focus on validating our results empirically through involvement of industry experts.

Keywords: cloud computing, Platform-as-a-Service, Software-as-a-Service, business model, market adoption, telecommunication industry, adoption factors

Page | ii

Acknowledgement

This thesis addressed a topic of market and adoption of a new technology. This time, I had the opportunity to come to an “unknown” territory of cloud computing, especially about Platform-as-a- Service. I have been really excited to work on this topic. This topic has given me exciting challenges that really motivates me to continue to learn more and more. This thesis conclude my two-years Master’s study of Management of Technology at TU Delft. I am really grateful to have the opportunity to conduct my thesis research at Accenture. It has given me invaluable experience and learning opportunities throughout my six months internship period in the company.

Foremost, I would like to say my gratitude to God. Furthermore, I would like to convey my special gratitude to Mark de Reuver for giving me invaluable time, feedbacks, and discussion in order to sharpen my thoughts in doing such an academic research. I also would like to convey my special gratitude to Peter Wegbrands for supervising me throughout my stay in Accenture and introducing me to the consulting world. I am also grateful to Harry Bouwman that helped me for initiating and realizing this project, also for the insightful inputs and comments on my thesis. I am also grateful to Martijn Warnier and Fatima Nikayin for their critical inputs and discussion.

Next, my gratitude goes to the thirteen interviewees that have cooperated and devoted their time for this research. Also my gratitude goes to Koen van de Biggelaar from Accenture for the exciting discussion on the topic of cloud computing and telecommunication world that has given me a lot of insight.

Most importantly, I would like to thank Lia for accompanying my journey during the whole 2 years of my Master’s study at TU Delft. Thank you for your endless inspiration, encouragement and supports. Thank you for always making me smile throughout this two-years roller coaster ride.

I would like also to thank my family in Indonesia: Papa, Mama, mbak Tia, and mas Dodi for their continuous prays and unconditional supports throughout the whole period of my stay in the Netherlands.

Last but not least, my special thanks go to Ramanda, Ryan, Kukuh, Tungky, Mas Prapto, Mas Andri, Ronald Halim and Uti for spending their time for all the discussion and proof reading my work. Thanks to Tipang for accompanying me during many “library” sessions. Also I would like to thank my classmates for all the friendships and adventures in and outside the class. Finally, I am indebted to NUFFIC for giving me this huge opportunity to continue my Master’s education and pursuing my dream.

Thank you all!

Delft, 4th of August 2010

Aryo Primagati

Page | iii

Page | iv

Table of Contents

Executive Summary ...... i Acknowledgement ...... iii List of Figures ...... vii List of Tables ...... ix List of Abbreviations ...... x 1 Introduction...... 1

1.1 Research Problem ...... 2 1.1.1 The New Challenge in Telecommunication Industry ...... 3 1.2 Aims and Research Questions ...... 6 1.3 Research Approach ...... 9 1.4 Relevance...... 12 1.5 Report Outline ...... 12 2 Cloud Computing Domain ...... 15

2.1 Cloud Computing Definition ...... 15 2.2 Software-as-a-Service ...... 18 2.3 Platform-as-a-Service ...... 19 2.4 Infrastructure-as-a-Service ...... 20 2.5 Public and Private Cloud ...... 21 2.6 Advantages and Disadvantages ...... 21 2.6.1 PaaS advantages over SaaS ...... 22 2.7 Cloud Security and Reliability ...... 23 2.8 Cloud Key Stakeholders ...... 24 2.9 Cloud Market Overview ...... 24 2.9.1 Dutch PaaS Market Overview ...... 25 2.10 Trends of Telecom Companies Activities in Cloud Computing ...... 27 2.11 Conclusion ...... 30 3 Theoretical Background...... 31

3.1 Related Works on Cloud Computing ...... 32 3.1.1 Application of the related works toward client PaaS business model ...... 33 3.1.2 Conclusion ...... 34 3.2 The Concept of Business Models ...... 34 3.2.1 Business Model and Strategy ...... 35 3.2.2 Business Model Definition and Concept ...... 35 3.2.3 The STOF Model ...... 36 3.2.4 Applying STOF Model to Client PaaS Business Model ...... 40 3.2.5 Conclusion ...... 45 3.3 Integrated PaaS Adoption Factors ...... 46

Page | v

3.4 Conclusion ...... 47 4 Towards a Validated Adoption Model ...... 49

4.1 Interview Methodology ...... 50 4.1.1 Why Interview? ...... 50 4.1.2 Interview Respondents Selection ...... 50 4.1.3 Interview Process and Protocol ...... 51 4.1.4 Data Management and Analysis Approach ...... 53 4.2 Results and Analysis...... 55 4.2.1 PaaS Market Condition ...... 55 4.2.2 Towards a Validated Model ...... 58 4.2.3 Further Analysis of Interview Results on ...... 70 4.2.4 Summary...... 76 5 Potential Role of Telecom Companies in the Cloud ...... 81

5.1 Telecom Companies Competences and Roles ...... 81 5.2 Towards Potential Role in Cloud Computing Services ...... 84 5.2.1 Role of Cloud Service Broker ...... 85 5.2.2 Role of Cloud Service Provider ...... 86 5.3 Conclusion ...... 88 6 Conclusions and Recommendations ...... 89

6.1 Main Findings ...... 89 6.2 Reflection ...... 92 6.3 Contribution to Literature ...... 97 6.4 Recommendation for Practitioners ...... 98 6.5 Limitations and Further Research ...... 99 6.5.1 Research Limitations ...... 99 6.5.2 Direction for Further Research ...... 99 References ...... 103 APPENDIX A - Telco Market Overview in the Cloud ...... 109 APPENDIX B - Telecom Market in the Netherlands ...... 110 APPENDIX C – Interview Protocol ...... 111

Page | vi

List of Figures

Figure 1 Structure of Chapter 1 ...... 2 Figure 2 Mobile service revenues in the coming years ...... 5 Figure 3 The traffic challenge that is faced by MNO ...... 5 Figure 4 Research questions construct ...... 9 Figure 5 Conceptual framework of the proposed research ...... 10 Figure 6 Cloud computing main player groups, adopted from ...... 11 Figure 7 Report structure ...... 13 Figure 8 Structure of Chapter 2 ...... 15 Figure 9 Three-layer cloud computing architecture ...... 17 Figure 10 Cloud business model ontology ...... 18 Figure 11 Western Europe cloud service market by segment in 2009 and forecasted market in 2014 ...... 25 Figure 12 Two standpoints of the client PaaS business model ...... 31 Figure 13 Structure of Chapter 3 ...... 32 Figure 14 The STOF model ...... 37 Figure 15 STOF model Critical Design Issues and Critical Success Factors relating to creating customer value .. 38 Figure 16 Critical Design Issues and Critical Success Factors relating to creating network value ...... 39 Figure 17 Applicability of CDI Targeting ...... 41 Figure 18 Applicability of CDI Branding and Creating Value Elements ...... 41 Figure 19 Applicability of CDI Customer Retention ...... 41 Figure 20 Applicability of CDI Security and Quality of Service ...... 42 Figure 21 Applicability of CDI System Integration ...... 42 Figure 22 Applicability of CDI Partner Selection ...... 43 Figure 23 Applicability of CDI Network Governance ...... 44 Figure 24 Applicability of CDI Pricing ...... 44 Figure 25 Applicability of CDI Cost ...... 44 Figure 26 Applicability of CDI Revenue ...... 45 Figure 27 The identified client PaaS business model design issues derived from STOF model ...... 45 Figure 28 The integrated PaaS adoption factors model ...... 47 Figure 29 Interview results ...... 55 Figure 30 Application selection factors for PaaS adoption ...... 59 Figure 31 Application selection factors for SaaS adoption ...... 59 Figure 32 Application selection factors for PaaS adoption from vendor perspective...... 61 Figure 33 Value elements for internal organization of PaaS adoption from client perspective ...... 62 Figure 34 Value elements for internal organization of SaaS adoption from client perspective ...... 62 Figure 35 Value of PaaS adoption for organization from vendor perspective ...... 64 Figure 36 Technology domain for PaaS adoption ...... 64 Figure 37 Technology domain for SaaS adoption ...... 65 Figure 38 Vendor selection factors for PaaS adoption ...... 67 Figure 39 Vendor selection factors for SaaS adoption ...... 68 Figure 40 PaaS vendor selection factors from vendor perspective ...... 70 Figure 41 Desired PaaS vendors ...... 71 Figure 42 Desired SaaS vendors ...... 71 Figure 43 Perceived barriers to adopt PaaS from client perspective ...... 72 Figure 44 Perceived barriers to adopt SaaS from client perspective ...... 72 Figure 45 Perceived risks from PaaS adoption (client perspective) ...... 74 Figure 46 Perceived risks from SaaS adoption (client perspective) ...... 74

Page | vii

Figure 47 Identified potential role of telecom companies in the cloud service provision ...... 75 Figure 48 The integrated conceptual model of PaaS adoption proposed in Chapter 3.3 ...... 76 Figure 49 Validated model of PaaS adoption factors ...... 77 Figure 50 Validated model of SaaS adoption factors ...... 78 Figure 51 Dutch mobile telephony market (left) and broadband market (right) as of 2010 Q3 ...... 110

Page | viii

List of Tables

Table 1 List of interview respondents from PaaS vendors ...... 11 Table 2 List of interview respondents from potential PaaS customers ...... 11 Table 3 Cloud services categories ...... 21 Table 4 Overview of PaaS vendors ...... 26 Table 5 Categorization of Cloud Offerings from Telcos ...... 28 Table 6 Cloud computing adoption strategies ...... 32 Table 7 Applicability of cloud computing adoption strategy towards Client PaaS business model ...... 33 Table 8 Applicability of cloud computing adoption factors toward client PaaS business model ...... 33 Table 9 Interview respondents ...... 51 Table 10 Example of interview questions ...... 52 Table 11 Classification of client interviews ...... 54 Table 12 Classification of vendor interviews ...... 54 Table 13 Consolidated list of telecom companies' resources and capabilities ...... 82 Table 14 Aggregated generic roles of telecom companies in the telecommunication industry ...... 84

Page | ix

List of Abbreviations

ADSL Asymmetric Digital Subscriber Line API Application Programming Interface CCC Communication and Collaboration CRM Customer Relation Management DSL Digital Subscriber Line ERP Enterprise Resource Planning IaaS Infrastructure-as-a-Service ISP Internet Service Provider ISV Independent Software Vendor MNO Mobile Network Operator MVNO Mobile Virtual Network Operator QoS Quality of Service SaaS Software-as-a-Service SCM Supply Chain Management SFDC Salesforce.com SI System Integrator SLA Service Level Agreement SME Small-Medium Enterprise STOF (model) Service Technology Organization Finance (model) PaaS Platform-as-a-Service PC Personal Computer SQL Structured Query Language Telco Telecommunication Company VAR Value Added Reseller

Page | x

This page intentionally left blank

Page | xi

1 Introduction

Today, companies in general have been shifting their way in acquiring computing power to support their business towards web-based applications. Cloud computing, as they call it, provides computer resources through the (Internet) networks instead of utilizing local computers to run software applications and store data.

It has been generally assumed that cloud computing offers benefits of more flexibility and lower upfront investment compared to the conventional system. Moreover, public IT cloud service has been estimated to grow at the rate of five times more compare to traditional IT products, with a forecast to reach $55.5 billion worth of market in 2014 worldwide (IDC, 2010b).

Cloud computing market is segmented as Software-as-a-Service (SaaS), Platform-as-a-Service (PaaS), and Infrastructure-as-a-Service (IaaS); in which they have different maturities and adoption levels. So far, SaaS and IaaS have been leading the cloud computing market. In 2009 in Western Europe, SaaS took 39.7% of the cloud computing market share, while IaaS took 51.8% market share (IDC, 2010a). Moreover, as an example, SaaS has been offered in the form of Customer Relationship Management (CRM, Enterprise Resource Planning (ERP), Communication and Collaboration (CCC), and Supply Chain Management (SCM) applications. Meanwhile, IaaS for instance has been offered in the form of cloud-based compute capacity and storage services.

In contrast, PaaS has yet to develop its market in Europe, as well as in the Netherlands. As an example of PaaS service, PaaS services have been offered in a form of a platform to develop SaaS applications. A market research by IDC (IDC, 2010a) shows that in 2009 PaaS had only 8.6% share in cloud computing market in Western Europe. However, IDC forecasts that PaaS would be the fastest- growing segment in cloud computing market within the next 5 years time. One of the reasons why PaaS market has not been well developed is because it is not yet known clearly what kind of application services that organizations and companies in general want to migrate to the cloud by utilizing PaaS technology. Companies in general lack guidance to adopt the technology in order to add business value (Marston, Li, Bandyopadhyay, Zhang, & Ghalsasi, 2011). This issue about PaaS adoption makes it an interesting question that we would like to set our focus on. Particularly, we are interested on PaaS adoption from the perspective of PaaS client, especially large organizations.

Furthermore, even though IaaS and SaaS have bigger share in cloud computing market, Gonçalves and Ballon (2011) and Rake-Revelant et al. (2010) argue that PaaS based service is also seen to be an opportunity to bring new business value and revenue source for the Telecom industry. Telecom industry has been struggling to offer new business value as the new 4G-network technology begins to emerge. The 4G- network technology, which characterized by all-IP based communication, is threatening telecom operators as the conventional business models and revenue sources of phone calling and SMS have shown the tendency to decrease for the coming years. Even though the data service becomes the new source of revenue for telecom operators, this revenue growth is not expected to compensate the voice and SMS revenue loss (T-Mobile, 2010).

Page | 1

Chapter 1

Therefore, along with our previously mentioned focus on PaaS technology market development in general, we would also like to focus our attention deeper on the implication of PaaS market development for Telecom industry in particular.

Furthermore, we would like to examine this problem from the perspective of customer-centric/client business model and multi-sided platform theories, in order to give companies guidance on how to successfully adopt the PaaS technology to create new business value.

In the following section, the domain about the cloud computing and the Telecom industry will be further elaborated. The relevant theoretical backgrounds that bring our point of view of the problem into perspective will also be elaborated. In the section afterwards, research objective and research questions are presented. Next to that section, the research methodology of how the research will be conducted, possible conclusion and outcomes, and the research schedule are discussed subsequently.

1.2 1.1 1.3 1.4 1.5 Aims and Research Research Problem Research Approach Relevance Report Outline Questions

Figure 1 Structure of Chapter 1

1.1 Research Problem

Based on the related domain and theory backgrounds mentioned in the previous paragraph, we would like to formulate our research problem and research objectives in this section. We identify several problems that we want to address in this research, which will be elaborated briefly below.

The first problem we would like to identify is based on the following arguments. Weinhardt (2009) argues that further research in the area of new business model of cloud computing is necessary. Furthermore, Marston et al. (2011) also argue that further research in the area of cloud computing technology adoption is essential. They believe that further research is needed to give organizations guidance in developing their technology in relation with cloud computing, particularly to decide which applications are best positioned to be moved to the cloud and how to adopt this technology in the least disruptive way. Therefore, we would like to focus our research to address the problem of the lack of guidance for organization in developing their cloud computing business model and capturing values out of the technology.

Secondly, out of the three generic services of cloud computing (SaaS, PaaS, and IaaS), IDC Research has shown that PaaS market in Western Europe has the lowest share of 8.6% only (IDC, 2010a). However, IDC has also forecasted that PaaS would be the fastest-growing market segment within the coming years (IDC, 2010a). This means that PaaS technology concept has yet to reach its maturity and under a rapid development. This becomes our reason to focus our research on PaaS over the other cloud computing services.

Thirdly, Rake-Revelant (2010) and Gonçalves and Ballon’s (Gonçalves & Ballon, 2011) have argued that PaaS brings a lot of potential in creating new business opportunities especially in the Telecom industry domain as has been discussed in previous paragraph. As the (Internet) network provider of all cloud-based services, telecom companies (both wire line and wireless) have a big stake and good

Page | 2

Introduction potential on the cloud business. Therefore, we believe Telecom companies could do more than just a bit piper for the Internet connection in the business. Like what has been discussed in the previous section, one way to do this is by utilizing the PaaS concept like what has been shown by NTT Com by its partnership with Salesforce.com (Salesforce.com, 2010), and also by O2, and Orange by their experimental development platform (Gonçalves & Ballon, 2011).

However, this kind of role of telecom companies has yet to be seen in the cloud market in the Netherlands. It is not yet clear how exactly Telecom industry could benefit the adoption of this PaaS technology concept in practice in the Netherlands. In addition to our initial idea to research on the business model topic of PaaS in generic level, this issue has attracted our attention to also focus deeper into Telecom industry in particular. We would like to investigate further what other role that Dutch telecom companies could play in the Dutch PaaS market and business other than being customers of PaaS service like the other companies. In the following section, we will elaborate further the reasons of bringing the topic of PaaS and telecommunication industry together.

1.1.1 The New Challenge in Telecommunication Industry This section will discuss the domain of Telecommunication Industry in correlation with the cloud computing domain. As discussed in the previous paragraph, we would like to pay a special attention on Telecom industry in relation to PaaS utilization in our research. This section will elaborate further why we think that telecom companies have a particular opportunity to develop their business by PaaS concept.

Why PaaS? First of all, being the (Internet) network provider, telecom companies play a big stake on the provision of cloud computing services because all cloud-based services use the Internet connection to be accessed. It has been further argued by Rake-Revelant (2010) and Gonçalves and Ballon (Gonçalves & Ballon, 2011) that telecom company can create new business value and new revenue source by utilizing cloud computing concept, other than just being a bit piper to provide data service only.

From resource-based point of view, telecom companies have the indispensable resources such as telecommunication network infrastructures (Internet, mobile Internet, etc.). Telecom companies also have large customer base, either public consumers as well as corporate customers. Moreover, telecom industry in general is one of the industries that have a lot of experiences with billing. Therefore, it is quite apparent that telecom companies have big opportunities to play a role in cloud-computing business.

Furthermore, Rake-Revelant (2010) and Gonçalves and Ballon (2011) also showed that some telecom companies has been shifting their focus towards PaaS (in comparison to SaaS or IaaS model) to make the best out of cloud computing. They argue that previous attempts have been made by telecom companies to diversify their service through SaaS-based concept such as WAP, but they have failed or are experiencing slow take off (Gonçalves & Ballon, 2011). This failure is argued due to unfavorable business model that does not have good incentives for application providers and users, and also because telecom companies faced too great risk

Page | 3

Chapter 1

in the implementation of this SaaS model (management of network of suppliers, deal with performance, scalability and security issues, etc.) compared to the small value delivered to customers, content, and application providers.

telecom companies have been showing their focus to PaaS by experimenting to create platform environment for development of mobile applications (Gonçalves & Ballon, 2011). This has been shown by the introduction of Orange Partner (Orange, 2011), O2 Litmus (beta) (O2, 2009), and Vodavone Betavine (Vodafone, 2010) platforms. Within this platform, the telecom companies are exposing their telecommunications network functionalities through application programming interfaces (APIs) to application developer that will take up service innovation. On the other hand, another telecom company has also seen the opportunity in PaaS from different angle. NTT Communications (NTT Docomo) Japan for example, they have seen opportunity in PaaS by partnering with Salesforce.com (Salesforce.com, 2010). Besides establishing data center in Tokyo to support Salesforce.com cloud computing service, NTT Docomo takes care of billing, customer account management and security of Salesforce.com’s PaaS service. NTT Docomo does the marketing and sales of Salesforce.com and the registered applications in the marketplace and earn commissions from that. The examples presented above shows that Telecom industry has started to think PaaS as a new source of revenue, other than just providing Internet connection. It also has been shown that there is more than one business model that telecom companies could adopt to utilize PaaS.

Even though it has been shown that mobile network operator (MNO) are shifting towards and experimenting with the PaaS concept, these platforms have not been showed success stories as far as we observed. Moreover, to the best of the author’s knowledge, there has not been much discussion on what exactly the best business model approach is and what the key ingredients are for telecom companies in order to successfully utilize the PaaS concept to create new business values. That is why and where we would like to put or attention to in this research.

Why Telecom Companies Need New Source of Revenue? To get one step back, we would like to bring up one important question why the telecom companies need to create new business value and new revenue source? The reason will be elaborated further in the following paragraphs.

Based on report from OPTA (2011a) telecom companies’ business of fixed telephony is losing ground to other forms of communication such as mobile telephony and the Internet. While on the other hand the mobile telephony service also received enormous pressure from the explosive growth of mobile internet usages. This condition gets more challenge by the introduction of new technology. Currently, Telecom industry is facing new technology emergence called 4G-network. Forge (2004) define 4G by four attributes: - Ubiquitous support for broadband-based media products, with adequate high performance and complete reliability at far lower cost (than 3G0 per packet - Lower cost of infrastructure and roll-out full national coverage - Handling the integration of previous generation of mobile, fixed wire lines and wireless networks - Use of newer radio technology for re-use of the spectrum without limits

Page | 4

Introduction

4G-network here is argued to be important because it “offers a highly focused solution to the problems of multiple heterogeneous networks” (Forge, 2004). Another important characteristic here is the extensive use of mobile data service. This implies that the infrastructure could threaten the traditional way of mobile phone calling, texting, and other business models. The network allows high-speed Internet connection in mobile phones. With the maturity of voice-over-IP (VoIP) calling, wireless VoIP (using VoIP via mobile Internet networks) could indeed threat the core business and revenue source of mobile operators (Accenture, 2010). Thus creating a big challenge for MNO to create new business value to stay competitive.

Furthermore, as depicted in Figure 2, it has been estimated that the revenue for MNO is decreasing for the coming years. The revenue growth from mobile data cannot compensate on the revenue loss from voice and text messaging (SMS). This problem creates an interesting challenge for MNO on how to make more revenue from their services.

Figure 2 Mobile service revenues in the coming years (T-Mobile, 2010)

Figure 3 The traffic challenge that is faced by MNO (Dogra & Orr, 2009)

Moreover, the increase traffic of mobile data usage introduces a so called the “traffic challenge” for MNO. As depicted in Figure 3 above, in voice domain, as the traffic of voice calls increase, the revenues will increase accordingly. On the other side, as more and more

Page | 5

Chapter 1

people are using Smartphone and exploiting the mobile data capabilities (mobile Internet), the increased traffic of mobile data is not followed by the increase in revenues in a proportional way. One of the causes is due to the flat-tariffs pricing scheme that MNO adopts in offering this data service.

Furthermore, there are other reasons why we would like to limit our research in the Dutch cloud computing market. Firstly, to the best of our knowledge, there is yet to be an in depth information of PaaS market in the Netherlands in particular. This means that PaaS market and business opportunity in the Netherlands is still, in the bigger part, unexplored. Also, this creates uncertainty among the organizations in the Netherlands on how to successfully adopt PaaS to increase their business value. Secondly, due to the geographic location of the author that is located in the Netherlands, and also due to the feasible organizations contacts that the author has in the Netherlands in order to collect the data.

Moreover we have another limitation that we would like to apply in this research. As has been discussed in the previous section, Sultan and Marston et al. (2011) assumed that startup companies and Small-Medium-Enterprises (SMEs) might have more eminent benefit in using cloud computing technology. They argue that significant technical, operational and organizational issues might hinder large organizations in capturing the values and benefit that are familiar for startups and SMEs. Therefore we would also like to scope our research to focus on large organizations to explore more in depth how those large organizations can capture value from PaaS in particular. Another reason of this choice is because the author has more feasible contacts with large organizations to do the research on than with startups and SMEs.

1.2 Aims and Research Questions

Based on the problems that are defined in previous section, this section will formulate the problems into a research objective and research questions. These objective and questions are going to be used as a guide to conduct the research activity.

Following the problem definition presented above, we would like to formulate the objective of this research as: To give guidance on how large enterprise companies as clients in the Netherlands can make use of the PaaS technology concept in order to successfully capture business value from the PaaS offerings, and to identify whether telecom companies could create particular new business values and opportunities from the PaaS service provision.

The research would begin with indentifying the current market of PaaS in the Netherlands. We would start by finding out in detail what kind of services and business applications that organizations want to migrate to the cloud, what business values they expected from utilizing cloud services and also what kind of services and applications the PaaS providers offer. We would then analyze how the demand and supply can be matched and what business model(s) is/(are) appropriate in order to create and bring the expected new values for the organization as PaaS consumers. Essential factors that the organizations and providers of PaaS think very important in the adoption of PaaS would also be deeply investigated. We would also like to know whether the type of applications they want from

Page | 6

Introduction

PaaS and the important factors surrounding its adoption are similar across organization in different industries in the Netherlands.

Furthermore, as the previous step gives the information of the essential measures in adopting PaaS addressed by the PaaS providers and (potential) customers in the Netherlands, after that we want to focus its implications for Telecom industry. We want to analyze what particular business model(s) is/(are) appropriate for telecom companies (as network providers) in order to successfully create new business value out of the generic PaaS service provision, especially within the Dutch PaaS market. This analysis is aimed to explore value creation for telecom companies beyond their role as PaaS costumers.

Moreover, to reflect and guide the objective of our research as mentioned above, we formulate the research questions in the following. The main research question is:

“What factors are important for large enterprise clients to adopt Platform-as-a-Service in order to capture the business value that is offered from Platform-as-a-Service offering, and what are the strategic implications of the Platform-as-a-Service adoption for Telecom industry?”

To help answering the main research question, several sub-questions are arisen. They are presented in the following paragraphs.

As mentioned before in the research objective paragraph, PaaS market in the Netherlands is going to be analyzed to give us information about the key players of PaaS service provision. This analysis would become the starting point to identify the adoption factors of PaaS adoption. Therefore our first sub-question is:

Sub-question 1 “Who are the key stakeholders in PaaS service provision in the Netherlands? What kinds of providers are offering PaaS service and what kind of organizations are the (potential) PaaS customers?”

Based on the stakeholder and market analysis from sub-question 1, we will proceed with analyzing what sort of business model is appropriate to adopt PaaS for enterprise companies. We want to know in more detail how exactly a client firm can capture the added value from PaaS offering. We want to know what factors are important for a client firm to take into consideration in order to utilize PaaS to its full extent. We will use the STOF model framework to derive the factors that are essential for the client firms in delivering the PaaS-based service towards their end-customers. This analysis will focus on the service, financial, and organization domain of the service delivery model. Service domain reflects what kind of PaaS service is wanted, financial domain reflects what kind of pricing scheme is appropriate, and organization aspect reflects who are the significant players that are crucial to deliver the wanted service and how to select them. Therefore, the second sub-question is:

Sub-question 2 “What sort of generic PaaS-related business model from client’s perspective is appropriate to deliver the wanted cloud services in terms of service, organizational and financial domain?

Page | 7

Chapter 1

Accordingly, we would like to validate our analysis from the sub-question 2. As PaaS is the underlying development environment of SaaS applications, we will validate our analysis by identifying what kind of business software applications the organizations across the industry want to migrate to the cloud. In other words, to understand the types of SaaS applications preferred by the organizations. By this means, we would like analyze how exactly the organization can utilize PaaS in order to create new business value. Based on this information, we will analyze whether those applications can make use of the concept of PaaS and whether our analysis model is representing the real practice. Therefore, the third sub question is:

Sub-question 3 “What sort of business software applications do client organizations want to build or migrate by utilizing PaaS to the cloud, and how this is related with the generic PaaS-related business model?”

Afterwards, analysis will be conducted to see whether single or different types of business models are arisen from the different organizations in utilizing PaaS. What kind of similarities and differences can be extracted and analyze in the provision of PaaS service among different organizations. This question will help us to identify and formulate the essential factors to successfully adopt PaaS technology. Therefore the fourth sub-question is:

Sub-question 4 “What sort of similarities and differences arise among the identified client PaaS-related business model(s)?”

By answering the sub-question 4, we expect to have the ideas of the generic essential factors in adopting PaaS technology. Furthermore, to answer the latter part of our main research question, we would like to identify the implication for telecom companies in relation to the provision of the PaaS service. We would like to know whether telecom company can make particular benefit from the PaaS service offerings and how, especially from being network providers instead of just being PaaS customers. Further examination on the answers of sub-question 3 will be conducted to help us answering this question. Thus our fifth sub-question is:

Sub-question 5 “What sort of potential role Dutch telecom companies can play in the (potential) PaaS service offerings?”

In order to elaborate the role of telecom companies in PaaS service provision even further, we would try to analyze an appropriately potential business model for telecom companies. We would try to identify a feasible business model for telecom companies in the Netherlands that is derived from the generic business model that we will identify from the previous sub-questions.

In conclusion, sub-questions 1,2,3, and 4 are going to help us in answering the first part of our main research question, that is “What factors are important for enterprise clients to adopt Platform-as-a- Service in order to capture the business value that is offered from Platform-as-a-Service offering?”. Sub-question 3 and 5 in particular will help us in answering the last part of our main research question, that is “what are the implications of Platform-as-a-Service adoption for Telecom industry?”. Figure 4 below the construction of the research questions as explained above.

Page | 8

Introduction

SQ1: Who are the key stakeholders in PaaS service provision in the Netherlands? What kinds of providers are offering PaaS service and what kind of organizations are the potential (PaaS) customers? SQ3: What sort of business software applications do client organizations want to build or migrate by using PaaS to the cloud, and how this is related with the generic PaaS- related business model? SQ2: What sort of generic PaaS- related business model from client’s perspective is appropriate to deliver the wanted cloud services in terms of service, organizational and financial domain? SQ5: What sort of potential role SQ4: What sort of similarities and Dutch telecom companies can play differences are arisen among the in the (potential) PaaS service identified PaaS-related business offerings? model(s)?

RQ1: What factors are important for large enterprise clients to adopt Research RQ2: What are the strategic PaaS in order to capture the Questions implications of PaaS adoption for business value that is offered from Telecom industry? PaaS offering

Figure 4 Research questions construct

1.3 Research Approach

This section will elaborate our approach in conducting this project to answer the research questions. This section will also contain the methodology we chose to carry out the project.

In order to gain deeper insight to the unexplored research problems and to answer the research questions, an exploratory qualitative research approach is chosen. The rationale behind this selection is based on the fact that business model and development topic in the PaaS domain is relatively a new field of research. Exploratory research will allow the author to gain deep information about the topic from the related stakeholders and experts (Verschuren & Doorewaard, 2005).

Furthermore, as can be seen in Figure 5 below, we adopts the STOF business model (Bouwman, de Vos, & Haaker, 2008) to frame our research. This research uses the perspective of business model in order to solve the research problems, and focus on the service, technology, financial and organization domain. We consider the Dutch market as the market development external force, and also only select PaaS technology development as the enabling technology. These two input factors will be the main input forces in our analysis. In our business model analysis, we will take the perspective of PaaS client. We will consider the cloud application (SaaS) that client wants to build for its end-users as the “service”. Then we would like to understand how PaaS impacts each of the four domain in the STOF model for that cloud application. Thus, as the outcomes, the adoption factors of the PaaS business model development for client will be analyzed in order to create values for client organizations in general, and for telecom companies in particular.

Page | 9

Chapter 1

Dutch market analysis

Cloud Computing Technology

Generic value creation for organizations

IaaS SaaS PaaS

Specific value creation in Telecom industry

Research focus

Figure 5 Conceptual framework of the proposed research

The research approach would begin with literature study and desk research. The desk research will be conducted to gain more information on the related subject of cloud computing, PaaS, and the condition of Telecom industry at the moment. The literature study will be conducted to gain theoretical insight in the field of platform theory and business model theory. As mentioned before, these theories will be used mainly as the point of view in answering the research questions. This phase would aim to build up the perspective of the authors on the related matter, and also to do preparation to collect data.

Qualitative approach is going to be conducted throughout the study and in terms of data collection, we will perform some interviews. Semi-structured interviews with representative and experts of industry organizations are going to be conducted. Due to the time limitation to conduct this research, interviews are only going to be conducted with two groups of the main cloud computing players. The group classification of the main players are based and adopted from Marston et al. (2011) as depicted in Figure 6. These two groups that we would like to focus on are: (1) the cloud computing providers/vendors, (2) the (potential) consumers/clients.

Page | 10

Introduction

System Integrator / Enabler

Cloud Computing End Users / Vendors PaaS Consumers

Regulators

Figure 6 Cloud computing main player groups, adopted from (Marston et al., 2011)

In the following paragraphs below, tables of the potential respondents for the interviews are presented. As mentioned in the previous section, the author has more feasible contacts with large organizations. Therefore this research will limit its scope by focusing primarily on large organizations as the research objects. Moreover, due to the preliminary nature of this proposal document, the selection of the industry sectors and the organizations presented here are tentative. Careful study and final selection will be performed during the development of the interview protocol. This interview protocol will be the outcome of the desk research.

Table 1 and Table 2 show the potential respondents from the PaaS providers and PaaS (potential) costumer groups. The company names mentioned in the table do not represent the real respondents. Multiple consumer organizations from different industry sectors are going to be interviewed to gain more insight and construct validity. The interviews with these organizations are intended to collect data to answer mainly the research sub-question 1,2,3, and 4, as these research questions are aiming to treat the issues of PaaS in more general level.

Vendors Type Companies PaaS Provider Company 1 (e.g. Microsoft) PaaS Provider Company 2 (e.g. Google) PaaS Provider Company 3 (e.g. Salesforce)

Table 1 List of interview respondents from PaaS vendors

Clients Industry Sector Companies/Organizations Financial Services Company 6 (e.g. ABN Amro) Company 7 (e.g. ING) Telecommunication Company 8 (e.g. T-Mobile) Company 9 (e.g. ) Resources Company 12 (e.g. Shell) Company 13 (e.g. Essent) Products Company 14 (e.g. Ahold) Company 15 (e.g. TNT)

Table 2 List of interview respondents from potential PaaS customers

Page | 11

Chapter 1

For targeting the PaaS providers, we would aim to interview the marketing managers as they might have the most relevant information about the service offerings as well as the knowledge of the market overview. In gathering the data from (potential) PaaS consumers/clients, we would like to interview a person within each of the organization that has a strategic role in the company for IT related decision, for instance the manager from the IT department.

Furthermore, we will use the analyzed data from the interview results also to answer the sub- research question 5, which is related to the role of telecom companies in PaaS provision. Besides the potential benefit that telecom companies might have from being a consumer like the other organizations, telecom companies might also extract different kind of value from its core competency. Telecom companies in general are typically the one who provide Internet network. On the other hand, cloud computing service provision is directly dependable on the availability of the Internet network. We will analyze the value networks of the client’s business model to find a gap that can be filled-in by telecom company to deliver the “missing” role in the value network.

Additionally, in order to help answering the sub-research question 5, benchmark analysis on the existing business model for telecom companies from other countries will be taken into account. Given the information on what important factors are essential to adopt PaaS in the Dutch market from the previous sub questions, we will see whether the benchmarked business model is fit and applicable to the Dutch telecom companies.

1.4 Relevance

This project contains several relevancies. First of all, is for practitioners. There are two outcomes aimed for practitioners. The result of this research can be used as guidance on adopting and utilizing generic PaaS services in order to create a new business values. Another outcome is a recommendation for telecom companies whether to involve their business in PaaS services in order to create new revenue source.

The second relevancy is from academic point of view. This research contributes to the development of the theory of business model. Particularly, this research contribute to the development of STOF business model framework, especially concerning its application to the cloud-based services. Moreover this research would also give insight into cloud computing adoption theory, particularly in the context of PaaS adoption for clients.

1.5 Report Outline

Following the Introduction chapter, Chapter 2 will present discussion on the domain of cloud computing. Furthermore Chapter 3 consists of elaboration on theoretical background that is being used as theoretical framework in conducting this research. Accordingly, chapter four will present the result and analysis of the conducted interviews, which includes the interview methodology and the validation of the conceptual model. Chapter 5 consists of discussion about the roles of telecom companies in cloud computing world, and finally Chapter 6 will elaborate the conclusion and recommendation. The Figure 7 below depicts the construction of this report.

Page | 12

Introduction

Ch. 2 Cloud Computing (PaaS)

Ch. 3.1.1 y s r g

Ch. 3,1 Application of the o o t l c o

Related work on related works toward a d F o cloud computing client PaaS business h n t o e model i t M p

o 4 1 w . . d e 3 4

i Ch. 4.2 A

. . v r h h S Result and Analysis e a t C C a n I

P /

Ch. 3.2.4 d Ch. 3.2 n e o t

STOF model i t Business Model a r application to client a g d i l

theory e t

PaaS business a n I

(STOF model) V model

Ch. 5 Role of Telecom Companies in Cloud Computing Service Provision

Ch. 6 Conclusion and Recommendation

Figure 7 Report structure

Page | 13

Chapter 1

This page intentionally left blank

Page | 14

2 Cloud Computing Domain

This chapter will present discussion on cloud computing domain. First, general definition of cloud computing will be presented. Furthermore elaboration on the three core concept of cloud computing services is presented. Some general advantages and disadvantages will also be discussed, continued by general current market analysis and forecast. Due to the limited articles that discuss PaaS in particular, we will carry out general perspective of cloud computing throughout this chapter, in which additional focus on PaaS will be added accordingly.

2.2 Software-as-a- Service

2.1 2.3 2.5 2.6 2.7 Cloud computing Platform-as-a- Public and Private Advantages and PaaS Key definition Service Cloud Disadvantages Stakeholders

2.4 2.8 2.9 Infrastructure-as-a- Cloud market Conclusion Service overview

Figure 8 Structure of Chapter 2

2.1 Cloud Computing Definition

The origins of cloud computing concept can be traced back in the early 1960s with the added benefits of universal access that is enabled through the Internet (Hirzalla, 2010). The terminology of cloud computing still brings a lot of confusion among the IT practitioners at this moment. Some academicians like Buyya et al. (2009) defines cloud computing as “a type of parallel and distributed system consisting of a collection of inter-connected and virtualized computers that are dynamically provisioned and presented as one of more unified computing resource(s) based on service level agreements established through negotiation between the service provider and consumers”. The key points in this definition are that cloud computing puts emphasize on its accessibility through the Internet from centralized nodes (virtual data centers), its scalability, and also its characteristic of on demand provisioning.

Moreover, research companies such as Gartner, IDC and Forrester also have their own definition of cloud computing. Gartner (2010) defines cloud computing as “a style of computing where scalable and elastic IT-related capabilities as provided as a service to consumers using Internet technology”. IDC (2010b) describe simply cloud computing in terms of cloud services as “consumer and business products, services and solutions delivered and consumed in real-time over the Internet”. IDC further describes cloud services by defining its eight key attributes and its two deployment models. The key attributes are: shared (standard service), solution-packaged, self-service, elastic scaling, use-based pricing, accessible via the Internet/IP, standard user interface (UI) technologies, and published

Page | 15

Chapter 2 service interface/API. While the deployment models defined by IDC are: public and private. The terminology written here will be elaborated further in the following section. On the other hand, Forrester (2009) defines cloud computing as “a standardized IT capability (services, software, or infrastructure) delivered via the Internet in a pay-per-use and self-service way”. Forrester definition implies broader sense of cloud computing because it includes all kind of Internet-based services such as blog, wiki, social network, etc.

In general Cloud computing is promising to provide services to users without reference to the infrastructure on which these services are hosted, because cloud applications are accessed with the Web Service technology through the Internet. Instead of purchasing their own computer servers rack and operating themselves, companies could benefit from cloud computing mainly in reducing the cost (including reduced capital expenses), a low barrier to entry, and the ability to scale up as demand requires. Cloud computing adapts multi-tenancy architecture technology, therefore it really brings the advantage of economies of scale in terms of the provision of the data centers.

From the technical point of view, cloud computing is based on an architecture called multi-tenancy. Multi-tenancy principle allows single application software hosted in a cloud server to be used by multiple clients, so that no duplicates of the same software are needed for every customer. In terms of memory and processing overhead, this principle allows better utilization of the computing resources (Marston et al., 2011), thus enabling efficient scalability.

From financial perspective, cloud computing changes the point of view of IT service provision from big upfront investment into a predictable operational cost. The reason is because the typical fee incurred in consuming cloud computing service is based on pay-as-you-go scheme or monthly subscription.

Moreover, some academicians (Buyya et al., 2009) also see that the introduction of cloud computing is transforming the provision of IT services in a manner similar to traditional utility like water, electricity, gas, and telephony. In this manner, IT services are delivered and offered based on the users requirements and demands, regardless of their usage location and also the host location.

Furthermore, according to Reynolds and Bess (2009), the cloud is a means to an end, not an end in itself. They argue that big enterprises will use cloud computing technologies to add value or reduce the cost of their business operations.

From the elaboration we presented, we see that cloud computing is more specific than just a software applications that are accessible through the Internet. Therefore, within this research we adopt the definition of cloud computing that aligns with the definitions from Buyya et al. (2009), Gartner (2010), and IDC (2010b). We are putting our emphasize of cloud computing as on demand computing services that are delivered through the standard Internet interface, which has the characteristic of simple scalability and utility-based pricing.

Page | 16

Cloud Computing Domain

Figure 9 Three-layer cloud computing architecture (Lin, Fu, Zhu, & Dasmalchi, 2009)

Cloud computing itself cannot be seen as only one thing. It is constructed in some layers. Like what has been defined by Lin et al. (2009), cloud computing can be categorized into three architectural layers, as can be seen from Figure 9. These three layers represent the on-demand software application delivery layer on the top, a middleware layer of platform service that provides runtime environment for cloud applications, and flexible infrastructure of distributed data center services delivered via the Internet on the bottom layer. These three layers are also commonly known as Software-as-a-Service (SaaS), Platform-as-a-Service (PaaS), and Infrastructure-as-a-Service (IaaS) consecutively.

Moreover, Weinhardt et al. (2009) has developed a business model ontology specific for cloud computing services. This ontology, as depicted in Figure 10, classifies different business models for the three layers cloud computing architecture. This ontology can be used as a framework to characterized and classify cloud offerings. Weinhardt et al. argue that this ontology would be useful for cloud users, as well as cloud providers, to mapping cloud-based service products whether they fit to the concept of IaaS, PaaS, or SaaS. Thus, this classification would help both users and providers to identify their key customers and suppliers to realize the cloud-based services.

Similar to Lin et al. (2009), the cloud business model ontology (Weinhardt et al., 2009) defines cloud computing services in three layers, namely applications, platform-as-a-service, and infrastructure layers. The ontology then classifies every cloud computing service layers into two types of categories. He classifies Infrastructure or IaaS offerings into the categories of storage and computing business models. Storage business model here means providing remote web-based storage capabilities, where computing business model means providing a service of remote web-scale computing power.

Page | 17

Chapter 2

Figure 10 Cloud business model ontology (Weinhardt et al., 2009)

PaaS is classified into development platform and business platform. Development platform is a platform where developers could write and test their cloud-based applications, as well as uploading these applications to the cloud. Business platform on the other hand is a platform where development, deployment, and management of tailored business application are enabled.

Moreover, Applications layer (SaaS) is classified into SaaS applications and provisioning of basic on- demand Web services. SaaS applications business model means offering software application package through the Internet and accessible through a Web browser, meanwhile on-demand Web services business model means aggregating basic web-based information services that could be viewed through the Internet as well.

On the following sections, each of the cloud layers, namely Software-as-a-Service, Platform-as-a- Service, and Infrastructure-as-a-Service are going to be elaborated further in more details.

2.2 Software-as-a-Service

SaaS is what most people recognize from cloud computing in general. SaaS offers the flexibility of accessing applications through the web, without the need to install in the user’s local computer. SaaS concept is to enable accessing software application regardless of time and location. This cloud service offers complete functionalities ranging from office productivity applications (word processor, spreadsheet, etc.) to Customer Relationship Management (CRM) and Enterprise Resource Management (ERM) applications.

Weinhardt et al. (2009) argue that SaaS can be distinguished as two types: (1) SaaS applications and (2) provisioning of basic on-demand Web services. Examples of SaaS applications here are the web applications such as Google Apps (Google calendar, Gmail, Google docs, etc.) (Google, 2011),

Page | 18

Cloud Computing Domain

Microsoft Office Live, and Salesforce.com. Examples of on-demand Web services are financial web services hosted by Xignite (Xignite, 2011).

SaaS providers typically adopting pay-as-you-go pricing scheme in delivering their products to the consumers. This is allowing the customers to use the resource as many as what they really require to use. As discussed previously, the pricing scheme is changing the economics of IT system provisioning by eliminating the needs to have high upfront investment in IT infrastructures.

In addition, according to Gonçalves and Ballon (2011), SaaS represents one-sided market model in developing and delivering IT services. The service flows from the PaaS provider/vendor to the customers and the revenue stream flows simply the other way around.

Moreover, in IT industry some people might have confusion between SaaS and Service-oriented Architecture (SOA). Besides the acronym soup that may be indeed confusing, the confusion might also arise because SOA has the same characteristic of flexibility and typical web-based application environment. However, the difference is that SOA is a means of designing and building software; it is a software-construction model or design principles. Whereas SaaS is a means of receiving software from external parties to your business; it is a software-delivery model (Natoli, 2008). SaaS is developed based on SOA design principle.

2.3 Platform-as-a-Service

PaaS is a software development platform and runtime environment that is delivered through the Internet (Lin et al., 2009). Utilization of PaaS would enable (cloud) application developers to focus primarily on development cycle of new application, without worrying on the provision of the physical underlying infrastructure (Gonçalves & Ballon, 2011). PaaS offers the platform to design, develop, test, deploy and host a web application. This means that the primary goal of PaaS is to support SaaS applications.

PaaS can also be defined from the aspect of coordination of subscribers and providers of cloud-based services on a platform. The key offerings for PaaS can be an application components or building blocks for fast and easy creation of new SaaS applications (Rake-Revelant et al., 2010). PaaS could be used to develop an application on top of functionalities of an existing SaaS (e.g. Force.com), or to develop completely new web applications (e.g. Google App Engine).

According to Weinhardt (2009), PaaS can be further distinguished as (1) development platform and (2) business platform. Development platform is a platform where developers could write and test their application, as well as uploading their application code into the cloud. This written application can be accessible and used by users via the standard Web interface. The examples of these development platforms are Google App Engine and Microsoft Azure. Typically each of these platforms has their own software development kit (SDK) to help application developers to fully use their platform through some application programming interface (API). Each of the providers/vendors brings their own compatibility with the existing programming environment. For instance, Google App Engine supports Java and Phyton programming language (Google, 2011), while Microsoft Azure supports SOAP, REST, XML, Java, XML, PHP and Ruby protocols and language (Microsoft, 2011).

Page | 19

Chapter 2

Business platform is a platform where development, deployment, and management of tailored business applications are enabled. Force.com is an example of this business platform, where they offer their SaaS applications (Salesforce.com) of CRM and sales-related business applications along with additional add-ons application. These add-ons applications can be used to modify the generic SaaS application, and were built only for this platform upon the same technology that underlying the generic SaaS application.

Moreover, as described from one of the Accenture case study on Origin Digital company (Accenture, 2009), PaaS offers benefits such as: (1) real-time scalability, (2) reduced administration, (3) quicker time to market, and (4) reduced cost. The elastic capabilities the PaaS offered could also accelerate the development time and reduce the application support cost. This means that developers can build web-based applications and run it in the cloud in a matter of weeks instead of months. Furthermore, as the user of an application grows, developers do not have to worry about the system scalabilities issues.

Similar to SaaS model, PaaS providers typically charge their customers based on subscription fee or pay-per-use scheme. This means that consumers will only pay the resource as much as they use it only.

In addition, in contrast with SaaS, PaaS represents two sided market model according to Gonçalves and Ballon (2011). This means that PaaS as a platform mediates the providers (supply side) and the customers (the demand side) by creating value and extracting revenues from both sides.

2.4 Infrastructure-as-a-Service

IaaS is a service that offers computer infrastructure via the Internet. The computer infrastructure here includes virtual computers, servers, and storage devices. Weinhardt et al. (2009) argue that IaaS business model can be classified into two categories: storage and computing. On one side, IaaS service can be understood as a service that provides remote web-based storage capabilities. This kind of service is what is shown by Amazon Simple Storage Service (S3) (Amazon Web Services, 2011b). On the other hand, IaaS could also provide remote web-scale computing power as a service. This type of service is also adopted by Amazon by offering their Elastic Compute Cloud (EC2) service (Amazon Web Services, 2011a). The essence of IaaS services is actually transferring the benefit of scale (of the IT infrastructure) towards the consumers.

Similar to SaaS and PaaS model, IaaS also adopt pay-as-you-go pricing scheme or monthly subscription fee. This means that customers only pay for what they use, in a matter of uploaded or downloaded data from the cloud services, or in terms of the amount of data stored in the cloud servers.

Page | 20

Cloud Computing Domain

Table 3 below shows the summary of the three cloud computing service categories. The table shows what are the difference between them in terms of functionality and typical services.

Cloud Service Functionality Type of Cloud Services Category Enterprise Resource Management Applications Supply Chain Management Applications Operation and Manufacturing Applications SaaS Software applications Engineering Applications Customer Relationship Management Applications Collaborative Applications Application (SaaS) Development Software Applications Server Middleware Application Information & Data Management Paas development and Enterprise Portals deployment Integration & Process Automation Middleware Other Applications Development and Deployment System and Network Management Software System infrastructure Security and Storage Software IaaS software and storage Servers Infrastructure Storage Infrastructure

Table 3 Cloud services categories (adapted from (IDC, 2010a))

2.5 Public and Private Cloud

There has been a lot of discussion about the terminology of public and private cloud. Public or Private cloud distinguish different model of cloud computing services deployment. IDC (2010b) defines Public cloud services as a cloud service that “open to a largely unrestricted universe of potential users; designed for a market, not a single enterprise”. Where Private cloud is defined as a cloud service that “designed for, and access restricted to, a single enterprise (or extended enterprise); an internal shared resource, not a commercial offering”. It can be understood that in Private cloud, the IT organization of a company is the “vendor” of cloud services with respect to its users.

The main difference here is that in Private cloud, the organization still owns the infrastructure. So the feature of flexible scalability here is limited. Whereas the benefit of having low up-font investment cost is also diminish. The main reasons of the existence of Private cloud is due to security issue, where some organizations do not have the confidence and trust to host their data on public server like what happen with Public cloud. IDC (Gens, 2009) furthermore sees that Private cloud is a step to transform IT organizations into service-oriented IT delivery providers.

2.6 Advantages and Disadvantages

In principle, there are several key advantages that cloud computing offer compared to traditional IT system (Marston et al., 2011):  Lower the cost of entry for companies that want to benefit from compute-intensive business applications

Page | 21

Chapter 2

 Lead to immediate impact and faster time to market as no upfront capital investment is needed  Lower IT barriers to innovation  Easier to scale application services  Lead to possible new classes of application and services delivery that were not possible before. For instance a location-, environment-, and context-aware services

Several authors like Sultan and Marston et al. (2011) have indicated that cloud computing solutions are very attractive, especially for startups and SMEs. However, some (Marston et al., 2011; Sultan) have expressed their hesitation whether cloud computing is beneficial for large enterprises at this moment. They argue that cloud computing services might not be cost-effective for large enterprises, especially for those enterprises that have achieved significant and state-of-the art efficiencies from their computing operations. SMEs see cloud computing as an opportunity to be more efficient and bring down total cost ownership of their IT resources, because they see that IT-related activities are not their core competences.

Nevertheless, despite all the benefits that cloud computing could offer, there are some concerns and challenges regarding the adoption of the technology. Sultan argues that the most important concerns are related to: control of the IT system and data, vendor lock-in, performance, latency, security, privacy, and reliability. It has been assumed in several articles (Marston et al., 2011; Sultan, 2010) that users are quite reluctant to migrate their applications to the cloud due to the lack of control feeling on their critical data. Dependency of the system with the Internet network and hesitation of the related privacy and security (Subashini & Kavitha, 2011) to establish secure Internet connection are also other factors that are argued to slow down the cloud computing adoptions. Therefore, to summarize, there are some disadvantages of cloud computing for enterprise companies as the following:  Cloud computing might not be cost-effective for large enterprises  Lose control of the data  Dependent on external party on the performance, security, and reliability of the IT system  Exposed to vendor lock-in

Furthermore, Reynolds and Bess (2009) argue that there are several strategies in adopting cloud computing. These strategies depend on two aspect: 1) whether the utilization of cloud computing is intended to reduce cost or add value, 2) whether the IT organization of a company or the business unit is the one who will benefit from the cloud implementation. Based on these aspects, four generic strategies are presented, namely: the replacement strategy, the augmentation strategy, the integration strategy and the exploitation strategy. Each of these strategy would result on different benefits from the adoption of cloud computing.

2.6.1 PaaS advantages over SaaS When we see SaaS offerings, typically SaaS is delivered as an off-the-shelf application to the customers. SaaS is typically offered as off-the-shelf Customer Relationship Management (CRM), Enterprise Resource Planning (ERP), Human Resource (HR), or Supply Chain Management (SCM) applications. The main application of SaaS is typically ready to use with minimum customization opportunities.

Page | 22

Cloud Computing Domain

On the other hand, PaaS brings different perspective of cloud service offerings. As described before, PaaS is a development environment of cloud applications (SaaS). Therefore, customers have full flexibility to build any kind of cloud applications in PaaS, which offer any organizations full customization of their own cloud-based business-related applications. Moreover, with PaaS, customers have more control over their own applications than SaaS.

2.7 Cloud Security and Reliability

Security is one of the major issues that reduces the adoption growth of cloud computing (Gartner, 2010). Different security risks pose a threat to the cloud environment. Subashini and Kavitha (2011) note that the security challenges include but not limited to accessibility vulnerabilities, virtualization vulnerabilities, web application vulnerabilities such as SQL (Structured Query Language) injection and cross-site scripting, physical access issues, privacy and control issues arising from third parties having physical control of data, issues related to identity and credential management, issues related to data verification, tampering, integrity, confidentiality, data loss and theft, issues related to authentication of the respondent device(s) and IP spoofing. Some of these fundamental security challenges can be categorized into four categories of data storage security, data transmission security, application security, and security related to third party resources (Subashini & Kavitha, 2011).

Security controls in cloud computing are similar to the security controls in any IT environments (CSA, 2009). These controls are implemented in several layers ranging from the facilities (physical security), to the network infrastructure (network security), to the IT systems (system security), all the way to the information and applications (application security). As argued by the Cloud Security Alliance (CSA) (2009), no list of security controls can cover all cloud service deployment options (including the different service models of SaaS, PaaS, and IaaS). It has been suggested by CSA that organizations should adopt a risk-based approach to move to the cloud and selecting security options.

Different cloud service models (i.e. SaaS, PaaS or IaaS) implies different level of security issues and responsibilities for both the cloud providers and the clients. In the context of SaaS service for instance, the provider is responsible for all the security controls from the physical security until the application security level. In contrast, within the IaaS service context, the cloud provider responsibility is limited only in the physical and virtualization environment. The clients are in turn responsible for the application and data security layers. While PaaS offering lies in a balance somewhere in between when it comes to security. PaaS provider is responsible to secure the platform, but securing the applications developed against the platform and developing them securely, both are the responsibility of the clients (CSA, 2009).

Cloud computing is also argued to be “hackers dream” (Visscher, 2011). Hackers have the same advantages as users. They can use “unlimited” computing power at the same time from cloud service to attack. Cloud computing can also be used to do clickfraud, because with cloud computing one can use as many computers and IP-addresses as possible. Nevertheless, it is a complex issue to prevent. Due to privacy issue, it is hard to do thorough and complete check of what goes on in the cloud. Also, it is typical that cloud service providers do not know who their customers exactly are due to “all- online” subscription procedure (Visscher, 2011).

Page | 23

Chapter 2

With regards to reliability, performance concerns is argued to stop some companies from using cloud computing for transaction-oriented and other data-intensive applications (Leavitt, 2009). Clients who are long distance from cloud providers could experience latency, particularly if there is a lot of traffic and their code is not optimized for efficient transmission (Leavitt, 2009). Moreover, cloud computing track record have not been perfect as well. For example, Salesforce.com and Amazon experienced outage for several days in 2008 (Leavitt, 2009). Even though some compensations from the vendors were offered, this issue still of clients’ concern to run their business in the cloud.

2.8 Cloud Key Stakeholders

Marston et al. (2011) argue that cloud computing changes the roles of the traditional stakeholders of computing setup, and adds new ones. These stakeholders are not restricted to providers and clients, but also regulators and enables, due to the nature of the delivery model of the service. Marston et al. (2011) describe these stakeholders in the context of cloud computing in general. However, we find that this analysis is also applicable in the specific context of PaaS because of the similarity in the nature of the technology.

In the context of PaaS, the clients or PaaS consumers are effectively subscribers, or users of the PaaS service. These clients can be corporate entity, which can be represented by its IT department, or independent software vendors (ISVs). The providers are the PaaS service owners that operate the cloud computing system to deliver to third parties. The providers will be responsible of all the infrastructure of the cloud service as well as their maintenance. Next are the enablers. Marston et al. (2011) use the term “enablers” to describe those organizations that will sell products and services that facilitate the delivery, adoption, and use of cloud computing. Enablers include consultancy firms that provide service in cloud implementation and adoption. Finally, regulators are included as the key stakeholders because regulators are the ones that make regulations that are relate to the cloud computing business. For instance, European Union has a regulation that requires a company to know where the data in its possession is at all times. This is in conflict with the nature of cloud computing where the hosted data could reside in a place that the client organization might be unaware of.

In addition, we would like to add another key stakeholders in the provision of PaaS offerings. They are the network providers, both fixed ( broadband operators or internet service provider) and mobile (MNO, mobile network operator). Our argument is based on the core resource and capabilities of these stakeholders, which is to deliver the Internet network access. Any cloud computing services are dependent on the availability of the Internet access to work. Therefore, we argue that internet service provider hold a key role on cloud computing ecosystem.

2.9 Cloud Market Overview

Based on the market analysis performed by IDC Research (2010b), cloud computing services exceeded $16 billion in revenue worldwide in 2009. This market is also forecasted to grow and reach $55.5 billion of revenue in 2014. This estimated growth rate is five times larger than the estimated growth rate of traditional IT services.

Western Europe in particular share around $3.4 billion worth of market in 2009, and estimated to grow to $17.4 billion in 2014 (IDC, 2010a), with UK as the biggest market. Moreover, the Netherlands

Page | 24

Cloud Computing Domain shares $161.2 million worth of market in 2009, and estimated to grow to $883.1 million worth of market in 2014. Thus with a compound annual growth rate (CAGR) of 40.5%, the Netherlands remains to be a potential market for public cloud services.

Furthermore, IDC Research (2010a) also shows that out of the three cloud services segments, SaaS is the most mature segment. However, IaaS is the segment that has the biggest share in the market with over than 50% share, followed by SaaS in the second place. Even though PaaS is the segment with the lowest market share at this moment, it has been projected by IDC Research (2010a) to be the segment that has the fastest-growing rate for the coming years. Figure 11 below shows the Western European cloud service market by segment in 2009 and the projected market in 2014.

2009 2014 Total = $3,446 million Total = $17,404 million

PaaS PaaS 8,6% 14,9% SaaS SaaS 31,7% 39,7%

IaaS 51,7% IaaS 53,4%

Figure 11 Western Europe cloud service market by segment in 2009 and forecasted market in 2014 (IDC, 2010a)

In conclusion, cloud computing is a promising enabling technology and has many opportunities to create new values for any businesses. PaaS segment in particular has big opportunity as its market is forecasted to grow rapidly in the coming years. However, to the best of the author’s knowledge, there are limited discussion on guidance and framework on how to adopt and utilize cloud-based service in order to successfully create value to the business. This is the gap that this research would like to explore.

2.9.1 Dutch PaaS Market Overview There are several providers of PaaS available in The Netherlands. The main players include Salesforce.com (SFDC) with its Force.com, Windows with its Azure, and Google with its App Engine. However, there are some smaller players such as Caspio, Cordys, Engine Yard, Heroku, LongJump, OrangeScape, and WorkXpress. Each of the providers has their own market strategy in offering their PaaS offerings. Table 4 below shows the overview of the different PaaS vendors.

Page | 25

Chapter 2

PaaS Providers/ Description Vendors Google App Engine enables you to build web applications on the same scalable systems that power Google applications. Partnering with Salesforce.com to give Google App Engine developers using Google App Engine native access to the capabilities of Salesforce.com platform including multi-tenancy, authentication, scalability, database, and workflow so that their software runs in a business environment. Platform the enables developers to create and deliver any kind of business application, entirely on-demand without software. Adopting their own proprietary SFDC Force.com standard called Apex. Currently with a partnership with Google App Engine. Through this partnership Salesforce.com developers get native access to Google's Bigtable distributed-storage system and the ability to talk to apps in Python. Platform hosted in Microsoft data centers. Supports a consistent development Microsoft Azure experience via integration with Visual Studio including .NET applications. Also supports third party tools such as Eclipse, Ruby, PHP, and Phyton. Offer solution focusing on the enterprise IT organization. Offer tool to build Web- based mashup applications. The types of applications deemed appropriate include “situational applications” and those that support dynamic business processes. Cordys Cordys also offers an online marketplace to sell applications to end customers, and it supports ISVs with a program that covers partner on boarding, training and support, as well as assistance with pricing models. Platform that targeting replacement of traditional development with a point-and-click platform for building applications. ISV market is not a current focus. Hosts the Caspio platform in a commercial data center facility based in the US. The main emphasis for Caspio is a high-productivity hosted environment for the development and deployment of two-tier, database-centric Web applications. Platform that manages organizations Rails application deployments, from EngineYard small/medium applications running on 3 slices to large applications that automate business operations. Ruby on Rails Platform with instant deployment. When developers create or import an app with Heroku, it’s already live on the web. No configuration or deployment Heroku necessary. Heroku is built on Amazon EC2, thus the apps have access to unlimited computing power. On-deman platform for creating and delivering business applications to manage data, streamline collaborative processes and provides actionable analysis. This robust PaaS LongJump solution reduces the time and costs of developing and delivering secure, database- driven applications, while also providing extensive features round security access, data analysis and visualization, and process automation – all on the web. Platform that position itself as offering a process-centric solution, suitable for situational applications and those with unpredictable volumes or spiky usage. OrangeScape Focuses firmly on reducing the time for its customers to develop applications, as well as the flexibility of deployment. Possible to deploy its service on more than one cloud infrastructure provider. Easy to program by means of visual user interface. Platform that focus on the creation of customizable solutions. Supports nonprogrammers via a drag-and-drop development model. Differentiate itself from integration with third-party services like FedEx. Have an appstore the offers access to WorkXpress its own CRM and ERP offerings. Provide choice of where the user’s applications is hosted, thus supporting deployment where data center location is important. Helping ISCs with market positioning. Support a range of multitenancy choice.

Table 4 Overview of PaaS vendors

Some of the providers carry proprietary programming language, such as Salesforce.com (SFDC). While the others adopt more common programming language like Microsoft Azure that accommodates many programming language such as .Net framework, Java, PHP, Ruby, etc. Some of them have a marketplace within their PaaS offerings, such as SFDC with its AppExchange and Azure with its App Market, that allows third party applications to be used as building blocks of a custom- made application. The PaaS offerings also differentiate in the degree of its complexity in programming. Some need low level programming skill to be able to develop application on the

Page | 26

Cloud Computing Domain platform, some uses a visual user interface approach that allows building application by means of drag-and-drop feature. Furthermore, in general PaaS providers employ their own data center in providing the service, but some providers enable customers to choose which infrastructure provider (IaaS provider) they would like to use to host their applications that are built on the platform.

2.10 Trends of Telecom Companies Activities in Cloud Computing

Within this section, we will discuss the global trends of telecom companies (telcos) activities in cloud computing. We will include telecom companies from Europe, North America and Japan.

Around the world, some telecom companies (both MNOs and Broadband operators) have started to put their foot on the cloud computing domain. As said by Evan Kirchheimer, practice leader for Enterprise Services at Ovum, telecom companies’ control over the network on which cloud services are offered gives telecom companies a “compelling advantage” (Weaver, 2011). Below we will elaborate the trends of roles that telecom companies have been playing in the cloud computing domain.

Some believe that telecom companies’ opportunity to monetize the cloud lies in being the cloud provider themselves (IDC, 2009; Weaver, 2011; White, 2011). Elizabeth White (2010) argues that telecom companies play an important role in bringing enterprise-grade cloud computing solutions to a broader base of users. She also argues that telecom companies can monetize the cloud computing as an extension of their managed networking solutions (i.e. VPN service), by offering on-net cloud computing capabilities backed up by end-to-end service level agreements (SLA). The quality of cloud computing is heavily dependent on the network (both whether it is public Internet or VPN). Therefore cloud computing services are a good fit for telecom companies since they own and manage network infrastructure and therefore can guarantee high quality end-to-end service (InformationAge, 2011). Additionally, telecom companies already have an operating model that is ready to support utility-based services such as cloud computing, because telecom companies are already used to bill their customers for the network service based on the usage and monthly subscription (InformationAge, 2011).

One of the examples, British Telecom (BT) in the UK has a number of cloud-related services on offer. The offering covers IaaS, SaaS and “Communication-as-a-Service” approach (InformationAge, 2011). BT has introduced BT OneVoice, a hosted unified IP-based communications service, as a communication-as-a-service offering. BT launched its Virtualized Data Centre (VDC) last year to allow their customers to provision infrastructure and on a month-by-month basis as an IaaS offering. BT has also various SaaS offerings in its portfolio that are focus on specific industry, such as BT Radianz (a hosted information aggregation platform for financial traders) and BT Expedite (a hosted retail application suite).

In general, other telecom incumbents in the world also have entered the cloud computing market as cloud providers. The Table 5 below shows a selected overview of cloud offerings from incumbent telecom companies in Europe, North America, and Japan. Further details of the cloud offerings can be found in the Attachment A.

Page | 27

Chapter 2

Telcos SaaS IaaS PaaS BT + + - Orange Business Service + + + Dutsche Telekom/ T-Systems + + + TellaSonera - + - Belgacom - + - Telefonica + + - AT&T + + + (planned) Verizon Business + + - Telstra + + + (planned) SK Telecom - + + NTT - + - Vodafone + - +

Table 5 Categorization of Cloud Offerings from Telcos (adapted from Buvat and Nandan (2010))

From the overview shown above, we can see that telecom companies’ offerings in cloud computing domain are centered around IaaS and SaaS with limited focus on PaaS. It has been argued that current IaaS offerings from these telecom companies such as BT (Virtual Data Center), Orange (Flexible Computing), Verizon (Computing as a Service), and Deuteche Telekom (Zimory) are as competitive as those from other IaaS providers like Amazon (Buvat & Nandan, 2010). As IaaS is not of our focus in this research, we will not discuss Telcos role in the IaaS domain in depth.

On the other hand, telecom companies have also offer a host of SaaS applications in partnership with ISVs. For example, NTT offers cloud-based security solution, and T-Systems offers SAP from the cloud. Other SaaS offerings are taking the form of communication and collaboration software such as BT OneVoice (InformationAge, 2011).

Nevertheless, this concept of being the cloud provider can also be approach from a different angle. Gonçalves and Ballon (2011) showed that telecom companies particularly MNOs have been experimenting with PaaS concept by being a PaaS provider for the development of mobile applications. Some examples are Vodafone Betavine and Orange Partner. Through this PaaS offering, MNOs expose their telecommunications network functionality through APIs (Application Programming Interfaces) to third parties (i.e. Independent Software Vendors (ISVs)). By means of the accessibility to the APIs, the third party developers are allowed to use limited number of network capabilities such as messaging, location, rich presence, charging, etc. These third parties then will take up the service innovation, bring services to market more rapidly, increase customer intimacy through personalized services towards a large MNO’s customer base, while MNO’s benefit from a revenue sharing model (Gonçalves & Ballon, 2011). However, Gonçalves and Ballon (2011) argued that the multiple approaches offered by multiple MNOs are still lack of standardization in programming language, network APIs, and application shops. This fragmentation can slow down the adoption of the offered PaaS solution. Thus the openness of such a PaaS solution is understood to be one of the success factors in promoting this approach as a new revenue source for telecom companies.

Page | 28

Cloud Computing Domain

Another way for monetizing cloud computing by telecom companies is by offering complementary service, or a so-called “telecom cloud services”. Verizon has been adopted this approach in the US. Verizon announced its joined telecom cloud services strategy with Terremark, an IaaS provider. Verizon will offer more security, reliability and advanced services to Terremark’s cloud computing service (Scarpati, 2011). Kerry Bailey, president of Terremark Worldwide Inc., believes that Verizon as a telecom company has built their reputation on the security and reliability. These issues are something that Terremark as a pure-play cloud provider has failed to offer (Scarpati, 2011).

Moreover, another approach to benefit cloud computing domain has been adopted by telecom companies in Japan and UK. NTT Docomo, a telecom company in Japan, and Salesforce.com (SFDC), an incumbent cloud service provider that provide PaaS and SaaS services (Salesforce.com, 2010). They arranged a strategic partnership in delivering the cloud service. In this case, NTT Docomo acts as a mediator between SFDC and end-customers. NTT Docomo also acts as a broker between various ISVs that have applications running in SFDC platform, with the end-customers. NTT Docomo take care of billing, customer account management and security of the provision of SaaS and also do the marketing and sales of SFDC and any registered applications towards their own customer base. As a company, NTT Docomo earns commissions as incentives.

This kind of partnership can also be found between BT and Salesforce.com (SFDC) in the UK (Salesforce.com, 2008) (Meyer, 2008). As the news quoted, “BT has signed a deal with Salesforce.com that will see the software-as-a-service company’s customer-relationship-management software sold through BT Business” (Meyer, 2008). For BT’s customer, they can find the CRM service as part of their supplier’s BT Applications portfolio.

Broadband operators are also believed to play critical role in Microsoft’s Office 365 cloud strategy. Microsoft expects Broadband operators (i.e. Telefónica, Telstra, Vodafone Group) to be a primary link to SME market (Wilson, 2011). Telecom companies, such as Broadband operators, have been having relationship with SME companies for many years, which involve billing relationships and customer relationships. Therefore Microsoft is willing to collaborate with local telecom companies to exploit this resource in order to build the markets of their SaaS product. As explained by Wilson (2011), part of the collaboration is the integration of the billing of Office 365 into the network operator’s billing systems to produce a single bill and also customer support from the telecom companies’ customer service departments. This approach will give the client companies better awareness and a lot of simplicity.

There are several benefits that can be gained from these kinds of partnership between telecom company and cloud vendors. The cloud vendor in one hand can make use of a large customer base to market their product with a trusted marketer, while on the other hand the telecom company can create a new business value from cloud service provisioning and thus generating new revenue source from their existing resources, which are their customer base and customer relation. From the customer perspective, customers can make use of their existing customer relationship, such as billing relationship, in order to use and consume easily some new services from the cloud.

In conclusion, there are several trends of telecom companies’ activities in the cloud computing domain. In general telecom companies are acting as Cloud service provider and Cloud service broker. With regards to the role of cloud service provider, mostly the cloud services are offered by incumbent telecom companies that have business in delivering broadband service as well as mobile

Page | 29

Chapter 2 service (i.e. BT, Orange, Deutsche Telekom). However, with regards to PaaS, the services offered by telecom companies are mainly within the mobile domain. The PaaS service is offered to give access to mobile network APIs and to develop mobile applications.

2.11 Conclusion

This chapter gives us clearer understanding of the cloud computing domain, especially about Platform-as-a-Service (PaaS). So far, this chapter has helped us to answer the following sub-question 1 as mentioned in Chapter 1.

“Who are the key stakeholders in PaaS service provision in the Netherlands? What kinds of providers are offering PaaS service and what kind of organizations are the (potential) PaaS customers?”

In answering the question, first, we have presented related works that discuss the key stakeholders of cloud computing in general. There are providers, enablers, clients/consumers, regulators, and network providers. These key stakeholders are tentative answer to our question. More definite answer, particularly in the field of PaaS in the Netherlands, will be explored though the analysis of the interview results.

Secondly, subchapter 2.8 has presented the available PaaS providers. Each of the providers has their own offering characteristics. The offerings differ in terms of user interface, pricing, standards, and technical architecture. Four providers have their in-country representative in the Netherlands, they are: SFDC, Microsoft, Google, and Cordys. The other providers are based in North America and India, without an official representative in the Netherlands.

Furthermore, we were not able to answer what kind of organization are the (potential) PaaS customers yet, as the information available in the literature is limited. We aim to learn more of who are the PaaS customers throughout our interview result analysis later on.

In conclusion, cloud computing is a promising enabling technology and has many opportunities to create new values for any businesses. PaaS segment in particular has big opportunity as its market is forecasted to grow rapidly in the coming years. However, to the best of the author’s knowledge, there are limited discussion on guidance and framework on how to adopt and utilize cloud-based service, particularly PaaS, in order to successfully create value to the business. This is the gap that this research would like to explore.

One point to keep in mind is that PaaS is built to support cloud applications like SaaS. Therefore, PaaS cannot be discussed in isolation of SaaS. We will carry out this perspective throughout our discussion in this report

Page | 30

3 Theoretical Background

Following the domain background presented in Chapter 2, this chapter concerns with the theoretical background of the thesis. This chapter aims to identify and conceptualize the PaaS adoption factors, which are the factors that influence client’s decision in designing its business model (BM) to adopt and utilize PaaS technology. What we mean by client business model in our discussion is the business model that client design to offer a (cloud) service towards its end-users by means of utilizing PaaS from certain PaaS provider. Since the objective of PaaS is to support cloud applications (SaaS), we will analyze the cloud applications that client wants to offer to its user. What we mean by cloud applications here are the business software application that are (or will be) migrated to the cloud and built on PaaS. Then, we want to understand how PaaS impacts the business model design for those cloud applications.

We will begin by presenting related works on cloud computing adoption in Subchapter 3.1. These works will give us insight on client adoption strategies and characteristics to adopt cloud computing in general. We will then identify the important lessons we can take to be applied in PaaS domain specifically.

Afterwards, we will discuss business model theories and concept in Subchapter 3.2. We will present discussion on business model definition, including elaboration on the STOF model. We will then use business model theory and apply it to analyze client business model in utilizing PaaS. From the analysis, we will identify the issues that are relevant to take into account in designing client business model in utilizing PaaS.

Regarding the client perspective in designing its PaaS business model, there are two important points of views involved: towards the end-users, and towards the PaaS provider. Thus we can take two distinct standpoints in our discussion: the standpoint of client as PaaS customer, and client as SaaS service provider that utilize PaaS in order to deliver the service. The standpoints are depicted in, as Figure 12 below. The first standpoint concerns with the value that the client perceives from PaaS offerings from the PaaS providers, while the second concerns with the value that client wants to propose and deliver towards its end-users.

Client as PaaS customer Client as SaaS service standpoint provider standpoint PaaS Provider Client End-users

Figure 12 Two standpoints of the client PaaS business model

Then, we will formulate adoption factors based on the business model concept and the lessons learnt from the related works on cloud computing in Subchapter 3.3. We will end this chapter by presenting a conclusion in Subchapter 3.4. The structure of this chapter is shown in the following figure.

Page | 31

Chapter 3

3.1 3.1.1 Related works on Application of the related cloud computing works toward client PaaS adoption business model 3.3 3.4 Integrated PaaS Conclusion Adoption Factors 3.2.4 3.2 STOF model application Business model to client PaaS business theory model

Figure 13 Structure of Chapter 3

3.1 Related Works on Cloud Computing

Several authors have discussed the issue of general cloud computing adoption from the perspective of enterprise client. These related works on cloud computing will be discussed and presented in this Subchapter. However, none of the literature is focusing on PaaS specifically. We will use these related works as starting point to gain better understanding of the possible adoption factors that might be relevant for PaaS.

We will begin with the work by Reynolds and Bess (2009). They discuss cloud computing adoption strategies by enterprise client. They classify four strategies based on two core principles. First, whether the reason to adopt PaaS is to reduce cost or to add value. Second, who will benefit from the use of cloud computing, whether it is the IT organization of the client, or the business unit. The four strategies are depicted in the table below.

IT Department Business Unit Add Value Augmentation Strategy Exploitation Strategy Reduce Cost Replacement Strategy Integration Strategy

Table 6 Cloud computing adoption strategies (Reynolds & Bess, 2009)

Those mentioned strategies are defined as the following: - The replacement strategy: uses the cloud to replace or re-hosting IT systems. - The augmentation strategy: uses the cloud to expand the capacity or reach of existing IT systems - The integration strategy: uses the cloud to create more efficient and effective connections within organization as well as with clients and suppliers - The exploitation strategy: uses the cloud to generate new revenue streams by leveraging existing corporate intellectual property and information

Moreover, Misra and Mondal (2011) discuss four adoption factors that an enterprise client should take into account before adoption cloud computing. Those four factors define the organization characteristic of the client that wants to use cloud computing, and also define the cloud application characteristic that a client wants to build. Those factors are: (1) the size of the company’s IT resources; (2) the utilization pattern of the resources; (3) sensitivity of the data they are handling; and (4) criticality of work done by the company.

Each of the mentioned factors above can be broken down as the following. First, the size of the IT resources of a company is influenced by the number of servers, the size of customer base (end-

Page | 32

Theoretical Background users), the annual revenue from IT offerings, and the number of countries IT resource is spread across. This factor defines the characteristic of the client organization that is willing to use cloud services. Second, The utilization pattern of the resources can be determined by average usage, peak usage, and amount of data handling/transaction done. Third, the sensitivity of the data handled concerns with the issue of data lock-in in the case of outages and data availability in the long run in case the cloud providers goes out of business. Finally, the criticality of work done by company concerns with how stringent is the needed requirements to deliver the application via the cloud. The more critical the application is, the more demanding are the requirements. These last three factors define the characteristics of the applications that client wants to migrate to the cloud.

3.1.1 Application of the related works toward client PaaS business model Those strategies discussed by Reynolds and Bess (2009) give us insight of the importance of goals in PaaS adoption. Also, the work give us insight on the type of goals that client might have by creating cloud applications. The goals could be based on the principle to reduce cost, or based on the principle to add value to the client’s business. Those strategies also give us insight on business model design issue of the actor that will benefit from the cloud applications, whether the IT department or the business unit of the client. The applicability of the theory is summarized in the box below.

Possible reason/goal to utilize cloud applications: - To reduce cost - To add value

Possible design issue - Who benefit from the cloud applications from client perspective (IT department or Business Unit)

Table 7 Applicability of cloud computing adoption strategy (Reynolds & Bess, 2009) towards Client PaaS business model

Furthermore, the work by Misra and Mondal (2011) about cloud computing adoption factors gives us insight on the potential business model design issue in creating cloud applications. First we have the characteristic of the adopting organization. Secondly, we have the characteristic of the applications that are to be migrated to the cloud. The applicability of the theory is summarized in the box below.

Characteristic of organization that want to use cloud - Size of customer base (end-users) - Number of servers - Number of countries IT is spread across - The annual revenue from IT offerings

Characteristic of applications to be migrated to the cloud - Utilization pattern of IT resources (average and peak usage, amount of data handling/transaction done) - Sensitivity of data handled (in case of outage and unavailability of access) - Criticality of the application

Table 8 Applicability of cloud computing adoption factors (Misra & Mondal, 2011) toward client PaaS business model

Page | 33

Chapter 3

The characteristics of applications to be migrated to the cloud can be reflected as design issues of client business model in utilizing PaaS, because the characteristic of the application can be influence by the client. In defining the type of cloud application the client has to take into account the issues of sensitivity of data handled, the criticality of the application towards the business, and the utilization pattern of the application. The degree of how critical the application and how sensitive the data to be handled will influence the technical requirement of the application service.

Moreover, the characteristics of organization that want to use cloud are not relevant to be included as business model design issues, because these characteristics are not variables that can be influenced by the clients. These characteristics are the given state of the organization/client.

3.1.2 Conclusion We have presented related works on general cloud computing adoption factors and adoption strategy. These works have given us insight on possible goals to utilize PaaS, and possible design issues of the cloud application service. Characteristics of the applications to be migrated to the cloud were also identified to be of relevance in discussing client business model to utilize PaaS. These identified issues are going to be combined with design issues extracted from business model theory in order to construct a framework of PaaS adoption factors at the end of this chapter.

3.2 The Concept of Business Models

Business model is one approach to capture and create new business value for an organization in general. Business model perspective is able to give indications on how organizations would propose their service value and how customers would perceive value from a service. This perspective is aligned with our objective to understand how client organizations can successfully capture and create business value by utilizing PaaS. Business model theory would be able to give us insight on how clients would perceive and capture the value from PaaS offering, and how clients would create and offer their services that are built on PaaS towards their end-customers. Therefore we will study business model concept in this Subchapter. We will then apply the business model concept to analyze PaaS adoption factors from the client perspective.

The focus on client perspective is a reason that makes us not to use other theory perspective such as multi-sided platform theory (Gawer & Cusumano, 2002b). Even though PaaS, by its name, is a form of platform and has the feature of multi-sided market in its marketplace (in some PaaS products), we find that multi-sided platform theory is not capable of analyzing the adoption factors of PaaS from the client perspective. The multi-sided platform theory concerns more about platform leadership, which is more focus on the provider perspective.

Moreover, multi-sided market theory concerns with interdependencies of multiple actors around the platform to innovate. The utilization of the platform is highly dependent on the third parties involvement. Typically this is realized in a form of a marketplace. On the other hand, PaaS is a development platform of cloud application, in which the users can do innovation on cloud application without the involvement of third party developers. The availability of marketplace in PaaS is just a complementary feature. Without the existence of a marketplace, PaaS is still capable to work as cloud application development platform and deliver its service. The nature of PaaS is not solely a

Page | 34

Theoretical Background platform to connect multiple actors together. Therefore we found that multi-sided platform theory is not relevant as theoretical background in helping us fulfilling our research objective.

Within this section, we will first discuss the concept and definitions of business model in section 3.2.1. After this we will discuss the STOF model, a business model framework that focus on service design. We will then apply the STOF model to PaaS domain from client perspective in section 3.2.2.

3.2.1 Business Model and Strategy Business model terminology is often misunderstood with the term “strategy”. People often use it interchangeably. However some scholars argue that those terms are different in meaning. Seddon and Lewis (2003) distinguish strategy as more concrete and rooted in the firm’s competitive environment, where they define business model as an abstract representation of a firm’s strategy. According to Casadesus-Masanell and Ricart (2009) business model is about “the logic of the firm, the way it operates and how it creates value for its stakeholders”. Whereas strategy refers to “the choice of business model through the firm will compete in marketplace.

When talking about which comes first, strategy or business models? Seddon and Lewis (2003) argue that combination of business models could be used for “designing” strategy. Thus they believe that business model comes first than strategy. In contrast, Casadesus-Masanell and Ricart (2009) argue that a firm’s business model is “a reflection of its realized strategy”. Thus they believe that strategy comes first than business model.

From the discussion presented above, we understand that strategy reinforce business models. The viability and the success of a business is not only determined by a good business model, but also by having a good strategy to stand out in the competitive market. Several companies might have similar business model to capture some business values, but each of the company always have distinct strategy. This strategy defines the company uniqueness in the market. Strategy is essential in the realization of business model in order to grasp competitive advantage in the market.

3.2.2 Business Model Definition and Concept Amit and Zott (2001) pointed out that an organization’s business model is the central source to capture value for the organizations and its business networks, particularly in electronic business (e- business). This point of view is in contrast with Porter’s (1985) view of value chain framework. Porter (1985) argues that value chain analysis identifies activities of a firm and the economic implications of those activities, which in turn will create value for the firm. However, Porter’s view has been argued to be only appropriate for production and manufacturing business than service-oriented business or e-business (Amit & Zott, 2001; Stabell & Fjeldstad, 1998), such as cloud computing services.

There is several definition of business model that have been formulated. Chesbrough and Rosenbloom (2002) define business model as “the way an organization or a network of organization intends to create and capture value from technological innovation”. Osterwalder and Pigneur (2003) define business model in relation to generate profitable and sustainable revenue streams. While Bouwman et al. (2008) define business model as “a blueprint for a service to be delivered, describing the service definition and the intended value for the target group, the source of revenue, and providing an architecture for the service delivery, including a description of the resources required and the organizational and financial arrangements between the involved business actors, including a

Page | 35

Chapter 3 description of their roles and the division of costs and revenues over the business actors”. On the other hand Shafer et al. (2005) relates business model definition with business strategy, where he states that business model is a “representation of a firm’s underlying core logic and strategic choices for creating and capturing value within a value network”. Moreover, Amit and Zott (2001) emphasize that a business model should depicts the design of transaction content, structure and governance in order to capture business value.

Those business model definitions presented above show that there is yet to be a single agreement on the definition of business model. Within our research we choose to adopt the definition of business model by Bouwman et al. (2008), because this definition focus on the concept of service by technological innovation and the network of organizations around the service provision itself.

Following the definition of business model discussed above, several business model elements have arisen. Amit and Zott (2001) conceptualize value creation among novelty, lock-in, efficiency and complementarities characteristics in e-business. Shafer et al. (2005) defines business model into several group of components of strategic choices, value network, capture value and create value. While Osterwalder and Pigneur (2003) defines business model in the context of product innovation, customer relationship, infrastructure management and financials. Bouwman et al. (2008) conceptualize its model around the Service, Technology, Organizations, and Financials domains.

In the next section, we will present discussion on the STOF model in more detail.

3.2.3 The STOF Model The STOF model (Bouwman et al., 2008) puts emphasize on the concept of “service” to become the central point of its definition. STOF model provides a descriptive conceptual framework for the design of business model. We find the STOF model to be suitable to be used as a framework to answer our research objective. Because our focus is to understand how client can deliver their cloud applications or cloud-based service toward their end-users using PaaS. The focus here is also to deliver “service”.

STOF model is described in terms of four elements: service, technology, organization, and financial domains (as shown in the Figure 14). Each domain contains important design variables in designing a business model.

Page | 36

Theoretical Background

Figure 14 The STOF model (Bouwman et al., 2008)

In order to design a viable business model, Bouwman et al. (2008) argue that it is required to balance the requirements and interest of the actors involved, within and between the four domains of service, technology, organization and finance. “Requirement in the service domain guide the design choices in the technology domain, which in turn affect network formation and the financial arrangements” (Bouwman et al., 2008). They further argue that to understand this mechanism, the interdependencies among the involved Critical Design Issues (CDIs) need to be figured out. A CDI can be defined as “a design variable that is perceived to be (by practitioner and/or researcher) of eminent importance to the viability and sustainability of the business model under study” (Bouwman et al., 2008). These CDIs will lead to the Critical Success Factors (CSFs) of business model viability. CSFs here can be defined as “the limited number of areas in which satisfactory results will ensure that the business model creates value for the customer and for the business network” (Bouwman et al., 2008). The CSFs are categorized into two: for customer and network value.

In the following paragraphs, we will present the elaboration about the CSF formulated in the STOF model. We will also discuss how CSFs and CDIs are linked and related. In section 3.2.3, the application of this model to client business model in utilizing PaaS will be presented.

CSFs and CDIs in the STOF model The CSFs that are related to the creation of customer value are (1) Compelling Value Proposition (service domain), (2) Clearly Defined Target Group (service domain), (3) Unobtrusive Customer Retention (service domain), and (4) Acceptable Quality of Service (technology domain). The construction of the CSFs and related CDIs are depicted in Figure 15.

Page | 37

Chapter 3

Figure 15 STOF model Critical Design Issues and Critical Success Factors relating to creating customer value (Bouwman et al., 2008)

The success factor of Compelling Value Proposition refers to the benefits that are delivered to the user of a service by its provider. This success factor is build up from several CDIs, namely the Creating Value Elements (service domain), Branding (service domain), and Pricing (finance domain).

The Clearly Defined Target Group concerns with the focused end-customers of the related service. This CSF is co-determined by the Accessibility for Customers (technology domain). The Accessibility for Customers to the target groups is further influenced by the choice of technology platforms, devices, and architecture.

The Unobtrusive Customer Retention concerns with how customer retention strategy will not reduce customer value. The CDI of Customer Retention (service domain) influences this CSF. This design issue should lead to unobtrusive mechanism, which can be supported by User Profile Management (technology domain). User Profile Management deals with personalization of a service towards user interest, preferences and behavior.

Finally, the Acceptable Quality of Service concerns with technology aspects that are related to the delivery of the service. The CDIs that are related to the technology domain such as the Quality of Service, Security, and System Integration should lead to an acceptable (quality) level. Quality of Service relates to the performance of the technological architecture in delivering the functionality. The Security aspect concerns with service accessibility and the security of communication and information processing. The System Integration deals with compatibility matters with the existing technical infrastructure and that between subsystems.

Furthermore, the CSFs that are related to the creation of network value are (1) Acceptable Profitability (finance domain), (2) Acceptable Risk (finance domain), (3) Sustainable Network Strategy

Page | 38

Theoretical Background

(organization domain), and (4) Acceptable Division of Roles (organization domain). These CSFs focus on finance and organization domains, particularly related to organization arrangement that provide essential resources and capabilities of a service. Figure 16 depicts the relation of CSFs and CDIs in creating network value.

Figure 16 Critical Design Issues and Critical Success Factors relating to creating network value (Bouwman et al., 2008)

The CSF of Acceptable Profitability concerns with an acceptable financial result (a positive result) that match the companies’ risk/return profile. This CSF is influence by the CDI Division of Cost and Revenues (finance domain), Pricing (finance domain) and Acceptable Customer Base (service domain). CDIs such as Customer Retention (service domain), Accessibility for Customers (technology domain), and Network Openness (organization domain influence the CDI Acceptable Customer Base.

The Acceptable Risks is related to the high uncertainty with respect to technology choices and market acceptance. This CSF is influenced by CDI Division of Investment (finance domain) and Division of Costs and Revenues (finance domain). These CDIs are enabled by the CDI Valuation of Contribution and Benefits (finance domain)

From the organizational domain, CSF Sustainable Network Strategy concerns with securing access to (inimitable) resources and capabilities. This CSF is influenced by CDI Network Governance, which concerns with rules of organization collaboration and partnership. CSF Acceptable Division of Roles also belongs to organization domain. This CSF refers to the distribution of roles among firms and the integration of roles within firms that participate in the business network of a service provision (Bouwman et al., 2008). This aspect is also related to resource and capabilities of the organization that develop the service. This is then related to the decision to in source or outsource. Which means the CDI Partner Selection (organization domain) and Network Complexity (organization domain) are

Page | 39

Chapter 3 here relevant. Partner Selection concerns with design issue in selecting and acquiring access to the needed resource and capabilities to realize the service offering. Network Complexity concerns with relationship and management of the value network, especially when many actors are involved. This is also related to technical architecture of the actors’ IT applications and systems.

The two groups of CSFs and their related CDIs mentioned in previous paragraphs form a causal model that explains the viability of business model. This model elaborate creation of customer and network value from a viable business model. Managers and business model designers can influence the CDI in order to make a feasible and viable business model.

In the next section 3.2.3, we will apply this model to PaaS client business model development. We will use the elaborated aspects and issues to analyze the adoption factors of PaaS client, in order to see how those aspects are seen from the PaaS client perspective.

3.2.4 Applying STOF Model to Client PaaS Business Model In this section, we want to understand how the CDIs and CSFs described in STOF model are seen from the client perspective in PaaS domain, in order to construct our adoption factor model, and give us insights of the important issues and variables in designing business model by utilizing PaaS. We also want to see how the service, technology, organization and financial domain are interrelated in client PaaS business model. Our definition of adoption factor is similar with the definition of CDI, therefore we will analyze the relevant design issues in client PaaS business model with respect to the STOF CDIs.

In applying the STOF model to PaaS domain, we take the perspective of cloud applications (SaaS) that are built on PaaS as the “service”. This is the service that a client wants to offer to its end-users. Then, we want to analyze and understand how PaaS impacts each of the four STOF components for that cloud applications (SaaS). As mentioned at the beginning of this chapter regarding the client perspective, we can take two standpoints in discussing the subject matter. First isthe standpoint of the client as PaaS customer, which concerns with the value that the client perceives from PaaS offerings from PaaS providers. Second is the standpoint of the client as SaaS service provider that utilize PaaS in order to deliver the service. This standpoint concerns with the value that client want to propose and deliver towards its end-users.

Service Domain The related CDI in the Service domain are Targeting, Creating Value Elements, Branding, and Customer Retention. To apply the first CDI of Targeting in our case, this issue is relevant to be seen from the second standpoint of the client as SaaS service provider. Targeting deals with the target group of the cloud application. Should the service offering be targeted at the client’s own employees or their end-customers or their suppliers of the business? Would the target group differ among different client organization across different industries? This issue of targeting is related with the goals of the cloud applications. The client’s purpose or goal to offer a cloud application would define the end-users as well. Briefly, this selection of design issue will relate to the organization, technical, and financial arrangement, for instance whether an aspect of revenue source is relevant in the business model design, or whether extra security of the service quality is needed. Figure 17 depicts the applicability of the CDI Targeting.

Page | 40

Theoretical Background

Targeting PaaS Provider Client End-users

Figure 17 Applicability of CDI Targeting

Secondly, the CDIs that relates to CSF Compelling Value Propositions is relevant to be discussed from both of the client standpoints. From the client as PaaS customer standpoint, the related CDI elements of Branding can be seen as the appealing value that influence the client to choose which provider that suits their business model, because branding is related with trustworthiness of the PaaS provider. While the CDI Creating Value Elements is more relevant to be discussed form the other standpoint. Creating Value Elements concerns with the particular added value that client could offer to its end-users by utilizing PaaS. This CDI is closely connected with the choice of the target group. Different target group might offer different added value. For instance, in case of the target group is client suppliers, building application in PaaS might give a sense of standard platform for system integration as there might be another application built by the supplier that needs to interact with the first application. While in case of the target group is client’s own employee, where the application is more to support business process and employees will not develop their own application, that kind of added value might be not relevant for the end-users. Additionally, the added value that is created by utilizing PaaS can also be applicable towards the internal organization of the PaaS client itself. What we mean hear is that the benefit of using PaaS technology might be not reached the end-users, but will benefitting the client only. For instance, the benefit of cost reduction or faster time to market by utilizing PaaS will not be felt by the end-users, but it will benefitting the client who offer the applications/service. Figure 18 depicts the applicability of CDI Branding and Creating Value Elements in our analysis.

Branding of PaaS Creating Value Elements: for provider influence end-users and client internal PaaS provider selection organization PaaS Provider Client End-users

Figure 18 Applicability of CDI Branding and Creating Value Elements

Next is the CDI Customer Retention. This issue is more relevant to be discussed from the standpoint of client as PaaS customer. This is related to how retention customer mechanism of PaaS provider influences the client in selecting the PaaS provider. This issue could refer to the aspect of the degree of lock-in that PaaS client find acceptable regarding the service that the client offers to its end-users. Thus this issue is related to the issue of Partner Selection, which will be presented in the following paragraph. The issue of customer retention is not relevant from the standpoint of the client as service provider, because this issue is not influenced by the utilization of PaaS. Figure 19 depicts the applicability of CDI Customer Retention in our analysis.

Retention mechanism from PaaS provider influence PaaS provider selection PaaS Provider Client End-users

Figure 19 Applicability of CDI Customer Retention

Page | 41

Chapter 3

Technology Domain The related CDI in the Technology domain are Security, Quality of Service, System Integration, Accessibility for Customers, and Management of User Profiles. The CDI that related to CSF Acceptable Quality of Service is relevant to be discussed from both of the client standpoints. First, from the SaaS service provider standpoint, the client would translate the service into technology requirements in terms of Security and Quality of Service. These technology requirements are influenced by the service/cloud application characteristics, such as the criticality of the application towards client’s business, the sensitivity of data handled in the application, and the utilization pattern of the application. At the end, these security and quality of service requirements would be reflected to the selection of PaaS provider. Thus the client will take the stand point of PaaS customer to assess whether the security level and quality of service of PaaS offered from the selected PaaS provider is acceptable to deliver the wanted application/service. This factor is then closely related to the CDI Partner Selection. Figure 20 depicts the applicability of CDI Security and Quality of Service in our analysis.

Security and QoS requirement influence PaaS The SaaS defines requirement provider selection on Security and QoS PaaS Provider Client End-users

Figure 20 Applicability of CDI Security and Quality of Service

Next is the CDI System Integration. This issue concerns with the extent that the cloud service can be integrated into the existing technical infrastructure. This issue is relevant from both standpoints. From the standpoint of the client as SaaS service provider, this issue concern with the software standard interoperability of the “new” cloud application with the “old” IT system that the end users use, in order to have seamless operation. From the standpoint of the client as PaaS customer, this issue can be translated into the requirement to choose which PaaS provider to be used. For instance, if the end-users’ old IT system is built around the Microsoft standard, then it would influence the client to have more tendency to choose Mircrosoft Azure as PaaS provider than the other providers that have different software standard. Figure 21 depicts the applicability of CDI System Integration in our analysis.

System integration influence the selection of software standard to be used, and also the selection of System integration with end- PaaS provider users’ old IT system PaaS Provider Client End-users

Figure 21 Applicability of CDI System Integration

Moreover, the issue of Accessibility for Customers is of less relevant here, because the cloud applications can only be accessed through the Internet. The Internet becomes the standard technical requirement that cannot be influenced by client. This also applies to the issue of Management of User Profiles. This issue is not relevant in our discussion. This issue concerns with the personalization of a service. Cloud application that is built on PaaS uses multi-tenancy concept as explained in Chapter 2. Thus, a client as SaaS service provider would not be able to build custom application for each of its end-user.

Page | 42

Theoretical Background

Organization Domain The related CDI in the Organization domain are Partner Selection, Network Openness, Network Governance, and Network Complexity. The first CDI to be discussed is Partner Selection. This issue concerns with acquiring access to resource and capabilities needed to realize the wanted cloud application service. It is apparent that the most important partner in client PaaS business model is the PaaS provider. With regards to this, we will take the standpoint of the client as the PaaS customer. The main issue in partner selection from this standpoint is how client decides which PaaS provider to select and what influence client in PaaS provider selection. From the discussion of the previous domains, we derived the issues that influence the selection of PaaS provider explicitly. We extracted the issues of Branding, and Retention Mechanism from the Service domain, and we got the issues of Security, Quality of Service, and System Integration from the Technology Domain. Thus, we can see here that the issue of PaaS provider selection is of significant in client’s business model.

From the second standpoint of the client as SaaS service provider, we can see Partner Selection not exclusively about PaaS provider selection. Partner Selection is related to the kind of cloud application service to be offered by client. Thus there might be needs to involve more actors in order to deliver the wanted service. For instance, if the security level required by a service application is really high in which the PaaS provider resource and capability cannot guarantee and deliver the requirement, other actor such as ISP might come in to contribute by providing complementary security. The logic here is because any cloud applications are access by Internet, and the ISP is the main player in providing the Internet network. This is only one example. Other kinds unforeseen involvement from complementary actors might emerge depending on the cloud service that the client (would) offers. Figure 22 depicts the applicability of CDI Partner Selection in our analysis.

Involvement of complimentary PaaS provider selection that partners w.r.t. the types of depends on several issues application PaaS Provider Client End-users

Figure 22 Applicability of CDI Partner Selection

The discussion of Partner Selection, particularly about PaaS provider selection, is also related to the resource and capabilities aspect of the client.

Next is about the CDI Network Openness. We find that this issue has no relevance in our client business model discussion. The power to influence this issue is at the hand of the client itself. In order to deliver the cloud service, all of the needed actors must join the value network from the first place, otherwise the service cannot be realized. On the other hand, the openness of the PaaS platform itself is not the concern of the client.

The issue of Network Governance is of importance in almost every business model. Here, both of the client standpoints are applicable. From the standpoint of the client as PaaS customer, we can see that PaaS provider is the dominant actor, because PaaS provider is the one who deliver the technology. Though, the client still has limited control on this matter, because the client has the power to select any PaaS provider that suits the client’s goal. From the other standpoint, the network governance is strictly depends on the involvement of other actors. The simplest application would only involve PaaS provider and end-customers. In a more complex value networks, the client has to

Page | 43

Chapter 3 find the best way to orchestrate the value networks so that the realization of the service can be sustainable. Figure 23 depicts the applicability of CDI Network Governance in our analysis.

Governance to maintain Governance towards the sustainable relationship with PaaS provider involved actors PaaS Provider Client End-users

Figure 23 Applicability of CDI Network Governance

Finally, the discussion about Network Complexity issue is closely related with the CDI Network Governance and Partner Selection. The complexity really depends on the number of the involved actors. As discussed before, the number of involved actors is influenced by service and technical requirements. Therefore, we will not discuss this issue in more detail because it will overlap with our previous discussion.

Financial Domain The related CDI in the Financial domain are Pricing, Division of Investments, Division of Costs and Revenues, and Valuation of Contribution and Benefits. The CDI of Pricing is relevant to discuss from the standpoint of the client as SaaS service provider. This issue concerns with the price that the client want to charge to its end-users. This issue depends on the type of the application as well as the target end-users of the cloud applications. For instance, if the end-users are the client’s employees, most likely there will be no financial flow from the end-users to the client as service provider. Or in other word the service will be consumed for free. But if the end-users are the client’s customer or suppliers, some pricing schemes that monetize the use of the service might be applicable. Figure 24 depicts the applicability of CDI Pricing in our analysis.

Pricing PaaS Provider Client End-users

Figure 24 Applicability of CDI Pricing

The CDI of Division of Costs and Revenues is of little relevance in our discussion of client PaaS business model. What really matter from the client’s perspective is the cost incurred by utilizing PaaS in order to realize an cloud application service. The client is the central actor of the cloud application business model. So the term “division” does not make any sense. In PaaS utilization context, the revenue would be directly related to the cost. PaaS could bring the benefit of cost reduction for the client, thus the more cost is reduced, the more revenue client would get. Figure 25 depicts the applicability of CDI Cost in our analysis.

Cost PaaS Provider Client End-users

Figure 25 Applicability of CDI Cost

The same case would apply for CDI Division of Investments and Risks. The term “division” is not relevant in our discussion, because the client is the central actor in our discussion of business model. The fact that PaaS providers are generally give trial period to use their PaaS at zero cost, reduce the financial risk of client to start using PaaS technology. The fact that there is no up-front investment

Page | 44

Theoretical Background needed to start using PaaS solutions would also eliminate the issue of Investment in designing client PaaS business model. However other risks of using PaaS might still emerge from the other domains other than financial.

Revenue PaaS Provider Client End-users

Figure 26 Applicability of CDI Revenue

Finally, the CDI Valuation of Contributions and Benefits in client PaaS business model would only relevant in terms of cost. As argued before that there is no division in the financial domain of the client PaaS business model. Therefore, from the client’s perspective, valuation of contribution from the involved actors would translate into the cost of the value activities themselves.

3.2.5 Conclusion From the analysis of STOF model application in the PaaS domain presented above, we have identified several issues that might influence the design of client’s business model in order to utilize PaaS. In the bigger picture, we can identify two kinds of issues. The first issue deals with aspect of the business model design around the provision of the cloud service by client towards its end users. The second issue deals specifically with PaaS provider selection in order to deliver that cloud service, which is part of the Partner Selection design issue. The identified issues are summarized in the following figure.

Technology Domain Organization Domain

Service Domain influence PaaS Providers’ Branding (S) Security PaaS Provider influence the Selection Targetting requirements for influence PaaS Providers’ Partner Selection Retention Mechanism Quality of Service Complementary Creating Value Actors PaaS Provider’s Elements Security Profile (T) System Integration PaaS Provider’s Quality of Service Profile (T)

Financial Domain PaaS Providers’ System Integration Profile (T)

IT Budget Pricing Cost Budget Value-added Revenue Model resources / capabilities

Figure 27 The identified client PaaS business model design issues derived from STOF model

Based on the two standpoints we used in our analysis, we can see two categories of design issues in designing PaaS business model for clients. The first category is related to the design of business model in order realize a cloud service towards client end-users. Thus the standpoint here is client as the service provider. The design of the business model starts with the issue of the service application. The issues are related to the targeted end-users and the creation of value elements towards the end- users as well as towards the internal organization of the client itself. These issues are the design choice of the client. The resulting design choices will have implications on the technology requirements to deliver the wanted application service. These technology requirements deal with the

Page | 45

Chapter 3 issue of security, quality of service, and system integration. Then these technology requirements will influence the organizational domain, particularly in the selection of partner in the value networks. The selection of partner includes the selection of PaaS provider and any additional actors to bring unique value activities. These unique value activities are depended on the type of the application and technology requirements. The partner selection is further influenced by the cost budget of a client company. This cost budget is derived from the IT budget for IT service provision in an organization and also the possible revenue model that might apply from the cloud service provision towards the end-users.

Furthermore, we would like to zoom-in on the issue of PaaS provider selection. The CDIs of the STOF model also gave us insight on the issues that might be influencing the selection of PaaS provider. Those issues are related to the PaaS product that the providers offer. They are: branding of the PaaS product, that could influence the client’s trust; the customer retention mechanism that relates to the degree of customer lock-in; the security profile, the quality of service profile, and the system integration profile. The last three issues acts as a checklist towards the technology requirements of the business model, whether a service provider fulfill the requirement or not. In case a service provider’s profiles on security, QoS, and system integration do not match with what is required, then the client has the control to select other PaaS providers that fulfill the requirements. In case that there is a required value activity for a particular service, that none of the PaaS service providers are capable to fulfill, client might want to involve other actors to bring this specific value activity in the value network. For instance, this can be in the form of extra security in the Internet network to access the cloud service. Network provider like ISP can play a role to provide and guarantee the needed extra security.

3.3 Integrated PaaS Adoption Factors

As mentioned at the beginning of this chapter, the aim of this chapter is to identify and conceptualize the PaaS adoption factors. PaaS adoption factors are the factors that influence client’s decision in designing their business model to adopt and utilize PaaS technology. From the theories discussed in Subchapter 3.1 and 3.2, we have pointed out some relevant issues with regards to PaaS adoption. In this subchapter we present the final selection of adoption factors that are derived from the theory discussed.

The derived design issues from STOF model serve as the main conceptual model to structure the adoption factors. The identified characteristics of cloud application from subchapter 3.1 are translated into design issues and used to refine the design issues in the service domain to be more PaaS specific (shown in grey boxes in Figure 28). The work from Reynolds and Bess (2009) put emphasize on the “goal” of utilizing the cloud computing. Therefore we put this emphasize as a design issue to start designing a cloud application service by PaaS. Furthermore, Reynolds and Bess (2009) also contribute to a design issue of who will benefit from the use of cloud solution. We integrated this issue with the issue of value element creation. We find that the value element created by the utilization of PaaS might not only beneficial for the end-users, but also for the client as service provider itself. The IT department or Business Unit can represent the client of PaaS service. Moreover, we translate the characteristic of the application to be migrated to the cloud defined by Misra and Modal (2011) into a design issues in the service domain. The level of criticality, data sensitivity, and utilization pattern of a cloud application are relevant issues that client have to take

Page | 46

Theoretical Background into account when designing its business model. These issues have direct influence on technological requirement. For instance, the level of data sensitivity might influence the required security aspect to realize the application service. These issues are also determined by the purpose of the cloud application creation. The integrated PaaS adoption factors are depicted in Figure 28 below.

influence Service Domain Technology Domain PaaS Providers’ branding (trust) Target of the Organization Domain application service PaaS Providers’ Security PaaS Provider customer defines influence the Selection lock-in mechanism Value Elements for requirements for influence end-users and Partner Selection PaaS Provider’s internal Quality of Service security profile organization Complementary Actors PaaS Provider’s The goal to Quality of Service Level of criticality of build cloud System profile work done by the application Integration applicattion service service PaaS Providers’ system integration Level of data profile (software Financial Domain sensitivity handled standard) by the application service Value-added Pricing IT Budget resources / capabilities Level of utilization pattern of the Cost Budget application service Revenue Model

Figure 28 The integrated PaaS adoption factors model

The integrated PaaS adoption factors depicted in the figure above begins with the conceptualization of the cloud application service. This domain is the core of the business model design. It starts with the goal or the main purpose of the intention to create cloud application service. The aim of the cloud application will influence client decision on targeted end-users, the offered value element for end-users and the client itself, the level of criticality of work done by the application service, the level of data sensitivity handled, and the level of utilization pattern of the application. Then, these factors will influence client decision on the needed technology requirements (with respect to PaaS) on security, quality of service, and system integration issues. The needed technology requirements then will influence client decision on the involved actors in the partner selection factor. The partner selection factor can be analyzed from two aspects, first is the PaaS provider selection. In selecting the PaaS provider, the client has to match the technology requirements with the PaaS provider’s profile on security, quality of service, system integration (software standard), customer lock-in mechanism, and branding. This selection concerns with the ability of PaaS provider to support the realization and delivery of the wanted cloud application service. If a PaaS provider cannot fulfill the technological requirements, the client might select another PaaS provider, or involve complementary actors to fill- in the needed capabilities that the PaaS provider cannot offer. The partner selection would also be influenced by the cost budget that is derived from the client company’s IT budget and possible income from the revenue model towards the client’s end users. The involvement of the revenue mechanism in the realization of the service depends on the aim and the type of service designed in the first place.

3.4 Conclusion

This chapter offers theoretical information on the related works on cloud computing and business models, particularly about the STOF model. These theories are used to construct a conceptual model of PaaS adoption factors, which are the factors that influence clients’ decision in designing their

Page | 47

Chapter 3 business model to adopt and utilize PaaS technology. The construction of this conceptual model helps us in answering the sub-question 2, namely:

“What sort of PaaS-related generic business model from client’s perspective is appropriate to deliver the wanted cloud services in terms of service, organizational and financial domain?”

This conceptual model furthermore will be used to help us answering the sub-question 3, namely:

“What sort of business software applications do client organizations want to build or migrate to the cloud by using PaaS and which PaaS providers are appropriate to that demands?”

The conceptual model of PaaS adoption factors will help us as a starting point in understanding the reasons and the type of software applications that clients want to build/migrate to the cloud by using PaaS, by means of indicating the relevant influencing factors. Also the conceptual model will help us in identifying the reasons of PaaS providers’ selection. The conceptual model will also help us in understanding the client PaaS business model in a more structured way, particularly in terms of service, technology, organization and financial domains.

Furthermore, the constructed conceptual model is going to be validated through interviews. To have reliable validity, interviews are going to be conducted across large enterprise companies in several industries, namely the financial service industry, the communication and high tech industry, the resources industry, the products industry. Elaboration of these interviews will be presented in the following chapters. Through this validation, the author expects to be able to draw important lessons to give general guidance of PaaS utilization and adoption by large companies as PaaS client.

Page | 48

4 Towards a Validated Adoption Model

This chapter aims to answer mainly the sub-questions 3 and 4 by means of validating the conceptual model proposed in Chapter 3.3. Sub-question 3 concerns with the sort of business software applications that client companies want to build or migrate to the cloud by utilizing PaaS and the model validation, while sub-question 4 concerns with the similarities and differences that are arisen among the results. This chapter also aims to answer a partial of sub-questions 1, which is to answer the type of organizations that are potential to be PaaS customers. Interviews with multiple clients and vendor companies were conducted to answer those sub-questions. Based on the interview results, the conceptual model is validated and the market potential of PaaS service in the Netherlands is also identified. Furthermore, barriers and risks that emerge from the PaaS adoption are also identified in order to have more understanding about the PaaS adoption issues.

The interviews were conducted with 9 client companies that have been using or considering to implement cloud services, and 3 vendor companies that have PaaS offering. The selection of the companies was done together with experts from Accenture that knew the condition of the firms regarding the possible interest on PaaS. Prior to each of the interviews, an introduction letter that explains the focus of the research were sent to the respondents. The 9 client companies responded positively to be interviewed. We expected that these 9 client companies were all interested with the public PaaS service offering. However, during the interview, we found out that only 2 client companies consider to use or have used public PaaS service, 3 client companies consider to use private PaaS concept, and 4 client companies are more interested on SaaS service offering other than PaaS service offering. This fact becomes a challenge in our analysis process. Therefore we decided to modify our analysis approach by validating the model from three different focuses, there are: - Approach 1: focus on PaaS adoption from client’s perspective - Approach 2: focus on SaaS adoption from client’s perspective - Approach 3: focus on PaaS adoption from vendor’s perspective

Those three approaches are taken based on the following reasons. Each of the client interviews began with a question about the company’s general strategy for any cloud computing services. Most of the respondents answered by describing their implementation of SaaS, or their plan to use SaaS in the near future. Then the question continued by asking whether the company respondent has used or consider using PaaS. When the company responded positively to consider using PaaS, then we proceeded the interview protocol with focus on PaaS. Otherwise, we improvised by proceeding the interview protocol with focus on SaaS. Throughout both of the interview types, the interview protocol used in framing the questions represent conceptual model proposed in Chapter 3.3. Thus, the two types of interview can be used to validate the same model but from different focus. During our analysis, we intend to compare the models from each focus to gain more understanding on the difference of PaaS and SaaS adoptions. Detailed elaboration on our analysis approach will be presented in section 4.1.4.

Page | 49

Chapter 4

Within this chapter, we will begin with elaborating the interview methodology. The choice of interview as a methodology, and the respondent selection will be explained. We will also elaborate the way we construct the interview protocol and the way we analyze the interview results subsequently. Furthermore the results and analysis of the interview are going to be presented in the following section. This section will contain detailed elaboration on the validation of the conceptual model proposed in Chapter 3.3.

4.1 Interview Methodology

This section will elaborate on the interview methodology that is used in collecting the data. First, the decision to go with interview as data collection method is explained. Then the selection procedure of the interview respondents is presented. Subsequently, the interview protocol and the data analysis approach are going to be elaborated.

4.1.1 Why Interview? Interview is chosen as data collection method in order to have in depth understanding of the problem at stake. This method is aligned with the nature of this research, which is an exploratory research. Conducting interview allow us to not only validating the conceptual model, but also for getting a chance to understand the reasoning behind respondents answers. To understand the reasoning behind each answer is crucial in our research because cloud domain, particularly Platform- as-a-Service domain, is a new concept in the IT world, where not much of research has been done in this area.

We took semi-structured interview approach with open-ended questions. This approach allow us to get elaborative answers from the respondents in a conversational way, in which on one hand deep understanding from the answers and interesting details can be gained, while on the other hand the structured questions can also be asked to validate the conceptual model.

4.1.2 Interview Respondents Selection We conducted interviews with two types of companies: client companies and vendor companies. For client companies, intentionally we aim to interview different companies across different industries in order to see if there is specific characteristic that can be identified based on the type of the industry. In general we select the companies based on the criteria that the companies are large company with 1000+ employees, and the companies are located in the Netherlands for practical reason. Furthermore, we select the interview respondents from the company based on their functional role. We aim to interview IT managers or IT architect that has the knowledge on their company’s IT landscape and has an influence on the company’s strategic IT decision. Table 9 below shows the overview of the respondents. Due to confidentiality requirement that the respondents address to participate on the research, the company respondents could not be named in this report. A detailed list of company respondents is only available on request. However, the types of the companies are presented on the table and the companies’ general description is elaborate in section 4.2.1. We interviewed respondents from 9 client companies. It is important to be noted that we had two interviews with two different people in Company C9. The way we process the data from these two interviews is explained in section 4.1.4.

Page | 50

Towards a Validated Adoption Model

While for vendor companies, we selected the top three PaaS vendors based on the report from Burton Group (Watson, 2011). Popular cloud players such as Amazon was not among the respondents, because Amazon cloud business is fall into the IaaS categorization. The respondents from the vendor companies were selected on the basis of their PaaS offering knowledge, specifically the knowledge about the proposed value and the differentiator of their PaaS offering. Therefore we conducted interviews with respondents that has marketing or sales role. Unfortunately, an interview with one of the top vendors was not feasible; therefore we decided to interview an expert on the respective vendor from Accenture.

Table 9 below shows the location and the date of the conducted interviews. All the interviews were conducted in May 2011.

Company Type of Company Industry Sector Functional Role Date Code Company C1 Dutch utility company Resources IT Architect & 11-05-2011 Project Manager Company C2 Telecommunication company Communication & IT & Business 13-05-2011 in the Netherlands High-Tech Architect Manager Company C3 Telecommunication company in Communication & IT I&O Manager 17-05-2011 the Netherlands High-Tech Company C4 Dutch retail company Product IT Operations 18-05-2011 Manager Company C5 Dutch investment banking Financial Service IT Development 20-05-2011 company Manager Company C6 Dutch insurance company Financial Service IT Architect 20-05-2011

Company C7 Dutch transportation company Product Enterprise 25-05-2011 Architect Company C8 Dutch incumbent bank Financial Service Innovation 25-05-2011 Manager Company C9a Dutch private investment bank Financial Service Demand Manager 26-05-2011

Company C9b Dutch private investment bank Financial Service IT Architect 26-05-2011

Company V1 Incumbent PaaS Vendor IT VP Sales 9-05-2011

Company V2 Incumbent PaaS Vendor IT Business Group 23-05-2011 Lead Company V3 Incumbent PaaS Vendor IT Expert 10-05-2011

Table 9 Interview respondents

4.1.3 Interview Process and Protocol We design the interview to serve the purpose of validating the conceptual model proposed in Chapter 3 and to identify the PaaS market potential in the Netherlands. There are two types of interview protocol were made. One interview protocol is intended for interviews with client companies, and the other interview protocol is intended for interviews with vendor companies. The two different protocols were made because each of the interviews with clients and vendors carry a slight different perspective. The interviews with clients focus on the perceived values from PaaS offering, and on the other hand the interviews with vendors focus on the proposed values from PaaS offering. Nevertheless, the interview questions were derived from the same conceptual model.

Page | 51

Chapter 4

Each of the interviews was conducted in one hour. Prior to the interview, an introduction letter was sent to each of the respondents. This letter contains the objective of the research as a whole, the expectation, and the context of the interview. At the beginning of the interview, the background and the objectives of the research were re-explained. The interview also began by asking permission to make a recording of the interview for administrative reasons, and also assuring that the interview result would be processed anonymously. The key concept such as the definition of PaaS was also explained in every interview to adhere general understanding of the concept among the respondents.

Before the actual interviews, our research project supervisors reviewed the interview protocols. The interview protocols consist of open-ended questions that guide the interviews. The full interview protocols can be found in the appendices, and examples of the questions topics are shown in the Table 10 below.

Topic Questions for Clients Questions for Vendors A. General questions 1. Have your organization used any kind of 1. Which client would you target regarding the company’s cloud based services? Why? with your PaaS offering? cloud strategy 2. Does your company consider using PaaS? 2. What is the differentiator of you Why? PaaS offering? B. Main Questions regarding 3. What is your underlying reason to adopt 3. How do you think customers the model PaaS?(S) could change their business model 4. What kind of applications do you consider by using PaaS? (S) to build on PaaS? (S) 4. How would you maintain your 5. What are the unique characteristics of the PaaS offering QoS? (T) application service that can only be 5. Which other party do you think is realized by the utilization of PaaS? (S) important in PaaS ecosystem? (O) 6. What is the most important technical 6. How would you differentiate you implication in your application service? offering in terms of pricing? (F) (T) 7. How do you choose the PaaS providers? Which criteria do you apply? (O) 8. How does your company’s IT budget influence your decision to use PaaS? (F) C. Questions regarding 9. What barriers and risks do you perceive in 7. What barriers and risks do you barrier and risks adoption and utilizing PaaS? see in PaaS adoption by enterprise? D. Question regarding the 10. What role would you think Telecom 8. What role would you think role of Telecom company company could play in PaaS service Telecom company could play in provision in general? helping PaaS adoption?

Table 10 Example of interview questions

The general questions were meant as warm up of the interview and to get insight on the current status of the company’s cloud strategy. The main questions regarding the service, technology, organization, and financial domains were the core interview questions that were derived from the conceptual model and meant as a validation of the conceptual model itself. These questions consist of some main questions and several sub-questions that refer to each of the factors in the conceptual model. The questions about barrier and risk were meant to identify the perceived barrier and risk in PaaS or Cloud adoption. While the last question about the role of telecom companies in the generic PaaS service provision was meant to help the researcher to get an insight to answer sub-question 5. In addition some lists of possible PaaS benefits, PaaS vendors overview, and degree of vendor lock-in were shown during the interview to help the respondents framing their answers. These lists are presented in the Appendix C.

Page | 52

Towards a Validated Adoption Model

Moreover, as explained at the beginning of the Chapter 4, not all of the client companies were desiring to use PaaS, therefore the main questions, mainly the questions in topic B and C, could not be addressed with focus on PaaS. Even though we asked in depth about the reasons why they did not desire to use PaaS, we have to improvise the questions in those cases in order to have useful data. The questions that were specific about PaaS were changed to be about SaaS and cloud service in general. Therefore we could still validate the model but with slightly different focus.

4.1.4 Data Management and Analysis Approach As mentioned before, with the consent of the interview respondents, audio recording were made for all of the interviews. Then, non-literal transcript was made for each of the recorded interviews. Each of the interview transcripts was sent back to the respondents for review and correction. The respondents were asked to check whether the transcripts are the truthful account of the interviews. The interviews result in 74 pages of transcripts. In order to process the data systematically, we use a software tool called Atlas.ti. The use of software tool in analyzing qualitative data can reduce analysis time, make procedures more systematic and explicit, and permit flexibility and revision in analysis procedure (Tesch, 1989). The Atlas.ti is a computer-aided tool for qualitative data analysis. Atlas.ti is able to help researchers in handling large amount of text, as well as managing annotations, concepts, and complex structures, including conceptual relationships that emerge in the process of interpretation of the texts (Bharosa, 2011).

The transcripts were then imported to the software to be processed. The software allows us to code some part of the texts into variables according to our analysis points. Coding is useful to assign a meaning to the descriptive or inferential information complied during the interview, so that a researcher can quickly find, pull out, and cluster the segments relating to a particular construct in the analysis process (Miles & Huberman, 1994). The conceptual model proposed in Chapter 3 was used as the preliminary coding structure for the data. More codes were made during the analysis as new specific issues began to surface. The coding allows us to group the variables and relate one interview with the other.

Moreover, Atlas.ti also helps in visualizing the analysis results in a “network view”. This feature allows us to visualize the relation between one variable to other variables. The tool is also able to give indication of the importance of a variable by showing how many times that respected variable has been quoted in the interviews.

Furthermore, as mentioned in section 4.1.2, we had two interviews with different people in different role in Company C9. One interview was conducted with the IT Demand Manager (Company C9a) which has more business-oriented view, and the other one was conducted with an IT Architect (Company C9b) which has more technical-oriented view. The time allocated for each of these interviews was less than the initial plan of one hour due to limited time the respondents have. Therefore, each of the interviews focuses on the different aspects of the interview protocol. The interview with Company C9a focused on answering topic A and partly topic B (question number 7 and 8) of the interview protocol. While the interview with Company C9b focused on the rest of topic B, topic C and topic D of the interview protocol. The results from these two interviews were than combined to complete the answers for one set of interview protocol.

Page | 53

Chapter 4

As explained at the beginning of this chapter, we had some variances in our interview results. Some respondents answers the question with focus on PaaS, since they were interested on utilizing the PaaS service, but some respondents answers the questions with focus on SaaS, since they were not interested on utilizing the PaaS service. However, during the interviews, some overlaps occurred in the answers (between the PaaS and SaaS focus). This is happening because some of the given answers took the perspective of cloud services in general, which automatically implies to both PaaS and SaaS. We accepted this perspective because there are a lot of similarities in terms of technology and service characteristics between PaaS and SaaS as explained in Chapter 2. Additionally, the reasons why some companies were not interested on PaaS can also be understood as factors that the client take into account whether to adopt PaaS, thus acting as an input for the model validation as well. Therefore, for each domain of the model, we classify the answers from each interview whether it focus on PaaS or SaaS, or whether their answers cover both of the focus. The classification is shown in the tables below. “Plus” sign means that the respective components of the model are covered, while “minus” sign means they are not.

Components of the conceptual model (Client perspective) Respondents PaaS SaaS S T O F S T O F Company C1 - - - - + + + + Company C2 + + - + + + + - Company C3 + + - + + - - - Company C4 + - + - + + + + Company C5 - - - - + + + + Company C6 + + + + + + + + Company C7 + + - + + + + - Company C8 + + + + + + + - Company C9 + + + + + + + +

Table 11 Classification of client interviews

Components of the conceptual model (Vendor perspective) Respondents Service Technical Finance Organization PaaS SaaS PaaS SaaS PaaS SaaS PaaS SaaS Company V1 + - + - + - + - Company V2 + - + - + - + - Company V3 + - + - + - + -

Table 12 Classification of vendor interviews

From Table 11, we can observe that with Company C2, Company C3, and Company C7 the conversation about PaaS covers the Service, Technology and Financial domain, but Organization domain. The reason is that those companies are adopting private PaaS approach, therefore discussion about the PaaS provider selection (Organizational domain) was not relevant to be discussed. The conversation with Company C4 was mainly about SaaS. However, their reason not to use PaaS can be analyzed as factors for PaaS adoption in Service and Organizational domains. Moreover, the Financial domain for SaaS adoption was not covered with Company C2, Company C3,

Page | 54

Towards a Validated Adoption Model

Company C7, and Company C8 because these companies are the ones who are interested on PaaS (either public or private approach). Thus, within the limited time, we spent the interview time to focus on PaaS more than the SaaS. While on the other hand, as can be seen from Table 12, the interviews with PaaS vendors cover the topic of PaaS only.

Nevertheless, the model will be validated from three focuses and finally will be compared to each other. The first model represents client PaaS adoption factor. Then, this first model will be compared to the second model that represents client SaaS adoption factor. The third model represents PaaS adoption factor from the vendor perspective. This third model will be compared with the first model to see whether the proposed value from PaaS offering that PaaS vendor try to propagate is matched with the value that the clients perceived.

4.2 Results and Analysis

The analysis will begin with elaboration on the market condition in the Netherlands based on the interviews with 9 client companies and 3 vendor companies. Subsequently, validation of the conceptual model will be discussed and compared from the three different analysis approaches as explained at the beginning of Chapter 4.

4.2.1 PaaS Market Condition In this section we will present general overview of each of the interview results. General description of the companies will be presented. Moreover, each of the company strategies and states regarding cloud computing adoption in general and PaaS adoption in specific will be elaborated.

The Figure 29 below shows the market condition regarding PaaS adoption in the Netherlands based on our interview results. The figure shows the number of the companies that in general consider using PaaS service, companies that consider tadopting private PaaS approach, and companies that do not consider using PaaS service.

Interview results 5

4

3

2

1 Number ofcompanies Number 0 Clients considering to Clients considering to Clients that do not use public PaaS service use private PaaS consider to use PaaS approach

Figure 29 Interview results

In the following paragraphs, general overview of the interview results is presented.

Page | 55

Chapter 4

Company C1 Company C1 is a Dutch utility company. Company C1 has not used any cloud services until this moment. One of their main concerns is due to the government regulation concerning the location of customer data. They see this issue as a primary barrier to adopt Cloud services in large scale. However, at this moment they are considering to do SaaS implementation for their employee communication system using Microsoft Lync. The reason they want to do this SaaS implementation is to renew their communication system with less capital expenditure, and also to have fast implementation. They do not consider using other SaaS application because they think they are not many mature products yet on the market other than Microsoft Lync.

Company C2 Company C2 is a telecommunication company in the Netherlands. In relation with cloud services, Company C2 has used several SaaS services for their CRM and workspace applications. However, these SaaS implementation are localized only to some countries where Company C2 operates. Regarding PaaS, Company C2 prefer to adopt private cloud approach to provide PaaS solutions to their subsidiary companies in different companies. This approach is chosen due to economy of scale reason. As a holding company, Company C2’s goal is to provide IT service towards their daughter companies at the most efficient way. Therefore, Company C2 does not see the benefit of using public PaaS service to fulfill their needs, because they do not want to pass the benefit of PaaS concept towards the providers. However, with regards to other cloud service like SaaS, in order to use the service they are driven by the functionality that the services offer. In general, they prefer to move only the non-critical business applications to the cloud, both either to SaaS or PaaS. They feel that they still need to learn whether these as-a-Service solutions can bring cost reduction to their business.

Company C3 Company C3 is a telecommunication company in the Netherlands. Company C3 IT department serves their other internal departments as internal customers. This company also considers a private PaaS approach in their organization. The reason the use a private PaaS approach is because they do not want to lose control of their critical applications. The utilization of PaaS itself allows them to serve their internal customers with higher availability and security, as well as higher agility for rapid development and customization of applications.

Company C4 Company C4 is a large retail company in the Netherlands. Their IT department serves their internal customers, which consist of several internal departments. With regards to their cloud strategy, Company C4 does not consider to utilize PaaS in their organization. One of the main reasons is because they think the functionality of PaaS is not mature enough to fulfill their IT service demand. They also see government regulations that prohibits them to put the data out of European Union as a barrier to adopt any cloud services. However, they intended to look at the possibility of using SaaS for their “standard” application. What they mean with “standard” application here is application service that is not the core application of their business process. Example of this application is like workspace application.

Page | 56

Towards a Validated Adoption Model

Company C5 Company C5 is a Dutch investment banking. Company C5 has implemented some small SaaS service for their HR system. They did not consider using PaaS service in the near future. The reason is because their current IT strategy is to reduce IT spending by means of standard application packages, which do not require too many customizations. In other words, they are not aiming to build their own SaaS, instead they are aiming to buy off-the-shelf SaaS products. Company C5 argues that “IT is not a differentiating factor in our business, therefore IT related matters should be standard package in order to be cost efficient”. In terms of selecting cloud provider, Company C5 put priority on product maturity and integration with their in house system. They do selection of cloud provider by means of internal assessment procedure.

Company C6 Company C6 is a large insurance company in the Netherlands. Company C6 has implemented some SaaS services. Their main implementation is CRM application implementation from Saelesforce.com. The reason they go for the cloud implementation in this area is to have cost reduction by means of standardizing their IT system. Therefore they aim to align their business process with the standard application package from Salesforce.com. In the future, Company C6 considers utilizing PaaS to migrate their applications to the cloud. They aim to have cost reduction as well as better performance with respect to security and availability by the utilization of PaaS. At this moment they consider to use the Force.com service from Salesforce.com.

Company C7 Company C7 is a large transportation company in the Netherlands. Company C7 has implemented several SaaS service. The biggest implementation is to provide workspace application from Google. Company C7 has also used PaaS service within one of their business unit. They implement Force.com from Salesforce.com. The reason they wanted to use PaaS is to have rapid development and faster time to market to serve their internal customers. They see the biggest challenge of this implementation is not about security, but about the integration issues. They argue that the lack of technical standardization in cloud computing world does not making the adoption of the technology goes as smooth as they expected.

Company C8 Company C8 is an incumbent Dutch financial service organization. Company C8 has not used any cloud services at the moment. But they are looking to adopt the PaaS concept approach to develop their new security application system together with other companies. They think PaaS approach is a good solution to provide common platform for such an open innovation process. Their goal is to achieve cost reduction in the development of the new system. However, they prefer to have private PaaS solution instead of using public PaaS service. The reason lies in two issues. The first issue is due to government regulation about data location, where they cannot put their data outside the country border. Secondly, they have to deal with the security profile of the available PaaS provider. They argue that the level of security they require to realize the security application is much higher then what the public PaaS providers offer at the market at this moment.

Page | 57

Chapter 4

Company C9 Company C9 is a private banking institution in the Netherlands. Company C9 currently has some SaaS implementations in the area of supporting applications, such as HR application from a Dutch provider called Exact. They differentiate their SaaS implementation as dedicated SaaS, which means that they are aware of the location of the data server that contains their SaaS application. At this moment, they do not consider using PaaS, even though they admit that there are some characteristics from PaaS that they find might be advantageous for their organization, such as scalability and no upfront capital investment needed. There are two obstacles that hinder Company C9 to adopt PaaS service. First, they see that, as a financial institution, there is a strict regulation that prohibits them to put data outside the border country. This regulation disable them to use any public cloud service. Secondly, utilizing PaaS means a company needs to build their own application. So far, Company C9 does not build their own applications, instead they buy off-the-shelf applications from vendors. The use of PaaS will introduce management complexity between the application vendors and the PaaS vendors. Therefore, Company C9 sees the opportunity to use public cloud offering for their non- critical business applications, but not for their critical business applications.

4.2.2 Towards a Validated Model In this section, the conceptual model will be validated based on the interview results. By means of Atlas.ti software, we are able to see the importance or the weight of the adoption factors that were mentioned by quantifying how many a factor has been quoted. We set the rule that a factor that was quoted more than one time is automatically approved and therefore validating the model. This limit is chosen due to the limited respondents we have, particularly with the PaaS provider companies. We only have three respondents with the provider companies, therefore setting the limit higher than one would make it difficult to validate the model. While factors that were only quoted once have to be supported by a relevant and strong argument to be approved to validate the model. We justify the relevancy and strength of the arguments based on our knowledge, particularly in cloud computing domain that we built from conducting literature studies at the early phase of this research project. Each factor is counted when every time different respondents quote the factor. In addition, the count also includes how many a factor was quoted from one respondent in order to show the level of importance of the respective factor. Moreover, We deal with the differences of respondents’ answers by weighing up which answer is more dominant among the respondents.

In our analysis of the interview data, we counted the factors that belong to PaaS and SaaS separately. This approach enables easier classification and hence easier comparison between PaaS and SaaS. Each factor that belong to PaaS is labeled with “PaaS” at the end of the factor name, while each factor that belong to SaaS is labeled with “SaaS”.

We will validate each component of the conceptual model proposed in Chapter 3.3. We will begin with service domain, and continue with technology domain, financial domain and organization domain subsequently.

Service Domain As discussed in Chapter 3, the Service domain concerns with application selection factors, which are the factors that client organization takes into account when they consider an application to be moved to the cloud environment. Figure 30 and Figure 31 below show an overview of the

Page | 58

Towards a Validated Adoption Model respondent answers with regard to application selection factors from the client perspective for PaaS and SaaS adoptions. Each of the factor codes shown in the figures contains two types of numbers. The first number (the front number) counts the groundedness of a code (i.e. the number of associated quotations). The groundedness number can be associated with how important the code is compared with the other codes. The second number (the last number) counts the density of a code (i.e. the number of links to other codes). Additionally, each code begins with a code that shows which domain it belongs to (i.e. (S) means the code belongs to the Service domain, while (O) means the code belongs to the Organization domain.)

Both models shown in Figure 30 and Figure 31 show that there are four factors that are important in deciding which application to be moved to the cloud by utilizing PaaS or SaaS. Those are the factors of data sensitivity handled by the application (code: (S) Data sensitivity handled (PaaS/SaaS)), the criticality of the application to the business (code: (S) Criticality of application (PaaS/SaaS)), the target end-users of the application (code: (S) Target end-users (PaaS/SaaS)) and the added value that client companies expect by utilizing PaaS or SaaS (code: (S) Value elements for internal organization (PaaS/SaaS)). The factor Value elements for internal organization ex[lains the clients perceive several reasons to migrate their application to the cloud. These reasons are differ to some extent between PaaS and SaaS adoptions. The different types of reasons (perceived value) that client organizations have are depicted in Figure 33 and Figure 34.

Figure 30 Application selection factors for PaaS adoption

Figure 31 Application selection factors for SaaS adoption

Page | 59

Chapter 4

The Criticality of application factor is the most important factor in the application selection. Almost all of the companies consider putting their non-critical applications to the cloud, both in utilizing PaaS and SaaS. The examples of the non-critical applications are workspace applications, HR applications and collaboration applications. Almost none of the companies consider having their critical application in the cloud. The main reason is because the critical applications mostly contain critical data about the client company and the customers. That is why the Data sensitivity handled factor and the Criticality of application factor are closely associated. In the Netherlands, there is law that prohibits these sensitive data to be hosted or located outside some specific country border. Some are not allowed to be hosted outside the Netherlands (i.e. some applications that contains customer information in financial service industry), and some are not allowed to be hosted outside European territory, (i.e. the data about the employees information of a company). These issues have become one of the main barriers behind the reluctance to adopt any cloud services.

Based on the groundedness number, the Target end-user factor was not seen as a crucial factor by many client companies. However, we argue that this factor is still valid and an important factor based on the interview conversations. Company C6 (for PaaS) and Company C9 (for SaaS) argue that this factor puts different requirements on the quality of service (QoS) of the application. The companies would require better QoS for end-customer facing applications than the applications that face internal end-users.

Moreover, from the figures above, we can also see that in SaaS adoption, the client companies have an additional factor to take into account. That is the functionality of the SaaS application (code: (S) Functionality (SaaS)). This factor is associated with the maturity of the SaaS product (code: (S) Product maturity (SaaS)). It can be seen that the functionality factor is a dominant factor in SaaS adoption. This result can be understood because the functionality of SaaS is comparable with the functionality of the IT applications that those companies currently use. Each SaaS from different SaaS providers offers different functionalities. While with regards to PaaS, almost all of the PaaS services serve the same functionality, which is to build a customized SaaS. Based on the interviews, the importance of functionality in SaaS adoption is also based on the fact that client companies are looking for standard application packages from SaaS. This is also one of the reasons why not that many client companies are interested in utilizing PaaS, because their company strategy is not to develop their own application and minimize customization. This argument can be found from the discussion with Company C4, Company C5 and Company C9.

Furthermore, we will identify the knowledge gap in the perception of the clients and the vendors. Figure 32 shows the application selection factors from the vendor perspective. Those factors are the criticality of applications to the client business (code: (S) Criticality of application), the data sensitivity handled by the application (code: (S) Data sensitivity handled), the utilization pattern of the application (code: (S) Utilization pattern), the target end-users of the application (code: (S) Target end-users) and the value perceived by client organizations (code: (S) Value for Organizations). Compared with the PaaS application selection factors from client perspective depicted in Figure 30, there are a lot of similarities. However, there is a difference. Vendor thinks that utilization pattern (i.e. peak utilization of an application during summer holiday) is another point of factors that is important in deciding whether to adopt PaaS or not, because this factor corresponds with the scalability feature of PaaS solution. In contrast, this factor has never been a point of discussion during the interview with the clients. Based on the interview conversations, we argued that this is related to

Page | 60

Towards a Validated Adoption Model the type of the applications that the clients want to migrate to the cloud. Most of the non-critical applications do not have the characteristic of unstable workload. There is no peak workload on those non-critical applications. Moreover the inclusion of Target end-users factor despite its number of groundedness is based on the reason explained before in the previous paragraph.

Figure 32 Application selection factors for PaaS adoption from vendor perspective

Furthermore Figure 33 and Figure 34 below show the different types of the value that are perceived as benefits by client organizations that are used as reasons to adopt PaaS and SaaS respectively. The main value that serves as the main reason for organization to adopt either PaaS or SaaS is the value to reduce cost of IT service provision (code: (S) Cost reduction). Client organization perceived the ability of PaaS and SaaS solution to reduce their cost of IT service provision as the main reasons to adopt these technologies. The benefit of cost reduction can be further distinguished between reduction of infrastructure maintenance cost (code: (S) Reduce infrastructure maintenance) and reduction of capital investment (code: (S) Reduce capital investment).

Page | 61

Chapter 4

Figure 33 Value elements for internal organization of PaaS adoption from client perspective

Figure 34 Value elements for internal organization of SaaS adoption from client perspective

There are many similar values perceived by client organization that are used as reasons to adopt PaaS and SaaS. One of them is the value of faster time to market (code: (S) Faster time to market). Client notices that by utilizing PaaS or SaaS rapid development and faster time to final implementation can be achieved. With regards to PaaS adoption, faster time to market is associated

Page | 62

Towards a Validated Adoption Model with bigger chance to cope and take advantage with new opportunities in the market (code: (S) Take advantage from new opportunities).

The following values are also seen as some similar benefitting values for both PaaS and SaaS adoptions. They are the global availability of IT application service through the Internet (code: (S) Availability of IT resources), enabling a standard business platform throughout the organization (code: (S) Std business platform (PaaS/SaaS)), providing a standard communication platform throughout the organization (code: (S) Std communication platform (PaaS/SaaS)), and the ability to quickly scale the IT demand (code: (S) Scalability (PaaS/SaaS)). Additionally, client companies also perceive that by utilizing PaaS and SaaS they could get better service in terms of availability, reliability and security, than their own on-premise solutions (code: (S) Service improvement (PaaS/SaaS)).

However there are some distinguish values perceived by organization from PaaS and SaaS adoption. First, PaaS adoption gives client organization to focus on the innovation of application development (code: (S) Focus on innovation (PaaS)). This value can be understood due to the nature of the PaaS service. PaaS service allows clients to develop their own SaaS application. Thus by providing the run- time environment around the application, one can focus to develop the functionality of the application instead of focus on maintaining the environment, while one cannot find this value from SaaS offering, because SaaS is offered in a form of standard package with little customization possibility. Clients do not have to develop or innovate anything when they adopt SaaS. The SaaS providers are the one that have to do the innovation.

Secondly, client companies see the value that PaaS solution provides the best skill in providing security, availability and reliability than in-house development (code: (S) Skilled resources (PaaS)). However, we expected this value to appear as well in the SaaS adoption perspective. Because besides SaaS solution also provide better security, availability and reliability than in-house implementation due to the same nature of cloud computing technology used in PaaS, SaaS solutions even have additional value. SaaS solutions are offered to provide specific functionality, which means that client organization would benefit from well-developed application. Thus client companies do not have to posses the resources to develop the functionality of the application itself. We see that this is a value that SaaS clients do not thoroughly recognize from SaaS offering. Nevertheless, based on our results, we can see that PaaS solution posseses more perceived values than SaaS solutions.

Moreover, we compare the value of PaaS offering perceived by clients with the value of PaaS offering proposed by the vendors (as shown in Figure 35). Most of the proposed values are similar with the perceived value shown in Figure 33, such as the value of Faster time to market, Cost reduction, Scalability, and Availability of IT resources. However, from the comparison, one can observe that vendors put emphasize on the value of business agility of their offering (code: (S) Business agility). Business agility means the ability to adapt and to change a business process rapidly by means of utilizing PaaS. It is closely related to, but not merely just the value of Faster time to market. While it seems that from the client perspective, what matters the most in PaaS adoption are always related to cost reduction.

Page | 63

Chapter 4

Figure 35 Value of PaaS adoption for organization from vendor perspective

The comparison also tell us that there is one value that is not well perceived by the client customers with regard to PaaS adoption (i.e. value of Business agility), while on the other hand there are also some values that were not really proposed by the vendors, but perceived by the clients (i.e. value of Std business process platform and Std communication platform).

Technology Domain This section concerns with the technology domain of the conceptual model. The important issue here is how the technology requirements and the service domain are related and influenced. The figure below shows an overview of answers that explain the relation of technology requirements and the service domain of the model. Figure 36 shows the model that focus on PaaS adoption, while Figure 37 shows the model that focus on SaaS adoption.

Figure 36 Technology domain for PaaS adoption

From the results above we can analyze that the relation between the factor components from the service domain with the technical requirements of PaaS service. It can be seen that besides the technical requirements of QoS (code: (T) QoS requirement (PaaS)) (i.e. service availability and

Page | 64

Towards a Validated Adoption Model reliability), the security (code: (T) Security requirement (PaaS)), and the system integration possibility with on-premise systems (code: (T) System integration requirement (PaaS)), the service domain draws an additional requirement of data location (code: (O) Data location requirement). Even though this is not really a technical requirement, but this requirement is seen to be important from the client organization perspective. Since Company C9 argued that “Online banking is very important and can be in such a platform (PaaS), but this is the core of the bank, it is going to be always within the bank itself”, we can conclude that the Criticality of the application factor influences where the data location should be. The Data location requirement thus further causing the selection between private (code: Private Cloud (PaaS)) and public cloud (code: Public Cloud (PaaS)) approach.

The Target end-users factor influences the QoS requirement and the Security requirement. Company C4 argued “Some applications must have high quality of service, especially when we have to deal with end customers. It also must be reliable and have no security breaches. We are not accepting any drawbacks in this area”.

Moreover, System integration requirement is driven by the type of the value that is perceived by customers, namely the value of Faster time to market. Based on the conversation with Company C7, one can argue that the attractiveness of PaaS, namely to have rapid development and thus faster time to market, cannot be achieved when the cloud and on premise system is not highly integrated.

Figure 37 Technology domain for SaaS adoption

Furthermore, the Figure 37 above shows the technology domain of SaaS adoption. Compared to Figure 36, the factors in technology domain of SaaS adoption are quite similar. They include the QoS requirement, the Security requirement, the System integration requirement, and the Data location requirement. Based on the interviews, we argue that the similarity is caused by the similarity in the technology principles between PaaS and SaaS. Both of the services has the same characteristics in technology domain, namely the multitenancy and internet accessibility natures. One difference occurs in the factor that drives the System integration requirement (SaaS). In SaaS adoption, system

Page | 65

Chapter 4 integration requirement is directly driven by the Criticality of application factor. As argued by Company C8 and Company C9, that when they implement SaaS for non-critical applications, the applications should be loosely coupled with the critical application to minimize interferences and dependencies.

Moreover, the interview results with the vendor companies regarding the Technological domain factors did not give indication of their relationship with the Service domain. The vendor companies see that the Technological domains such as security and QoS are the kind of requirements that they have to fulfill in high standard continuously. They maintain these high standards by complying with some world standard (i.e. SAS 70 type II for security compliance). Some vendors see that their QoS as a differentiator from their competitors. With regard to the latter, this issue is more relevant to be discussed from the perspective of Organizational domain, which will be further elaborated in the following paragraphs.

Financial Domain The financial domain concerns with the cost structure and the particular revenue model that client companies apply when they adopt the cloud services. Apparently based from the interviews, none of the respondents has specific revenue model scheme towards their end users when they consider utilizing any cloud services. Most of the client companies also argue that the financial reason behind the decision to use or to not use cloud services are the same with other IT investment decision. Which means that the financial aspect of the business case of the cloud solution is the most important factor in the financial domain. From the financial perspective, they argue that the solution that can provide the lowest cost is all what matters at the end. Moreover, the budgeting to deliver the cloud services, either PaaS or SaaS, (will) adapts the same scheme as the way they run their current on-premise IT systems. The cost will be taken from the departmental budget that is already allocated for their IT operations. Thus we argue that financial aspect of PaaS and SaaS adoption is a necessary issue, but not a sufficient condition to differentiate cloud adoption from the other type of IT investment. As quoted from the discussion with Company C5, “The aim is to reduce cost. The cloud seems to be a vehicle or a solution to reduce the cost. So it’s the enabler to the IT strategy”.

Nevertheless, the only relevant issue in financial domain is related to the Pricing of PaaS and SaaS services. The pricing is relevant when the client companies are about to select their desired cloud providers. Therefore, the depiction of the pricing factor is presented in the following section of organization domain.

Organization Domain The organization domain concerns with the partner selection to realize the public PaaS and SaaS services. First of all we will elaborate on the factors that determine the selection of public cloud vendors or providers. Subsequently we will also identify the third party partners that clients believe to be important in adopting PaaS or SaaS services.

The Figure 38 below shows the decisive factors in the PaaS vendors’ profile that clients take into account in selecting the PaaS vendors/providers.

Page | 66

Towards a Validated Adoption Model

Figure 38 Vendor selection factors for PaaS adoption

From the results shown above, one can observe that branding of the PaaS provider towards the client companies (code: (O) Branding (PaaS)) plays important role in vendor selection. This factor of Branding is associated with trust (code: (O) Trust (PaaS)). Based on the interviews, we understood that branding is important because using public PaaS service means trusting their application data to be handled by the vendors. Most of the client companies argue that the incumbent PaaS vendors such as Google, Microsoft or Salesforce.com are associated with good branding due to their good track records, and that is why those vendors are trusted. The client companies also believe that vendors with good brand typically are too big to be out of their business, thus the perceived risk is less when they choose to partner with these type of vendors. That is also a reason why those big vendors are trusted.

Another interesting point of discussion is about the security profile (code: (O) Security profile (PaaS)). What really matters from this issue is whether the vendor is compliant to the world security standards (i.e. SAS 70 type II). So in selecting the vendors, client companies do not differ one vendor to the others as long as the vendors have fulfilled security standard compliancy. That is why the factor of Security profile is associated with the factor of Security standard compliancy (PaaS).

The ability of the platform to integrate with on-premise system (code: (O) System integrator profile (PaaS)) is also an important factor in PaaS vendor selection. The type of lock-in that a PaaS provider introduce towards their customer (code: (O) Vendor lock-in profile (PaaS)) is also a determinant factor in selecting PaaS provider, because each of the vendors contain different level of lock-in (i.e. programming language lock-in, or data lock-in). Another differentiating factor that influences vendor selection is the service availability, stability and reliability (code: (O) QoS (Quality of Service (PaaS)) profile (PaaS). This factor corresponds with the requirement in Technology domain, namely the factor of QoS requirement. When high QoS required, client companies would demand SLA that guarantee the high availability and reliability of PaaS service. The QoS profile is seen as one of the most important factor in vendor selection, because the reliability of the platform enables the quality of the service that can be delivered after all.

Page | 67

Chapter 4

Moreover, there is one factor from financial domain that determines the vendor selection. This factor is pricing (code: (F) Pricing (PaaS)). Pricing of the PaaS solutions would definitely determine which vendor is going to be selected. Even though all of the vendors adopt the same pricing scheme (i.e. pay-as-you-go pricing scheme), but ultimately the unit price will differ from one vendor to the other. Even though Pricing is not the most important selection factor, but the price of PaaS solution will ultimately justify the economic reason behind the PaaS adoption.

Furthermore, comparing the PaaS vendor selection factors with the SaaS vendor selection factors (as depicted in Figure 39), most of the factors are quite similar. Those similar factors are Branding, Pricing, Security profile, System integration profile, QoS profile, and Vendor lock-in profile.

Some differences lie in the programming language factor (code: (O) Programming language (PaaS)), the flexibility profile (code: (O) Flexibility profile (PaaS)), the functionality profile (code: (O) Functionality profile (SaaS)), and the installed base factor (code: (O) Installed base (SaaS)). Based on the conversation with Company C6, the Programming language factor is relevant to select PaaS vendor because it is an important factor when a company wants to migrate their existing IT system from on-premise infrastructure to the cloud. When the client company wants to implement some on- premise applications to the PaaS environment, the company will be able to reuse the same application data that is developed in a certain programming language, if the desired vendor adopts that particular programming language as well. While the Flexibility profile corresponds with how flexible the platform can be offered to the customers. As argued by Company C9, PaaS is a standard platform offered in the cloud environment. The ability of a platform to be able to customize their offering towards the client companies would also determine the vendor selection.

Figure 39 Vendor selection factors for SaaS adoption

The functionality profile (code: (O) Functionality profile (SaaS)) becomes important in selecting the desired SaaS vendor. As argued by Company C2 that the most important one in selecting vendor will always be its functionality. This selection factor is aligned with the Functionality factor in the Service domain. As quoted by Company C4, “First we are looking at what we need in terms of functionality,

Page | 68

Towards a Validated Adoption Model then we are looking what kind of providers there are offer the functionality we want”. Furthermore, the Installed base factor also becomes a distinguishing point in SaaS vendor selection. The installed base here is related to the familiarity of the client company with a certain technology. This factor becomes apparent based on the discussion with Company C1. Company C1 stated that the reason they are only considering Microsoft to provide their collaboration application in the cloud is because their employees are already used to using the old version of collaboration application software from Microsoft.

Moreover, the Figure 40 below shows the PaaS vendor selection factors from the vendor perspective. This result can be seen as the factors that vendors see as differentiators of their PaaS offerings in comparison with their competitors. From the results we can observe that there are some similarities with PaaS vendor selection factors from the client perspective, namely about Branding, Pricing, Programming language, Vendor lock-in, QoS profile, Security profile and System integration profile. Here, Branding is also associated with Trust, and furthermore Trust is associated with Transparency (code: (O) Transparency). Vendors see that being transparent about their activities behind their service provision is the key to gain trust, and therefore to gain good branding. Moreover, other than reliability and availability, vendors see processing power (code: (O) Processing power) is also a part of QoS profile.

However there are some additional factors that were not mentioned by client companies like shown in Figure 38. These additional factors are more technical oriented such as the Service modularity (code: (O) Service modularity), and the difference in the PaaS Technical functionality (code: (O) Technical functionality). During the conversation, Company C3 stated that the detail functionalities behind the PaaS offering, such as relational database and back-up & restore functionality are differing from one PaaS offering to the other. These technical functionalities is then a relevant point to consider which PaaS offering to use depending on what client wants to do with PaaS in technical level. While the Service modularity factor deals with the ability of a platform to offer only specific functionality of their platform instead of offering only the full functionality of the platform.

Nevertheless, based on these additional factors, we argue that the clients have not perceived all factors that distinguish one PaaS offering with the others, such as the Service modularity and the Technical functionality. Even though the vendors see that these factors are quite important in deciding which vendor to partner with. The nature of these last two factors that are more technical oriented could be the reasons of the knowledge gap, because mostly our respondents are business people that are not too familiar with the detailed technical intricacies of such a technology offering.

Page | 69

Chapter 4

Figure 40 PaaS vendor selection factors from vendor perspective

Furthermore, from the interview results we also identify who the most important third party actor is in PaaS and SaaS provision. The result shows that System Integrators are the most desired third party that is important in helping client companies in adopting PaaS service. This result can be well understood because PaaS gives client companies the ability to have custom development. But, the client companies does not necessarily have the resources to do that custom development by themselves. Thus System Integrators can play their role in helping the client companies in utilizing and customizing PaaS solutions and acting as implementation partner

In addition to the clients’ perspective, despite the role of system integrators, PaaS vendors also acknowledge the importance of ISVs (Independent Software Vendors) in PaaS adoption. Vendors see that ISVs play important role in accelerating the adoption of PaaS through their involvement in the PaaS marketplace. Each PaaS offering has its own marketplace where clients can buy developed applications or applications building blocks. These developed applications and building blocks are built by ISVs that use the marketplace as marketing channel. Company V1 argue that the existence of marketplace is important for client companies when they utilize PaaS, because within the market place one can find code and samples that he can use to accelerate his application development. On the other hand, clients do not recognize the importance of marketplace and therefore ISVs. This issue is another example of the knowledge gap between vendors and clients with regard to PaaS offerings.

4.2.3 Further Analysis of Interview Results on Within this section, we will analyze the interview results further than just validating the adoption factor model proposed in Chapter 3.3 as presented in the previous section. We will start with analyzing the pattern that occurs in the selection between PaaS and SaaS vendors. Subsequently we will analyze the perceived risks and barriers from the adoption of PaaS and SaaS services. Since we noted that some companies were still hesitant to move to the cloud, therefore discussion about the perceived risks and barriers are relevant.

Page | 70

Towards a Validated Adoption Model

Desired Vendor Selection Patterns During the interview, the respondents were asked about their general cloud adoption. Most of the companies are using or going to use SaaS in a short term. We list the names of SaaS vendors that have been mentioned, and compare it with the names of PaaS vendors that have been mentioned by the companies that are interested in using public PaaS. Figure 41 and Figure 42 depicted the names of the desired PaaS and SaaS vendors.

Figure 41 Desired PaaS vendors

Figure 42 Desired SaaS vendors

There are only three names of the desired PaaS vendors that were mentioned during the interviews, they are Salesforce.com (with its PaaS offering: Force.com), Cordys, and Google (with its PaaS offering: Google App Engine). While there are eight desired SaaS vendors were named, namely Stonebranch, Proquro, Risk matrix, Exact, Centric, Google (i.e. Google Apps), Microsoft (i.e. Office 365), and Salesforce.com (i.e. the CRM application from Salesforce.com)

From the results shown, one can observe that there are small and niche vendors that were named for SaaS providers. While on the other hand, we do not see many of small and niche vendors named to deliver PaaS service. Company C9 said during the interview, “In selecting PaaS providers, branding is for sure very important. I don’t think it would be a very small provider. But in selecting SaaS providers, it could be a smaller company or a specialist”. From this quotation we can understand that the functionality of SaaS service is really a determinant factor in selecting SaaS providers. This issue allows smaller vendors that offer specialist solutions to have a market share. The SaaS vendor selection factors model depicted in Figure 39 supports this argument. The factor of the functionality profile is one of the most weighted factors in SaaS vendor selection.

In contradictory, PaaS market gives a little room for small vendors, particularly from the large enterprises perspective. The reason that Company C7 chooses big PaaS vendors is due to their branding. Company C7 wants to have the least risk possible in terms of security and reliability of the

Page | 71

Chapter 4

PaaS offering. It is also argued by Company V3 that the concept of PaaS service is relatively newer than SaaS service, and also the level of abstraction of PaaS offering is harder to grasp than SaaS offering. Therefore it does make sense that for the companies that are willing to adopt PaaS, they tend to choose the “big” names on the market rather than specific functionalities.

Barrier Figure 43 and Figure 44 depict the perceived barriers to adopt PaaS and SaaS from the client perspective respectively. There are a lot of similarities between the two results. The results show that the clients perceive the regulation of data location (code: (Barrier) Regulation of data location (PaaS/SaaS)) to be the main barrier for both PaaS and SaaS adoption. Most of the client companies are not putting their critical applications in the cloud because of this specific reason. In addition, this finding makes clear that Regulators have an important role in the cloud computing world.

Figure 43 Perceived barriers to adopt PaaS from client perspective

Figure 44 Perceived barriers to adopt SaaS from client perspective

Page | 72

Towards a Validated Adoption Model

Moreover, losing the control of their application and data becomes the other main barrier (code: (B) Losing control (PaaS/SaaS)). However, based from the discussion with Company C2 and Company C6, they believe that this issue will fade away in the future as people will be get used to cloud solutions. Yet, at this moment, this issue is still a big barrier from the internal organization itself in order to adopt PaaS and SaaS solutions. The lack of technical standardization in the cloud computing world (code: (B) Lact of technical standardization (PaaS/SaaS)) that relates to the creation of vendor-lock (code: (B) Vendor-lock in (PaaS/SaaS)) is also perceived as barrier that makes companies hesitant to adopt PaaS and SaaS. The client perceived that early adoption would increase the potential of unnecessary vendor-lock in, because there are no easy exit strategies due to the lack of technical standardization.

There are some other similar barriers perceived in both PaaS and SaaS adoption. Even though some client respondents admit that PaaS and SaaS solutions might offer higher level of security, but those client admit that the companies’ higher ups sometimes still feel unsecure for putting their data outside their control (code: (B) Feeling unsecure (PaaS/SaaS)). Moreover, there is also some human factor that was perceived as obstacle to adopt PaaS and SaaS (code: (B) Organizational issues (PaaS/SaaS). As argued by Company C7, some IT people of an organization see that the cloud development is something that may cost their jobs at some point. These two issues are perceived as barrier that comes from the internal organization.

Nevertheless, the adoption of PaaS highlights the manageability issue as the other barrier (code: (B) Manageability issue (PaaS)). As argued by Company C9, when companies are looking for standard off-the-shelf applications, it is hard for them to implement those standard applications in the PaaS environment. Because it means that they have to deal with two types of providers, namely the SaaS provider that manages the applications and the PaaS provider that manages the runtime environment. Both of the providers might have different update cycles that are cumbersome for the companies to manage. Additionally, the companies that take private PaaS approach also see that the economic reason of PaaS solution is a barrier to use public PaaS offering (code: (B) Economic reasons: economy of scale (PaaS)). They see that as long as private PaaS solution is cheaper than public PaaS solution, they will not use public PaaS solution. The nature of these companies that serve large scale of internal customers (as a holding company) make it feasible for them to make their own economy of scale with their own infrastructure, thus making private PaaS solution is cheaper for them.

Moreover, the adoption of SaaS also has a particular barrier. Some companies see that the nature of SaaS offering that has limited possibility of modification is a barrier (code: (B) Standard offering (SaaS)). The standard package that SaaS offers might not fulfill their customer demand, because typically the companies’ customers (i.e. the business units) demand some modification with their applications so that the application can be aligned with their particular business process.

Furthermore, Company V3 also argued that the nature of the PaaS service itself becomes a barrier of its adoption by large enterprises. PaaS service allows custom development of cloud application, while on the other hand a lot of companies are not doing custom development anymore. They want to go to IT vendors to get their standard off-the-shelf application in order to reduce customization and thus reduce cost.

Page | 73

Chapter 4

Risk Figure 45 and Figure 46 depict the perceived risks from PaaS and SaaS adoption respectively. Similar to the discussion of barriers previously, there are some similarities between the risk from PaaS and SaaS adoption perceived by the clients. Clients are still hesitant on the security level (code: (Risk) Security (PaaS/SaaS) and the performance level (in terms of reliability and availability) (code: (R) Reliability and availability (PaaS/SaaS)) of the service provision. They still see that these issues are the main risks they face when they adopt cloud solutions.

Figure 45 Perceived risks from PaaS adoption (client perspective)

Figure 46 Perceived risks from SaaS adoption (client perspective)

Some companies are worried with their reputation (code: (R) Reputation (PaaS/SaaS)). As they are losing the control of their applications when they using PaaS and SaaS service, failure in the service delivery can cost them their reputation. This is also another reason why clients are still hesitant to put their critical business application in the cloud. The challenge to manage an organization’s cloud portfolios is also seen to be a risk in adopting both PaaS and SaaS solutions (code: (R) Manageability (PaaS/SaaS)). Additionally, the risk whether the investment on cloud solution will pay off is perceived as common risk with any IT investments (code: (R) Economic reasons (PaaS/SaaS)).

The difference between SaaS and PaaS domain in terms of risk is that PaaS domain contains a risk of early vendor lock-in (code: (R) Vendor lock-in (PaaS/SaaS)). As the market and the technology of PaaS still develop, commitment to one vendor can cost an unnecessary lock in to a certain vendor. This issue is closely related to the lack of standardization issue raised in the previous paragraphs.

Page | 74

Towards a Validated Adoption Model

Roles of Telecommunication Companies in PaaS and SaaS Provision During the interviews, we ask a question regarding their opinion of the possible role that telecom companies can play in the provision of PaaS and SaaS services. We ask what kind of value-added activities that telecom companies can offer in order to enhance the adoption and service of PaaS and SaaS.

Based on the conversations with both client companies, we identified four types of roles that they think telecom companies can offer a specific value-added activity in relation to the cloud computing domain. These identified roles are depicted in the Figure 47, namely the role to provide reliable Internet network connection for cloud service usage (code: Providing reliable network); the role as broker/mediator between PaaS/SaaS provider and the clients (code: Broker/Mediator); the role to enhance the security through the Internet network (code: Enhancing security); and being the PaaS/SaaS providers (code: PaaS/SaaS providers). One important thing to note, the last role was mentioned by Company C2 and Company C3 during the interviews, which are both telecommunications companies. Thus expressing a sign of feasibility of this role.

However, providing a reliable network connection seems to be an obvious role of telecom companies, not only in cloud computing domain, but also in generic telecommunication domain. We argue that this role is still making telecom company as “bit piper”, thus not addressing our main concern of our research question. We will discuss the role of telecom companies in cloud computing domain in more detail in Chapter 5.

Figure 47 Identified potential role of telecom companies in the cloud service provision (based on interviews)

Additionally, we can interpret this finding that telecom companies is also acknowledged to be an important party in PaaS/SaaS service provision. The result shows that the most important role of telecom companies in cloud computing service provision is to provide reliable (Internet) network connections (code: Providing reliable network connection). Thus this result addresses the importance of Network Provider as key stakeholders in cloud computing.

Page | 75

Chapter 4

4.2.4 Summary The generic model of adoption factor that was presented in Chapter 3.3 is shown in Figure 48 below for comparison purpose. Within this section, summary of the validation of this model is going to be presented. Conclusion of the general analysis of the interview results is also presented subsequently.

influence Service Domain Technology Domain PaaS Providers’ branding (trust) Target of the Organization Domain application service PaaS Providers’ Security PaaS Provider customer defines influence the Selection lock-in mechanism Value Elements for requirements for influence end-users and Partner Selection PaaS Provider’s internal Quality of Service security profile organization Complementary Actors PaaS Provider’s The goal to Quality of Service Level of criticality of build cloud System profile work done by the application Integration applicattion service service PaaS Providers’ system integration Level of data profile (software Financial Domain sensitivity handled standard) by the application service Value-added Pricing IT Budget resources / capabilities Level of utilization pattern of the Cost Budget application service Revenue Model

Figure 48 The integrated conceptual model of PaaS adoption proposed in Chapter 3.3

Validated Model Figure 49 shows the validated factors of PaaS adoption model. The grayed components are new factors in addition to the generic conceptual model. The model shows four different domains, namely Service domain, Technology domain, Financial domain, and Organization domain. The Service domain concerns with application selection factors, which are the factors that client organization takes into account when they consider an application to be moved to the cloud environment. Some of this factors influence factors of requirement in Technology and Organizational domains. These factors of requirement then influence the factor of Partner Selection within the Organizational domain. Partner Selection consists of two aspects: PaaS Provider Selection and selection of Complementary Actors. The PaaS Provider Selection is influenced by some factors of profile that are possessed by the PaaS vendors.

In comparison with the generic conceptual model, the factor of Value elements for end-users and Level of utilization pattern of the application service within the Service domain are eliminated. The respondents have never raised these issues as an important issue during the interview. They see that there are no added values for end-users when they choose to implement PaaS service, and the respondents also did not acknowledge the utilization pattern (i.e. peak of IT resource utilization due to summer holiday) to be the main important factor in deciding which application to be moved to the cloud. The latter issue might happen because the PaaS values might not be all propagated to the clients. Another reason could also be something to do with the type of the application that clients typically consider moving to the cloud at this moment. Clients only consider moving non-critical applications, and theses non-critical applications might not have unique characteristic with regard to its utilization patter. Moreover, the components in the Financial domain such as Revenue model and IT budget were not found to be decisive to influence Partner Selection, thus these components are also eliminated.

Page | 76

Towards a Validated Adoption Model

influence Technology Domain PaaS Providers’ Service Domain branding (trust) (application selection factors) influences System Integration Organization Domain PaaS Provider’s lock- requirement PaaS Provider Value elements for in profile influences Selection internal Partner Selection organization Quality of Service PaaS Provider’s requirement security profile y g e t

a Target of the PaaS Provider’s r t

s Security Quality of Service

application service T

I requirement profile Complementary s

’ influences y

Actors n s a e PaaS Providers’ p c n m Level of criticality of system integration e o u

l work done by the C f Data location profile n

i applicattion service requirement Programming language Level of data sensitivity handled influences is a is a by the application influences service Flexibility profile

System Network Financial Domain Regulators Integrator Provider

Pricing Important Stakeholders

Figure 49 Validated model of PaaS adoption factors

The validated model shows that each component of the Service domain (i.e. Value elements for internal organization, Target of the application service, Level of criticality of work done by the application service, and Level of data sensitivity handled by the application service) draws requirement for the Technical domain, namely System integration requirement, Quality of Service requirement, and Security requirement. Additionally, the component in the Service domain also draws another requirement of Data location (Organizational domain). The Data location requirement exists due to regulation influences. These all technical and data location requirements influence the factors to select the desired PaaS providers. These factors include similar factors that were defined in the generic conceptual model, such as PaaS Provider’s Branding, PaaS Provider’s lock-in profile, PaaS Provider’s QoS profile, PaaS Provider’s system integration profile, PaaS Provider’s security profile and a factor from the Financial domain: Pricing. In addition to those factors, PaaS Provider’s Programming language and Flexibility profile also become important factors of PaaS vendor selection with the reasons that have been discussed previously in the section 4.2.2 (organization domain). In addition to the PaaS providers, client organizations also think that the role of System integrator is a necessary complementary actor to help them adopting the technology. The reason is because the client companies do not necessarily have the resources to do the PaaS implementation, while system integrator has the needed resource. PaaS clients also acknowledge other stakeholders, such as Regulators (i.e Government and Dutch National Bank) and Network Providers (i.e. telecom companies) to be important in the ecosystem of PaaS service provision.

Moreover, Figure 50 below shows the validation model of SaaS adoption that is comparable with the PaaS adoption model shown before. This model was also validated based on the interview results with client companies using the generic conceptual model shown in Chapter 3.3.

Page | 77

Chapter 4

influence Technology Domain PaaS Providers’ Service Domain branding (trust) (application selection factors) System Integration Organization Domain requirement PaaS Provider’s lock- influences SaaS Provider Value elements for in profile influences Selection internal Partner Selection organization Quality of Service PaaS Provider’s requirement security profile

Target of the PaaS Provider’s

y Security Quality of Service

g application service e

t requirement profile

a Complementary r influences t

s Actors

T I

PaaS Providers’ s ’ Level of criticality of

y system integration

n work done by the a profile s p applicattion service Data location e m

c requirement o n C e influences u l f Installed base n i Level of data sensitivity handled is a is a by the application influences service Functionality profile

System Network Financial Domain Regulators Application Integrator Provider functionality Pricing Important Stakeholders

Figure 50 Validated model of SaaS adoption factors

From the SaaS adoption model above, one can observe that there are a lot of similarities with the PaaS adoption model. The Service domain also contains factors such as Value elements for internal organization, Target of the application service, Level of criticality of work done by the application service, and Level of data sensitivity handled by the application service. A distinguish factor that differ this model with PaaS adoption model is the Application functionality factor of SaaS. The utilization of SaaS relies heavily on the functionality of the SaaS offering. This factor is not really apparent in PaaS adoption factors.

Moreover, the relationship between the Service domain and the factors of requirement (i.e. System integration requirement, Quality of Service requirement, Security requirement, and Data location requirement) in Technical domain and Organization domain also differ from the PaaS adoption model. Here the Value elements for internal organization do not influence any of the factors of requirement. Another difference occurs in the factors that influence the SaaS Provider Selection. In addition to the generic model, Installed base and Functionality profile perceived to be decisive in the selection of SaaS providers.

Furthermore, similarly the SaaS adoption model also takes into consideration the role of important stakeholders such as System integrators, and Network providers to perform complementary role.

Conclusion Towards the Research Sub-Questions Based on the PaaS market condition, we can conclude that there are not many client companies in the Netherlands which or considering using PaaS in the near future. Even though they see a lot of benefits from the utilization of PaaS, they still perceive several barriers related to regulation and security that prohibit them to adopt any cloud computing solutions in a large scale. Some companies take private cloud approach as a way out, while some other companies prefer to slowly adopt the cloud by implementing it for non-critical and standard applications (i.e workspace and communication applications). Hence, the SaaS offering for these non-critical and standard applications showed better adoption than PaaS offering in the near future. Based on the analysis of the companies that already have considered PaaS offering, we suggest that SaaS adoption is a logical step towards PaaS adoption in the future. This argument is based on the fact that those companies considered to utilize PaaS in the near future from the same provider of their current SaaS solution as

Page | 78

Towards a Validated Adoption Model a mean of extending the functionality of their SaaS solution, and also to provide better integration among their newly migrated/developed cloud solutions.

Our suggestion that SaaS adoption might be faster than PaaS adoption is supported by the discussion with Company V2, that said, “The concept of SaaS is easier to grasp. Business users can go to the cloud and manage the application directly online (i.e. to add more users to the CRM system). While PaaS requires more understanding how you build the applications. Thus it takes longer to understand the PaaS concept ”.

Some differences occur in the discussion of Service domain (i.e. value for organizations) and Organization domain (i.e. vendor selection factors) between the client perspective and the vendor perspective as discussed in section 4.2.2. The occurrence of these differences can be argued as a sign that not all of the value proposed by the PaaS providers is well understood and perceived as an important value by the PaaS clients. This issue seems to be acknowledged by some PaaS vendors, as Company V1 and Company V2 argued that educating the market is still one of their top priorities regarding the marketing of PaaS service. Company V2 mentioned, “As an industry we have a task to educate the market”.

Furthermore, with regard to vendor selection, our analysis also showed that small providers are more desired in the SaaS domain compared to the PaaS domain. In PaaS domain, clients consider Branding of the provider to be the most important factor in the selection. While in SaaS domain, small players can penetrate the market easier because clients consider Functionality of the SaaS solution to be as important as Branding.

Nevertheless, some risks and barriers of the PaaS and SaaS adoption have been identified as well. There are many similarities between the main risks and barriers of PaaS and SaaS adoption perceived by the clients. We argued that these similarities are caused by the similarities of the core concept of PaaS and SaaS offering that are based on cloud computing technology.

Finally, this chapter has helped us to answer a partial of sub-question 1 regarding the potential customers of PaaS offerings and sub-question 3 and 4 as the following:

Sub-question 3 “What sort of business software applications do client organizations want to build or migrate by using PaaS to the cloud and how this is related with the generic PaaS-related business model?”

Sub-question 4 “What sort of similarities and differences arise among the identified client PaaS-related business model(s)?

By understanding the type of the software applications client organizations wants to build or migrate to the cloud we are able to validate the generic conceptual model. Furthermore, due to the non-ideal result of our interviews, we are able to validate two adoption models of PaaS and SaaS. Even though we are not directly answering the sub-question 4, but we are able to a give deeper insight on the comparison between PaaS and SaaS adoptions.

Page | 79

Chapter 4

This page intentionally left blank

Page | 80

5 Potential Role of Telecom Companies in the Cloud

This chapter concerns with the research sub-question 5, which is to identify the role of Dutch telecom company in the provision of PaaS service. However, due to the change of scope as explained in the beginning of Chapter 4, we will include a broader perspective that includes SaaS service as well in our discussion

As explained in the Introduction chapter, the reason we focus on the telecom sector is due to the downfall of voice-based revenue stream that telecom operators (both fixed and mobile) faced. Fixed telephony is losing ground to other forms of communication such as mobile telephony and the Internet. While on the other hand the mobile telephony service also received enormous pressure from the explosive growth of mobile internet usages (OPTA, 2011a). Therefore, we would like to investigate whether telecom companies can exploit the cloud computing field to generate a new revenue source and not just being a “bit piper”. Within our discussion, we will only consider Broadband providers and MNOs as telecom companies/telcos (see Appendix B for general information of Dutch Telecommunication market), because these companies are the ones that are able to provide Internet connectivity (either fixed or wireless).

We will start this chapter by elaborating the competences and generic roles of telecom companies based on literature. Accordingly, we will discuss our interview results and confront them with the literature and the domain discussion about the trends of telecom companies’ activity in cloud computing (as discussed in Chapter 2.10) in order to identify the potential role of telecom companies in the cloud computing domain. We will also bring insight of Dutch telecommunication market in our discussion.

5.1 Telecom Companies Competences and Roles

Literature that will be presented in this section will serve as a starting point of our analysis on the telecom companies’ potential role in cloud computing domain. In the following section we will use literature on resources and capabilities of telecom companies to confront it with our interview results. This is performed in order to analyze the potential role that can be played by telecom companies in the Netherlands. We will also use the literature about telecom companies’ generic roles in order to see whether similar roles can be played in the cloud computing domain.

Telecom Companies’ Resources and Capabilities In order for telecom companies to generate new revenue source, telecom companies need to generate new service within the generic cloud computing solution. To provide a revenue-generating service, value-adding activities are required (Bouwman et al., 2008). These value-adding activities should be performed and executed by actors that poses specific resources and capabilities (De Reuver, Visser, & Bouwman, 2011). Resources and capabilities are the sources to have value creating

Page | 81

Chapter 5 strategy that will lead a firm to have a competitive advantage (Barney, 1991). Resources is defined as all assets, capabilities, organizational processes, firm attributes, information, knowledge, etc. that are valuable to a firm and controlled by a firm that enable the firm to conceive of and implement strategies that improve its efficiency and effectiveness (Barney, 1991).

In the telecommunication domain, De Reuver and Bouwman (De Reuver & Bouwman, 2008) classify mobile telecom company resources into access to the network, access to content, and access to applications. Moreover, De Reuver et al. (2011) argue that mobile telecommunication operators, or MNOs (Mobile Network Operators), have important resources and capabilities that can be categorized into tangible resources, intangible resources, and organizational capabilities (De Reuver et al., 2011). The tangible ones cover resources such as network infrastructure, frequency spectrum, third-party access to network, billing system and customer services channels. The intangible resources are customer relation, customer base, customer data, convenience, and trustfulness. While the organizational capabilities include interoperability, standardization, front-to-end control, and guaranteed QoS.

Kuo and Yu (2006) acknowledge some critical resources that are possessed by Telecom companies, especially MNOs. The critical resources include network infrastructure, the dominance of related technologies, popularity of existing brands, established resell channels, a large amount of ultimate users, users’ profile, bill integration ability, and price control ability.

Based on those resources and capabilities, we argue that resources that are related to network infrastructure, network access, and customer ownership are the most relevant resources and capabilities in our cloud computing discussion. Network infrastructure and network access are important because they are related to the provision of the Internet service, which is the backbone of cloud computing service provision. While customer ownership related resources are relevant because these are important type of resources in the ICT-service industries as argued by Weil and Vitale (2001). Customer ownership here for instance includes access to the customers, the customer data, and also customer transaction that involve charging and billing capabilities.

In order to simplify our analysis, we will discern and group the aforementioned resources and capabilities that are similar and relevant in cloud computing discussion under two main categories. Those categories are related to Internet network infrastructure and access, and customer ownership. We will present them in the following consolidated list shown in Table 13.

Categories De Reuver et al. (2011) Kuo and Yu (2006) Internet network Network infrastructure Network infrastructure infrastructure and access Frequency spectrum Dominance of related technologies Third-party access to network Interoperability Standardization Front-to-end control Customer ownership Customer service channels Popularity of existing brands Customer relation Established resell channels Customer base and data A large amount of ultimate users Trustfulness Users’ profile Convenience Bill integration ability Billing system Price control ability

Table 13 Consolidated list of telecom companies' resources and capabilities

Page | 82

Potential Role of Telecom Companies in the Cloud

Nevertheless, we are fully aware that the literature only shows resources and capabilities that are specific for mobile telecom operators (MNOs). When we were looking into the scientific database such as Scopus and ScienceDirect, We found that there are limited number of literature about resource and capabilities of broadband operators. Therefore, we make an assumption that broadband operator companies have the same resources and capabilities as the MNOs such as resources and capabilities that are related to customer ownership and billing, but not including network infrastructure. The reason is that the operation model of those companies are similar, because both of them are telecommunication companies that have huge customer base, have their own customer relations, and charge their subscribers in periodic time. However, each of them has different network infrastructures. Broadband operators have mainly fixed network infrastructure, while MNOs have mainly wireless network infrastructure. So they have different kind of Internet network infrastructures and accesses. We will carry out this assumption throughout our discussion in this chapter.

Telcom Companies’ Roles in Generic Telecom Industry Role-based approach is important in analyzing organization (Barley, 1990). Roles are technologically separable value-adding activities in a business network (Kambil & Short, 1993). Apfelbeck (1998) argue that a player in the telecommunication industry plays roles that encompasses a set of functions, whereas the provision of these functions my depend on the existence of certain resources. Apfelbeck classifies the roles in communication network into four domains: transport, access, service, usage and mediation. These roles include a role of Fixed Access Provider, Mobile Access Provider, Long Distance Carrier, Network Operator, Service Provider, Content Provider, Database Operator, Billing Partner, Subscriber, User, and Trusted Third Party.

Within the context of mobile service platforms, Ballon and Walravens (2008) distinguish four value- adding roles of Telecom companies: Profile/Identity Management, Service Provisioning/Service Brokerage, Charging and Billing, and Service Creation Environment. Profile Management is a component that manages general user data and user preference for various services. Service Brokerage represents the reference point for end users to retrieve, subscribe, and use services. Charging and Billing are essential components when considering service architecture development., while Service Creation Environment represents a set of development and hosting tools for third-party service developers.

Kuo and Yu (2006) derive several roles in mobile telecommunication industry that may be played by MNOs based on their critical resources. They are the roles as value-added service provider, transaction support, mobile portal, solution provider, and intermediator. The first role allows telecom operator to offer value-added service to their customer, for instance in the form of personalized service. The role of transaction support allows telecom operator to make use of its billing capacity to enable billing integration with other services. The mobile portal role enables telecom operator to provide more concise and more personalized information to their customers through their mobile portal interface. The solution provider role refers to a center that provides an overall solution to users that includes multiple service integration. Finally, the intermediator role enables telecom operator to be a place for a “one stop shopping” place for all kind of services for customers, whereas the advantages of brand awareness, large size of customer base, and billing integration are made used.

Page | 83

Chapter 5

Based on the literature surveys we discussed above about the roles of telecom companies in telecommunication industry, we found there are some similarities among the identified roles in different literature. We will discern some roles that are aligned with the scope of this research to focus on the telecom company’s roles only. In order to simplify our analysis, we will aggregate the roles that are closely related. In the following Table 14, we present a consolidated list of generic role of telecom companies in the telecommunication industry (based on (Apfelbeck, 1998; Ballon, 2009; Kuo & Yu, 2006)).

Aggregated Roles Apfelbeck (1998) Ballon & Walravens (2008) Kuo & Yu (2006) Network access provider Fixed access provider Service provisioning Value-added service and operator Mobile access provider provider Long distance carrier Network operator Billing partner Billing partner Charging and Billing Transaction support Mediator Trusted third party Profile/identity management Intermediator Database operator Service Brokerage Value-added service Service Provider Service Provisioning Value-added service provider Content Provider Service Creating environment provider Mobile portal Solution provider

Table 14 Aggregated generic roles of telecom companies in the telecommunication industry

These literature from Ballon & Walravens (2009) and Kuo & Yu (2006) defines role that are typically played by MNOs, while Apfelbeck defines roles that are more general for both Broadband operators and MNOs. The obvious difference we can discern between the two types of operators is that MNOs are used to play a role as mobile access provider, where Broadband operators are used to play a role as fixed access provider. Furthermore, we assume that both types of operator are familiar with the roles of billing partner, mediator and value-added service provider (i.e. personalized bundling service).

Furthermore, in order to create business value that can lead to a sustained competitive advantage firms must base their strategy on their strengths and can only develop their future strengths from today’s strength (De Reuver et al., 2011). Therefore, we will analyze the potential roles of telecom company in the cloud service (PaaS and SaaS), within the next section, by trying to match the opportunity gap we found based on the interviews with the telecom companies resources, capabilities and generic roles we have presented.

5.2 Towards Potential Role in Cloud Computing Services

Our interview result in section 4.2.3 has shown that clients think telecom companies role mainly is to provide reliable network connection (code: Providing reliable network). We argue that this is an obvious role of telecom companies, not only in cloud computing domain, but also in generic telecommunication domain. We argue that this role is still making telecom company as “bit piper”, thus not addressing our main concern of our research question.

Within this section, we will confront our knowledge on telecommunication domain and literature with some of our interview results. By this means, we aim to identify and reason the feasibility of other telecom companies’ roles in cloud computing domain, that will bring business value to the telecom companies and not being just a “bit piper” only.

Page | 84

Potential Role of Telecom Companies in the Cloud

Based on our domain discussion in section 2.10 about the trends of telecom companies’ activities in the cloud, we identified two roles of Cloud service broker and Cloud service provider. We will use these two roles as basis of our discussion.

5.2.1 Role of Cloud Service Broker We argue that telecom companies can play the role of Cloud service broker in the cloud computing domain. In the following paragraph we will elaborate our argument by confronting the literature with our interview results. Accordingly we will elaborate some reflection of this role.

Branding Selection Factor vs. Resource of Customer Ownership First of all, we will discuss our interview results about vendor selection factors for PaaS and SaaS as depicted in Figure 38 and Figure 39 (Chapter 4) respectively. Our results show that the factor of branding (code: (O) Branding (PaaS/SaaS)) is one of client’s most important factors in selecting the cloud vendors they want to use. Branding is closely related to the trust of clients towards the cloud vendors (code (O) Trust). The importance of branding is reflected in our result about the desired vendor that was selected in section 4.2.3. Many clients tend to choose big incumbent vendors such as Microsoft and Salesforce.com.

We can relate this result with telecom companies’ role for being cloud service broker (as mentioned before in section 2.10). In our context, a broker is the single point of contact for an enterprise for all cloud computing requirements such as service provisioning, service level agreements (SLA) and compliance. A broker sits between the enterprise and a single or multiple cloud service vendors and provides a layer of abstraction (Buvat & Nandan, 2010). As discussed in the literature, telecom companies are in possession of customer ownership-related resources (i.e. broad customer base, customer data, customer relation, customer service channel, etc.). By providing generic telecommunication service, telecom companies establish this customer ownership–related resources. Through this generic service, telecom companies also gain their resources of trustfulness and brand popularity. Small PaaS and SaaS vendors could exploit these resources to convey a sense of better trustfulness and better branding towards the clients. PaaS and SaaS vendors can also take advantage of telecom companies’ established marketing channel to reach the extensive customer base that telecom companies have. Telecom companies’ billing capabilities can also be exploited. Billing of the standard telecommunication service and the cloud services can be integrated, hence giving a sense of simplicity toward the clients to manage their service portfolio.

Reflection to the Barrier of Vendor Lock-in In relation with the cloud service broker role, we would like to relate this role with our interview results about the barriers that are perceived by clients in order to adopt PaaS and SaaS solution. These barriers are depicted in Figure 43 and Figure 44 (Chapter 4). The results show that vendor lock- in (code: (B) Vendor-lock in (PaaS/SaaS)) is one of the obstacles for utilizing both PaaS and SaaS. This barrier can be reduced in relation with telecom companies’ role of cloud service broker. As explained before, in our context, a broker can syndicate multiple vendors together towards a single client and manage the provision of all the desired cloud services. In this case, it will enable clients to switch cloud vendors without worrying about the operational details. Thus, brokering would also benefit clients by means of reducing the perceived vendor lock-in.

Page | 85

Chapter 5

Reflection to the Generic Role of Intermediator Additionally, to support our argument of the cloud service broker role, we will relate this with the literature on generic telecom companies’ role. Being a cloud service broker is being a mediator. We argue that this role is feasible for Telecom companies because telecom companies have been playing similar role in the generic telecommunication industry. Telecom operators have been playing a role of service brokerage (i.e. Vodafone Live! and Proximus) (Ballon & Walravens, 2008) and intermediator (i.e. NTT Docomo) (Kuo & Yu, 2006). Therefore, we are convinced that playing a role as broker is a great opportunity for telecom companies to create new business value in the cloud computing domain.

Reflection to the Dutch telecom market Furthermore we argue that this role of being cloud service broker should give the same opportunity for both Broadband operators and MNOs. This is because both of the operator types have the same relevant resources and capabilities for this role, namely the resources that are related to customer ownership, trustfulness and branding, and billing capability.

Looking into the Dutch telecom market (as presented in Appendix B), all of the main players such as KPN, UPC, , Vodafone and T-mobile have the same opportunity to play this role. The thing that might matter to the successfulness of this role is the size of each operator’s market share. The larger the market share, the larger the customer base is. The more customers can be mediated, the more revenue can be obtained. However each operators need to have strategic collaboration with one or multiple cloud providers to be able to adopt this approach successfully. Dutch telecom companies have an edge to be cloud service broker in the Netherlands, because from the perspective of international cloud vendors, Dutch telecom companies have better and deeper understanding of the local market.

5.2.2 Role of Cloud Service Provider We argue that telecom companies can also play the role of Cloud service provider in the cloud computing domain. Similar to our discussion before, in the following paragraph we will elaborate our argument by confronting the literature with our interview results. Accordingly we will elaborate some reflection of this role.

Risks of Security, Reliability and Availability vs. Resource of Network Infrastructure We would like to point out our interview results about the risks that the clients perceived to adopt PaaS and SaaS as depicted in Figure 45 and Figure 46. It is clear based on the results that security (code: (R) Security (PaaS/SaaS)), reliability, and availability (code: (R) Reliability and availability (PaaS/SaaS)) of the PaaS and SaaS service provision are the most notable risks that the clients perceived. Higher level and assurance of security, reliability and availability of PaaS and SaaS service will definitely reduce the risk, and hence increase the adoption of the technology.

Telecom companies’ involvement in cloud computing domain can reduce this risk. Telecom companies have resource of network infrastructures to deliver Internet services that they can fully control. On top of that telecom companies also have control over the technology of their network as a resource. Since Internet is the backbone in accessing any cloud computing services, therefore telecom companies can exploit this resource to enhance the cloud computing service security,

Page | 86

Potential Role of Telecom Companies in the Cloud reliability and availability. This can be done by playing a role of cloud service provider (either PaaS or SaaS).

By being cloud service provider, it will allow telecom companies to deliver a guaranteed high quality and high security end-to-end cloud computing services. On top of the generic PaaS or SaaS service provision, telecom companies can offer more added value through their network connectivity such as enhancing the Internet line security, and also guaranteeing the reliability and availability of the Internet connection through a stringent service level agreement (SLA). This added value activity is something that is hard to be realized by other “pure” cloud providers such as Microsoft and Salesforce.com, because they are not in control of the Internet network infrastructure and Internet service provision.

Reflection to the PaaS and SaaS Service Provision With regards to the role of cloud service provider for SaaS solution in particular, telecom companies can further exploit their billing system as a cloud application to offer. As argued by Company C2 (a telecom company) in our interview, at this moment they already provide billing service towards their affiliates in different countries by means of private cloud approach. In the future, there is an opportunity to provide this billing service toward external parties as a cloud service. Thus it will make the telecom company a cloud service provider.

Moreover, with regards to the role of cloud service provider for PaaS solution, telecom companies can make use of the access to their network to be offered in a form of APIs. Telecom companies can create a PaaS environment where developers or clients can develop their own cloud application and also can make use some functionalities of the network. In the mobile domain, these functionalities can be SMS function or location based function for instance. To play this role, telecom companies can target broader range of market than only their local market, so that international developer community can be attracted to the build application on the platform. Hence, in order to create a successful platform, theory about multi-sided platform (Gawer & Cusumano, 2002b) might be relevant.

Reflection to the Dutch Telecom Market Looking into the Dutch market, we argue that MNOs and Broadband operators can serve different types of PaaS or SaaS services. MNOs can focus on delivering PaaS for mobile service development, because they have the resource of mobile network infrastructure and accessibility. Moreover, the functionalities that MNOs can offer in PaaS (in the form of APIs) are more attractive, such as SMS or location based functionality. The functionality in broadband (other than higher access speed) is quite limited. On the other hand Broadband operators can focus on delivering SaaS for CRM or billing systems, because their fixed broadband network might be more capable in handling the bandwidth of those heavy enterprise applications. The only telecom company that can be active in both of the focuses is KPN, because KPN plays in both mobile and broadband markets. In addition, KPN could also benefit from having the opportunity to make a standardization that integrates their fixed and wireless functionalities within their cloud services. For instance, if KPN build a PaaS environment, KPN can provide the APIs of warless network functionalities (i.e. SMS or GSM-based location service) to be used for non-mobile based application service. So that one can send SMS through one’s CRM application from their PC for example.

Page | 87

Chapter 5

5.3 Conclusion

Several potential roles of telecom companies in the cloud service domain have been identified. This chapter therefore answers the following research sub-question 5,

“What sort of potential role Dutch telecom companies can play in the (potential) PaaS service offerings?”

The telecom companies certainly play a vital role in the cloud computing world, including in the provision of PaaS and SaaS. There is a big dependency on the Internet network connection utilizing cloud services. The interesting point here is how the telecom companies can exploit their role in order to generate new business value and new revenue source.

We identify the problems of cloud computing adoption based on our interviews. Then we analyze the generic role of telecom companies and the relevant resources and capabilities that telecom companies have based on the literature. By confronting the identified problems and the literature, we identify two potential roles that telecom companies can play in cloud computing domain. The first one is to play a role of cloud brokers. This role will position Telecom companies as a mediator between the cloud vendors and the clients. Secondly, telecom companies can play a role of a cloud provider. This role will enable Telecom companies to utilize their Internet network infrastructure and accessibility resources to deliver high level end-to-end cloud computing services, in terms of security, reliability and availability.

We argue that both of the roles are real opportunities for telecom companies to capture business value from cloud computing world, which are supported by the telecom companies’ existing resources and capabilities. The first role exploits the telecom companies’ resource of customer ownership and billing capability, while the second role exploit the telecom companies’ network infrastructure and its capabilities to deliver Internet service. Additionally, we have showed that these cloud-related roles are transformation from their generic roles in the telecommunication industry, whereas they should already have the basic knowledge and experience

With regards to the Dutch telecom market, we envision that Broadband operators such as KPN, UPC, and Ziggo will have bigger opportunity to adopt the role of SaaS providers due to their network superiority in terms of bandwidth and speed compared to wireless Internet infrastructure. While MNOs such as Vodafone and T-Mobile will have more opportunity in offering PaaS solution for mobile application development, because mobile infrastructure and technology have more attractive functionalities to offer (in the form of APIs) such as SMS and location based functionality. Moreover, both types of Dutch telecom companies would have the same opportunities to discover new service as cloud service brokers. Each of the telecom companies has their own market share in the Dutch telecommunication market. Thus each of them has better understanding over their own customer base, so that they can act as specific cloud service broker for their own market.

Page | 88

6 Conclusions and Recommendations

In this chapter, we will conclude this study by presenting answers for each research questions. Subsequently we will discuss some reflection and the implications of the results to literature and practitioners. Finally, we will present the limitations of this study and our recommendations towards future studies.

6.1 Main Findings

This thesis has two main objectives. First of all, this thesis aims to give guidance on how large enterprise companies in the Netherlands can adopt PaaS technology. Secondly, this thesis also aims to identify whether telecom companies in the Netherlands could create particular new business from the PaaS service provision. The findings of this thesis serve as a basis to answer the main research questions that will be outlined below. Answers to each of the research sub-questions will be presented to lead us to the final answers of our main research questions.

We will begin with our first sub-question:

SQ1: “Who are the key stakeholders in PaaS service provision in the Netherlands? What kinds of providers are offering PaaS service and what kind of organizations are the (potential) PaaS customers?”

In answering our first sub-question, we employed literature research and interviews. As discussed in detail in Chapter 2 and Chapter 4, we discovered that there are five key stakeholders in PaaS service provision in the Netherlands:

1. The Providers. There are some international incumbent IT companies that provide PaaS service in the Netherlands, such as Google, Salesforce.com, and Microsoft. There are also some smaller PaaS providers such as Cordys, Caspio, EngineYard, Heroku, LongJump, OrangeScape, and WorkXpress. Out of all of these providers, it is only Cordys that is a Dtuch- based company. The other providers are US- and India-based companies.

2. The Clients. Based on our interviews with client companies in the Netherlands, we found that there are not so many client companies that are interested to utilize PaaS yet at this moment. We were only able to identify that Company C6 and Company C7 are the ones that are considering to use PaaS solution in the future. However in general, we found that the demand of PaaS solution is not yet established among large enterprises in the Netherlands.

3. The Regulators. Regulator organizations (i.e. EU, Dutch Government, and Dutch National Bank) play a significant role in the adoption and provision of any cloud computing services in the Netherlands, including PaaS offering. In fact, the existence of a regulation that prohibit companies to put their sensitive data outside particular country border holds back the adoption of cloud computing services. For instance, a strict control and regulation from the

Page | 89

Chapter 6

Dutch National Bank is admitted to be a main barrier for companies in financial service industry to adopt cloud computing solutions, as argued by Company C9.

4. The System Integrators. Our study found that the system integrators play an important role in the implementation of IT solutions in larger enterprises, including the implementation of cloud computing solutions. System integrators have the required resources to deal with the technical issues of cloud computing implementation, including in PaaS implementation.

5. The Network Providers. Network providers such as Broadband operators and Mobile Network Operators are also important stakeholders. Our findings explain that these parties are playing a simple but important role in providing and delivering reliable internet connectivity to access any cloud computing services, including PaaS service.

Next, we will elaborate our answer to our second sub-question:

SQ2: “What sort of generic PaaS-related business model from client’s perspective is appropriate to deliver the wanted cloud services in terms of service, organizational and financial domain?”

In answering our second sub-question, we employed literature research to build a conceptual PaaS adoption model for client companies in order to understand how client companies can adopt PaaS and capture the value of PaaS offering. This model is mainly based on Business Model theory, particularly employing the STOF model framework. Our generic model of PaaS adoption is classified in four domains: Service, Technology, Organization, and Finance domains. Each of the domains contains adoption factors. What we mean by adoption factors are the factors that influence client company’s decision in designing its business model to adopt and utilize PaaS technology. The conceptual model is shown and explained in detail in Chapter 3.3.

Furthermore, we will answer the next two sub-questions:

SQ3: “What sort of business software applications do client organizations want to build or migrate by using PaaS to the cloud and how this is related with the generic PaaS-related business model?”

SQ4: “What sort of similarities and differences arise among the identified client PaaS-related business model(s)?

In answering our third and fourth sub-questions, we employed interviews with client companies. Answering sub-question 3 also serves as a mean to validate the conceptual model as proposed in answering sub-question 2. Nevertheless, due to the non-ideal interview results we had as explained in Chapter 4, we have to modify our approach in answering sub-question 3 and sub-question 4. We modify our approach by also taking the perspective of SaaS adoption in our answers, besides taking the perspective of PaaS adoption.

We found that the sort of software applications that client wants to migrate to the cloud are differ from one company to the other, depends on the type of business the company do. Each company has a particular software application that they use to support the company’s business process. However, some similarities of the software applications can be found between the companies within the same industry sectors.

Page | 90

Conclusions and Recommendations

Our study shows that the sort of business software applications clients want to migrate to the cloud (for both PaaS and SaaS) can be classified into several factors. Those factors are the criticality of the application to the business, the sensitivity of the data handled by the application, the value elements for the internal organization that are perceived by migrating the applications to PaaS or SaaS, and the target end-users. The validation shows that the utility pattern of the application is not a crucial factor for both PaaS and SaaS utilization. In addition, SaaS adoption model employs one additional important factor that is the functionality of the application itself. Moreover, PaaS adoption model elicit the criticality of application to be the most dominant factor, while in SaaS adoption model elicit the functionality of application to be the dominant one, even though it still acknowledges the importance of the criticality of the application.

Our finding also shows a tendency that non-critical business applications are most likely to be moved to the cloud, either by means of utilizing PaaS or SaaS. The reason is that those non-critical applications often do not contain critical data information about the company and the company’s customers. In SaaS domain especially, these non-critical applications are closely related to standard applications such as workspace/office applications, communication, and collaboration applications. These standard application offerings are recognized to be mature products that can offer specific functionalities required by the clients.

Furthermore, our findings show the validation of the other factors in Technology, Financial, and Organization domains as discussed in Chapter 4. We also found that regulations plays a significant role in PaaS and SaaS adoption. On top of that, clients also acknowledge the importance of system integrators and network providers as key stakeholders in PaaS and SaaS service provision.

In addition, in relation with the generic conceptual model, our finding shows slightly different relationship between the characteristics of the applications (Service domain) with the needed technical requirements (Technology domain). This difference, however, did not bring significant impact to our findings.

Based on the answers to the research sub-questions elaborated above, we can now answer the first part of our main research question

“What factors are important for enterprise clients to adopt Platform-as-a-Service in order to capture the business value that is offered from Platform-as-a-Service offering?

As the result of this study, we have developed a model that contains the important factors that enterprise clients need to take into account to adopt PaaS and SaaS (please refer to Figure 49 and Figure 50 in Chapter 4.2.4 for the complete model). This model contains several adoption factors in the Service, Technical, Organizational and Financial domains. This model serves as guidance for large enterprise companies to help them in assessing their decision whether to adopt PaaS and SaaS solutions. There are two main stages in using the model. The first stage is to access what requirements (technical and organizational) are needed to move an application to the cloud. These requirements are derived based on the application characteristic factors within the Service domain. Secondly, the model serve as a framework to access the available PaaS and SaaS offering from the cloud providers whether these offerings can fit and fulfill the requirements drawn from the first stage. This framework will help client companies decide whether to adopt the PaaS concept and the

Page | 91

Chapter 6

SaaS concept for particular applications. In addition, clients also need to take into account the regulation factor to be able to implement cloud solutions legally.

Furthermore, our study also answers the following sub-question 5:

SQ5:“What sort of potential role Dutch telecom companies can play in the (potential) PaaS service offerings?”

In answering our sub-question 5, we employed literature research and interviews. We found that there are two roles that Dutch telecom companies can play potentially in the cloud service domain, by taking into account the type of telecom companies in the Dutch market. Those are namely the roles of Cloud service providers and Cloud service brokers. Based on these answer, we can now answer the second part of our main research question:

“What are the strategic implications of the Platform-as-a-Service adoption for Telecom industry?”

Following our identification and analysis of the potential role for telecom companies we argue that not only PaaS adoption, but also SaaS adoption, imply a promising new revenue source for telecom companies. The Internet network infrastructure and broad customer base ownership are the Telecom companies’ strategic resources to expand their business in the cloud computing domain. The exploitation of these resources in the direction of cloud computing, by playing the role of Cloud service providers and Cloud service brokers, would create new business opportunities and makes telecom companies come out of the “bit piper” paradigm.

Final conclusion Our study concludes that clients need to take into account factors in the four domains of Service, Technology, Organization, and Finance in order to capture the value of PaaS and SaaS offering. On top of that, clients need to take into consideration the applied regulations in order to legally implement cloud computing solutions. Clients also need to acknowledge the added value activities offered by system integrators and network providers in implementing cloud computing services. Moreover, our study also concludes that there are two potential roles that telecom companies can play in the cloud computing domain to create new business value. Those roles are Cloud Service Providers and Cloud Service Broker.

6.2 Reflection

We conducted an exploratory study on a new domain of Platform-as-a-Service. We have developed a mental model for PaaS adoption in large enterprise from the client perspective. Within this section we will elaborate reflection of our findings.

Results Representativeness Our findings have shown that the market of PaaS service in the Netherlands still needs to develop, which resulted in small set of samples. This result alter the expected outcome of our study. However, given the relatively small sample size and the richness of our interview data, we are still able to portray a substantial portion of our interview data directly to serves as the first validation of our PaaS adoption model. This approach of using small set of interview data to validate a model has been

Page | 92

Conclusions and Recommendations adopted in the validation of system dynamics modeling (Diker, 2003; Luna-Reyes & Andersen, 2003; Rich, 2002; Richardson & Pugh, 1981). Diker (2003), Luna-Reyes (2003), Rich (2002), and Richardson & Pugh (1981) made intensive use of qualitative data during their whole modeling process in order to increase their general understanding of a phenomenon. In this way, such a small set but rich qualitative data can offer some reflections about the validity of a model as general theories (Luna- Reyes & Andersen, 2003). As examples, Rich (2002) used only 11 respondent to test his model of sustainability of knowledge management projects, while Luna-reyes (2003) only used four respondents to validate his model of collaboration in international IT projects. However, we are totally aware that more extensive validation of the model still needs to be done. We will elaborate more on these issues within the recommendation section.

Industry Sectors Due to the unequal spread of our interview samples across different industries, we are not able to discuss the difference of PaaS adoption among different industries. However, because we had four financial companies in our samples, we will discuss more insight from this particular industry. We discussed in our result analysis that one of the main barriers in cloud adoption is the regulation. The Dutch financial service companies, in particular, face more stringent regulations with regards to cloud computing implementation in general compared to the other companies. While other industries mainly concern with general regulation that are derived from EU directives, on top of that the Dutch financial service companies have to obey local regulation from the Dutch National Bank. This has become one of their main concerns for cloud computing adoption and explicitly reflected by our conversation with Company C8. “The problem with banks is that we are very closely controlled by the Dutch National Bank. We have to follow very strict rules from the National Bank for having control over all the activities we are doing, and also the way we put our data and applications somewhere, especially if it is not in the Netherlands”, said Company C8.

We also learned that there are some similarities among the financial industry companies when it comes to their goal of utilizing PaaS or SaaS solutions. We learn that their IT strategy is mainly concerning about operational excellence, which means that they are aiming to reduce cost of their IT operations. Based on the interviews, we understand that this aim is achieved by means of standardizing their IT applications, where any customizations of the applications are really minimized in order to reduce cost incurred from IT implementations and operations. In one hand this can explain why public SaaS offering is more attractive than public PaaS offering in the financial service industry, because financial service companies do not want to develop their own SaaS. SaaS solution is then very appealing for them, but they still see that not many SaaS applications can fit to all the functionalities they are looking for. That is why the SaaS adoption is restricted to non-critical and standards applications only (i.e. office or workspace applications). On the other hand, it can be seen that cloud computing is considered to be the enabler of their IT strategy. It means that in the cloud computing domain in the financial service industry, they consider their IT strategy to come first before their cloud computing business model. This is in contradiction with our discussion about business model and strategy in Chapter 4.2.1 that told us otherwise. This case reinforce Casadesus- Masanell and Ricart (2009) argument that a firm’s business model is “a reflection of its realized strategy”.

Page | 93

Chapter 6

Regulation As discussed before, this study has shown that there are many concerns related to regulation that clients perceive in order to adopt cloud computing service in general. This concern impacts clients’ strategy by holding back implementation of critical applications in the cloud environment.

This issue about regulation, triggers us to investigate more deeply about what regulation exactly it is that hinder the clients in general to adopt cloud computing. From our conversation with the clients, we understood that there are two main regulations that are of their concern. The first regulation is the EU Privacy Directive 1995/46/EC. According to the interpretation of this directive, personal information may be transferred to countries within the European Economic Area (EEA) and Canada, Argentina, Switzerland, Israel and Andorra, because these countries’ legal systems offer adequate protection. Transfers to the US are possible (under certain conditions) if the receiving entity is Safe Harbor-certified. Transfers to any country are possible if the organization has implemented corporate binding rules (Casper, 2011). The second regulation is the USA Patriot Act that gives U.S. authorities unlimited access to personal data, hence making a European company more reluctant to engage with a US cloud service provider (Casper, 2011). Based on our interpretation of the regulation, data transfers outside the country border is not impossible from the legal perspective. However, there are a lot of compliances to be met in order to make the move to the cloud is legal, while these compliances might be not well known or too cumbersome for companies.

Due to language barriers and limited publication in English, researchers was not able to find out how the EU Privacy Directive is interpreted in the Dutch law. Further investigation need to be conducted to address more relevant reflection in the Dutch cloud market. However, in general, we argue that regulators must start to look closely on regulations in cloud computing specifically. Regulators have to make clear the implications of the existing law with regard to cloud computing adoption by enterprise companies. The EU Privacy Directive was made in 1995, where cloud computing technology might not be out of concern back then. Regulation might need to be changed not to only focus on data location issue, but also to focus on increasing the level of security in the cloud environment, thus more sensitive information can be put in the cloud. One approach can be done by setting up a Code of Practice like in the UK (Burton, 2011). The Code of Practice enforce cloud service providers to demonstrate accountability, transparency and capability to end users to be self-certified. The cloud providers must adhere to their SLAs. Therefore allowing the potential customer to make an informed decision about the most suitable service for them. As a result, end users can be educated and encouraged to invest in reputable cloud service providers, to encourage a wide-spread adoption of cloud (Burton, 2011). From the cloud industry side, both of the provider and the client side need to engage with local regulators more closely in order to develop the cloud computing market (whether PaaS, SaaS, or even IaaS). The importance of regulators needs to be acknowledged and handled in stakeholder management when a company wants to implement cloud computing solution. In conclusion, specific regulations concerning cloud computing should be made in order to allow adoption of such a state-of-the-art technology but still keeping the security and privacy level high.

Security and Reliability Our research findings show that technical requirements such as security (code: (T) Security Requirement (PaaS/SaaS) and QoS (code: (T) QoS requirement (PaaS/SaaS)), in terms of reliability

Page | 94

Conclusions and Recommendations and availability, are the most important technical requirements in adopting PaaS and SaaS. Therefore cloud providers have to maintain their security and reliability of their service as number one priority. Even though clients consider the factor of branding to be important in their selection of cloud provider, a single outage on security and reliability issues can really damage the cloud provider’s reputation and branding (Leavitt, 2009).

Moreover, as discussed before, security issues might need to be more strictly regulated to protect customers and also to boost the cloud market. Regulations should not only restricted and focus on data location issues. Additionally, as our result shows, telecom companies can play added value activities to ensure more reliability of cloud service by exploiting their Internet network infrastructure resource.

Reflections to Theories and Related Work We have several reflection towards the theory. Firstly, we have shown that Business Model theory especially the STOF model is a fine framework to analyze client’s adoption of cloud computing services. The STOF model has served as relevant basis to analyze the cloud computing environment. Most of the STOF’s critical design issues (CDI) in the four domains of Service, Technology, Organization, and Finance are applicable to the concept of cloud computing adoption. However, it is important to note that its applicability to cloud computing is limited. Some adaptations and adjustments on specific issues of cloud computing, especially in the Service and Organization domains, need to be added in order to make the STOF model fits to the cloud computing domains. These adjustments mainly related to the specific characteristics of the cloud computing itself, such as the characteristics of the applications (i.e. its utility pattern, criticality to the business, and data sensitivity).

With regards to the Financial domain, we found that the general views of our respondents about cloud computing investments are the same with the other IT investments. It means that cloud computing solution must deliver competitive and attractive business case financially, compared to the other IT solutions, in order to be selected. Given the rapid development of cloud computing forecasted in the future (IDC, 2010a, IDC, 2010 #29), we speculate that client companies should adapt other approach regarding cloud computing investment. We think that Real Option theory might be relevant, because it gives the client companies flexibility on their investment on cloud computing as the market develop. Client companies can decide on the later stage whether cloud computing solutions is keeping up their promise of their technology superiority, and hence decide whether to continue and expand the investment on cloud computing based on that.

Secondly, we want to reflect our findings with the related work in cloud computing. Reynolds and Bess (Reynolds & Bess, 2009) argue that there are two possible reasons to utilize cloud applications: to reduce cost and to add value. It is clear from our findings that reducing cost is one of the main goals for enterprise companies to utilize either PaaS or SaaS. Moreover, our findings also able to specify what added values (other than cost-related) that client companies perceive in utilizing cloud computing. From our discussion in Chapter 5, we specify that there are nine different added values that clients perceive in utilizing PaaS, and seven different added values that clients perceive in utilizing SaaS. An important point to make is that clients really see that PaaS and SaaS will bring them the benefit of faster time to market and also improve their service quality.

Page | 95

Chapter 6

Thirdly, we want to reflect our findings to the other related work in cloud computing. Misra and Mondal (Misra & Mondal, 2011) defines several characteristics of organization that want to use cloud computing. Our finding confirms part of this work. We found that the size of customer base (factor of Target end-users) is a relevant characteristic in PaaS and SaaS adoption factor. However, we found that the other characteristics such as number of servers, number of countries IT is spread across, and the annual revenue from IT offerings is not relevant in PaaS and SaaS adoption in the Netherlands. We never encounter these issues during our interviews. One possible reason is that the client companies we interviewed did not see these issues as important factors when it comes to PaaS or SaaS adoption. Moreover, Misra and Mondal (Misra & Mondal, 2011) also define some characteristics of applications to be migrated to the cloud. Our findings confirm most of the characteristics they defined. Our findings show that criticality of the application and sensitivity of data is indeed important issues for clients. However, we found that utilization pattern of the application is not a critical issue. One possible reason is that the other two characteristics/factors over-weigh the utilization pattern issue. Because most of the applications to be migrated are the non-critical applications, where most of them have flat utilization pattern over the years.

Fourthly, the results of this study also confirm and enrich the existing literature on cloud computing key stakeholders. In particular, we contribute to the work of Marston et al. (2011) by adding one more key stakeholders in their stakeholder analysis. Marston et al. (2011) defines four key stakeholders in cloud computing world, there are Providers, Clients, System Integrators, and Regulators. Our findings confirm the importance of these four key stakeholders in cloud computing ecosystem. Additionally, we argue that Network Provider (i.e. telecom companies) is also an important key stakeholder in the provision of cloud computing services such as PaaS and SaaS.

Morevoer, the characteristic of cloud computing, particularly SaaS offering has resemble similarities with outsourcing concept. However, we argue that cloud computing differs from traditional outsourcing. In traditional outsourcing, it is still very much standalone computing. In many cases a client knows where exactly his data/host is and what resources are shared with others (Wang, 2009). While in cloud computing, data is decoupled from infrastructure and obscures low-level operational details. Multitenancy is the main differentiator between traditional IT outsourcing and cloud computing. These differences give rise to unique set of security and privacy issues that not only impact client’s risk management practices, but also stimulated a fresh evaluation of legal issues in areas such as compliances and auditing. This concern is really shown in our research findings regarding the main barriers. The main barriers to adopt cloud computing are regulation, security and privacy issues.

Additionally, our result shows that PaaS vendors acknowledge the importance of ISVs in PaaS environment (section 4.2.2). PaaS vendors see that ISVs play important role in accelerating the adoption of PaaS through their involvement in the PaaS marketplace. In contrast, we found that ISVs are not seen to play such important role from the client’s perspective. One explanation to this result is because most of the clients are not interested in PaaS in the first place, so they oversee the role of ISVs. However, from the vendor’ perspective, multi-sided platform theory (Gawer & Cusumano, 2002b) might be relevant to further expand their PaaS business model in relation with the ISVs involvement. It is interesting to take into account how the openness of the PaaS platform might attract ISVs in order to accelerate the adoption of the PaaS service. It is also interesting to see whether PaaS service could benefit from network externality effect (Katz & Shapiro, 1986) to

Page | 96

Conclusions and Recommendations outperform competitors that provide similar PaaS services. Multi-sided platform theory can be used to create vendors’ strategy to gain competitive advantage among other PaaS providers. Nevertheless, this theory is not covered deeply in this study because this theory takes more of vendor’s perspective to be applicable than the client’s perspective.

Telecom Companies’ Roles in the Cloud With regards to the telecom companies’ role in cloud computing, it needs to take into consideration some critical issues. In relation with the role of cloud service broker, telecom companies have to be careful in choosing which cloud providers they syndicate. Telecom companies have to apply stringent qualification toward the cloud providers, so that the cloud services they resell do not fall short under security and regulation issues. Otherwise, the telecom companies’ brand and reputation will suffer. Together with the trust they receive from the client, telecom companies should act as an entity that helps clients to check whether the delivered cloud services comply with the regulation and able to deliver high standards of security and reliability.

6.3 Contribution to Literature

This study adds several contribution to the literature. Firstly, our study has contributed to the cloud computing literature. Our study has highlighted one specific topic of Platform-as-a-Service to be one important domain in ICT related studies. The results of this study enrich the existing literature on cloud computing. We contribute to the work of Marston et al. (2011) by adding one more key stakeholders in their stakeholder analysis. The additional key stakeholder is the Network Provider. With regards to the related work in cloud computing of Misra and Mondal (2011), our study confirms part of their findings that criticality of application and sensitivity of application data are important factors of applications to be migrated to the cloud. However, we found that the other factor of application utilization pattern is falling short. Our study has also contributed to the work of Reynolds and Bess (2009) by confirming that reducing cost is indeed a possible reason to utilize cloud services, and also by specifying the value added of utilizing cloud services.

Secondly, we also enrich the business model theory, particularly the STOF model (Bouwman et al., 2008). STOF model was developed particularly for mobile service design. Our study extend the STOF model by applying it into the domain of cloud computing. STOF model is applicable for the basis of our study, however, some adaption need to be performed based on related works on cloud computing to fit to the need of our research.

Thirdly, our study has taken the business model theory within the internet-based service domain into a different perspective. Commonly, business model theory is used from the perspective of service provider that owns the technology that is offered. Our study takes the perspective of the client of a technology service. We analyze how clients can make use of an offered technology service to create new service towards their own organization or end-customers by using the business model perspective. Our study has given a different dimension into business model theory.

To conclude, our study has made a first attempt to construct a unified model for cloud computing adoption that makes explicit the important factors in the service, technology, organization and financial domain. Our study contributes to the theory of adoption of PaaS in particular, which to the best of our knowledge is still lacking in the literature. Additionally, we present some insights of the

Page | 97

Chapter 6 differences between PaaS and SaaS adoptions that show the distinguishing factors between the adoptions of these two cloud-based services.

6.4 Recommendation for Practitioners

Our study on the market condition and the adoption factor of PaaS lead to several implications for managers. First of all, we derive implications for the IT managers at the client side. Although the study findings are still far from generalization, the validated model presented in Chapter 4 served as guidance for the managers at the client side in considering the adoption of PaaS in their IT strategy. There are several factors in different domains that are interdependent to each other (Service, Technology, Organization and Finance domains). These factors in different domains are important to take into account when considering the utilization of the technology.

The validated adoption model showed in Chapter 4 also serves as a framework for decision making process for enterprise clients in order to decide whether to implement PaaS or SaaS solutions. As explained in previous sections, the first stage of the model (Service domain) serve as a framework to assess whether an application is suitable to be moved to the cloud environment, and derive what technological requirements are needed (Technology domain). The second stage of the model serves as a framework to assess and decide which cloud vendors/providers are suitable to deliver the wanted application into the cloud.

Moreover, the market analysis shows that most of the companies in the Netherlands feel that the cloud service is not yet mature enough, and it is a bit too risky to be the first players in this field, especially with PaaS adoption. However, based on the discussion with Company V2, we argue that it is important for enterprise companies (as clients) to start thinking and planning about public cloud implementation, because the future of public cloud computing do not lie only for the purpose of migrating legacy system, but also to developed new applications that are really developed for the cloud environment. Many studies believe that cloud computing is the future of IT. As argued by (Narasimhan & Nichols, 2011), technical concern such as security is the number one of misconception about cloud applications and platforms. Thus, it is important to start doing experiment with small cloud service to gain the experience and prepare the organization towards bigger implementation of cloud in the future. Based on our findings, client companies could do this by start doing SaaS implementation for their non-critical standard application.

Secondly, we derive implications for the Managers at the cloud provider companies’ side. From our analysis of the market potential of PaaS, we understood that a lot of companies are still hesitant to go to the cloud in general, even though our results are limited to the geographical area of the Netherlands. Based on our findings presented in Chapter 4.2.2, we can observe that not all of the values that were proposed by the PaaS vendors are well perceived as important values by the client organization. Thus, as supported by the conversation with Company V1 and V2, market education of PaaS in particular, and cloud service in general still needs to be carried on more intensively to boost the adoption of these cloud services, particularly PaaS service.

Thirdly, for the telecommunication companies, manager may consider to take the potential roles as identified in our study to the next level to see the feasibility of these roles. Developing and playing one of these roles may create new business value for telecom companies, which in turn will create

Page | 98

Conclusions and Recommendations new revenue source. Thus our study could serve as the first foundation of telecom companies’ strategy making in the cloud computing area.

6.5 Limitations and Further Research

Within this section we will elaborate several limitations of our research that can be improved. Accordingly, we will present some recommendation for the future research.

6.5.1 Research Limitations This study has several limitations. Firstly, despite many efforts, our study is based on limited number of interview respondents. We only had the opportunity to investigate PaaS and SaaS adoption from 9 client companies due to the limited number of companies that were possible to contact and were willing to participate on the research. However, given the relatively small sample size and the richness of our interview data, we are still able to portray a substantial portion of our interview data directly to serves as the first validation of our PaaS adoption model. This approach of using small set of interview data to validate a model has been adopted in the validation of system dynamics modeling (Diker, 2003; Luna-Reyes & Andersen, 2003; Rich, 2002; Richardson & Pugh, 1981). They made intensive use of qualitative data during their whole modeling process in order to increase their general understanding of a phenomenon. In this way, such a small set but rich qualitative data can offer some reflections about the validity of a model as general theories (Luna-Reyes & Andersen, 2003). However, we are totally aware that more extensive validation of the model still needs to be done. We will elaborate our suggestion on how to further validate our findings in the following section.

Secondly, our samples were not fairly distributed across different industry sectors. If our sample were more representing different industry sector equally, a cross analysis among the industry sectors cloud have been done and might give deeper insight on the issues at stake.

Thirdly, our study is mainly based on Dutch companies. If the scopes of geographical area of the companies respondents are broaden to include other companies in Europe, the results might have been more valid to be generalized in the European level. Moreover, as our findings show that one of the major barriers in the adoption of PaaS and SaaS is related to government regulations, applicability of this study in different region in the world needs to be further investigated due to different regulation that might apply in different region in the world.

Additionally, we did not able to empirically validate the result of our analysis about the role of telecom companies in cloud computing domain in the Netherlands. The reason is due to the time limitations we had that to conduct this study that did not allow us to conduct another round of interview sessions with the experts from telecommunication industries for verification of the analysis.

6.5.2 Direction for Further Research To address the aforementioned limitations, future research cloud be focused on the following issues. First of all, further research could focus on extensive validation of the model that we developed. A methodology of quantitative survey research can be adopted to conduct the research. The sampling strategy might include large number of samples of enterprise companies throughout European

Page | 99

Chapter 6

Union. The samples must also be fairly distributed among different industry sectors to enable comparison analysis among the different sectors. The targeted respondents might include IT managers that have involvement in the company’s IT strategy making. IT architects might also be included, since often they have more technical insights on IT-related technology developments such as cloud computing.

Furthermore, the questionnaire of the surveys can be designed based on the adoption factors that are included in our adoption model. Due to the newness of cloud computing technology, the questionnaire can start with some questions to test the respondents’ understanding on the concept of “cloud computing”. This way, an insight can be gained whether the samples have the same and correct understanding of cloud computing technology (Gartner, 2010). Accordingly, the questionnaire can represent different factors in the different domains as shown in our model. First, a question to investigate the reasoning to adopt PaaS can be asked. Next, questions regarding the factors of the applications that want to be migrated (refers to the Service domain) can be asked. The questions of the required technical issues (refers to the Technical domain) can follow. Finally questions regarding how would respondent choose their providers could be addressed (refers to the Organization domain), which include the financial aspect as well (refers to the Finance domain). As the analysis techniques, statistical analysis might be analyze the quantitative data in order to validate the model.

However, there is one important thing to note. As our study shows that the PaaS market is still immature, it would be wise to conduct this further research after some years when the PaaS market is more developed. By this means, the data collected could represent the client companies experience in adopting PaaS, thus providing a more relevant validation of PaaS adoption.

Secondly, this study was conducted at the level of large enterprises. Hence, our findings focus mainly on how large enterprise could capture some advantageous values from the PaaS and SaaS offering. As suggested by Marston et al. (2011), startup companies and SMEs might have more eminent benefit in using cloud computing technology. Therefore, further research could focus on comparing the possible differences in the values that are perceived by larger enterprises and SMEs with regard to the cloud computing adoption, particularly PaaS adoption.

Thirdly, our result shows that cloud computing market in the Netherlands is still immature. Instead of using the business model theory in studying the cloud market in the Netherlands, further research could focus on investigating the strategy roadmap of enterprise companies in order to adopt PaaS or cloud computing in general. Despite a lot of hesitancy at this moment to utilize cloud computing, there are a lot of optimism in the future towards full utilization of cloud computing technology in the enterprise IT landscape. This optimism at least was shown from the conversation with Company C6, Company C8 and Company C9. Therefore a research that focuses on roadmap approach would be interesting in order to have deeper knowledge on the transition steps. What we mean by transition steps are the important steps that enterprise companies should take to move from the current traditional IT strategy towards the future all-around cloud-based IT strategy.

Fourthly, our study has given an indication about the importance of ISVs as third party actor in PaaS service provision from the vendor perspective. ISVs are seen as the accelerator of PaaS adoption. Therefore, further research could focus on investigating the application of multi-sided platform theory (Gawer & Cusumano, 2002a) towards PaaS business model from the vendor perspective. A

Page | 100

Conclusions and Recommendations relevant research question might be whether the multi-sided platform theory is application in cloud computing domain, particularly in the PaaS context.

Finally, the findings on the telecom companies’ role are tentative, because it has not been empirically validated. Further research could focus on empirically validating our findings by means of elaborating the business model for each of the roles and testing them towards the industry experts. Elaborating the business model might give deeper understanding of the feasibility of each role , and validating them through set of interviews with industry experts might give practical insight and feasibility of each roles. Qualitative research approach can be adopted for this further research.

Page | 101

Chapter 6

This page intentionally left blank

Page | 102

References

References

Accenture. (2009). Microsoft Azure Solutions from Avanade and Accenture.

Accenture. (2010). Wireless VoIP: Threat or Opportunity.

Amazon Web Services. (2011a). Amazon Elastic Compute Cloud (Amazon EC2). Retrieved February 22, 2011, from http://aws.amazon.com/ec2/

Amazon Web Services. (2011b). Amazon Simple Storage Services (Amazon S3). Retrieved February 22, 2011, from http://aws.amazon.com/s3/

Amit, R., & Zott, C. (2001). Value creation in e-business. Strategic Management Journal, 22(6-7), 493- 520.

Apfelbeck, J. (1998). A role model for telecommunication networks, Global Telecommunications Conference (Vol. 2, pp. 1212 - 1217): IEEE.

Ballon, P. (2009). Control and Value in Mobile Communications: A Political Economy of the Reconfiguration of Business Models in the European Movile Industry. SSRN.

Ballon, P., & Walravens, N. (2008). Competing platform models for mobile service delivery: The importance of gatekeeper roles.

Barley, S. R. (1990). The alignment of technology and structure through roles and networks. Administrative science quarterly, 35(1), 61-103.

Barney, J. (1991). Firn Resources and Sustained Competitive Advantage. Journal of Management, 17(1), 99-120.

Bharosa, N. (2011). Netcentric Information Orchestration: Assuring information and system in public safety networks. Delft University of Technology, Delft.

Bouwman, H., de Vos, H., & Haaker, T. (2008). Mobile Service Innovation and Business Models. Berlin- Heidelberg: Springer.

Burton, A. (2011). UK cloud regulation services are a step ahead of the Australian government. Retrieved January 14, 2011, from http://www.businesscomputingworld.co.uk/uk-cloud- regulation-services-are-a-step-ahead-of-the-australian-government/

Buvat, J., & Nandan, P. (2010). Cloud Computing: The Telco Opportunity: Capgemini.

Buyya, R., Yeo, C. S., Venugopal, S., Broberg, J., & Brandic, I. (2009). Cloud computing and emerging IT platforms: Vision, hype, and reality for delivering computing as the 5th utility. Future Generation Computer Systems, 25(6), 599-616.

Casadesus-Masanell, R., & Ricart, J. E. (2009). From Strategy to Business Models and onto Tactics. Long Range Planning, 43(2-3), 195-215.

Casper, C. (2011). Privacy in the Cloud: Gartner.

Page | 103

Chappell, D. (2010). The Windows Azure Platform and ISVs: A Guide for Decision Makers (White Paper). San Fransisco: Chappell & Associates.

Chesbrough, H., & Rosenbloom, R. S. (2002). The role of the business model in capturing value from innovation: Evidence from Xerox Corporation's technology spin-off companies. Industrial and Corporate Change, 11(3), 529-555.

CSA. (2009). Security Guidance for Critical Areas of Focus in Cloud Computing V2.1: Cloud Security Alliance.

De Reuver, M., & Bouwman, H. (2008). Resource dependencies in mobile services value networks.

De Reuver, M., Visser, A., & Bouwman, H. (2011). The emerging ecosystem for future mobile communication services, 18th Biennial Conference of the International Telecommunications. Tokyo.

Diker, V. G. (2003). Toward a dynamic feedback theory of open online collaboration communities. . University at Albany, SUNY, Albany, NY.

Dogra, R., & Orr, S. (2009). Managing Through Challenging Times: How communications companies in Europe and Latin America should respond to today's challenges and tomorrow's opportunities: Accenture.

Forge, S. (2004). Is fourth generation mobile nirvana or ... nothing? info, 6(1), 12-23.

Forrester. (2009). Cloud Computing for the Enterprise.

Gartner. (2010). Survey Analysis: European Adoption of Cloud Computing, Patterns and Preferences.

Gawer, A., & Cusumano, M. A. (2002a). The elements of platform leadership. MIT Sloan Management Review, 43(3), 51-58.

Gawer, A., & Cusumano, M. A. (2002b). Platform Leadership: How Intel, Microsoft, and Cisco Drive Industry Innovation. Boston: Harvard Business School Press.

Gens, F. (2009). Defining "Cloud Services" - an IDC update, IDC Exchange (Vol. 2011).

Gonçalves, V., & Ballon, P. (2011). Adding value to the network: Mobile operators' experiments with Software-as-a-Service and Platform-as-a-Service models. Telematics and Informatics, 28(1), 12-21.

Google. (2011). Google App Engine. Retrieved February 22, 2011, from http://code.google.com/appengine/

Hirzalla, M. (2010). Realizing business agility requirements through SOA and cloud computing.

IDC. (2009). Telcos in the Cloud: Claiming a Seat at the Table. Framingham, MA.

IDC. (2010a). Western European Cloud Services 2010-2014 Forecast and Analysis. London.

IDC. (2010b). Worldwide and Regional Public IT CLoud Services 2010-2014 Forecast. Framingham, MA.

Page | 104

References

InformationAge. (2011). Will BT's telco roots help it win in the cloud? , from http://www.information-age.com/channels/the-cloud-and-virtualization/company- analysis/1641698/will-bts-telco-roots-help-it-win-in-the-cloud.thtml

Kambil, A., & Short, J. E. (1993). Electronic integration and business network redesign: A roles-linkage perspective. Journal of Management Information Systems, 10(3), 59-83.

Katz, M. L., & Shapiro, C. (1986). Technology Adoption in the Presence of Network Externalities. The Journal of Political Economy, 94(4), 822-841.

Kuo, Y.-F., & Yu, C.-W. (2006). 3G telecommunication operators' challenges and roles: A perspective of mobile commerce value chain. Technovation, 26(12), 1347-1356.

Leavitt, N. (2009). Is Cloud Computing Really Ready for Prime Time? Computer, 42(1), 15-20.

Lin, G., Fu, D., Zhu, J., & Dasmalchi, G. (2009). Cloud computing: IT as a service. IT Professional, 11(2), 10-13.

Luna-Reyes, L. F., & Andersen, D. L. (2003). Collecting and analyzing qualitative data for system dynamics: methods and models. System Dynamics Review, 19(4), 271-296.

Marston, S., Li, Z., Bandyopadhyay, S., Zhang, J., & Ghalsasi, A. (2011). Cloud computing -- The business perspective. Decision Support Systems, 51(1), 176-189.

Meyer, D. (2008). BT starts reselling Salesforce CRM. Retrieved June 30, 2011, from http://www.zdnet.co.uk/news/it-strategy/2008/10/03/bt-starts-reselling-salesforce-crm- 39497768/

Microsoft. (2011). Windows Azure. Retrieved February 22, 2011, from http://www.microsoft.com/windowsazure/

Miles, M. B., & Huberman, M. A. (1994). Qualitative Data Analysis: An Expanded Sourcebook (Second ed.). Thousand Oaks, California: Sage Publications.

Misra, S. C., & Mondal, A. (2011). Identification of a company's suitability for the adoption of cloud computing and modelling its corresponding Return on Investment. Mathematical and Computer Modelling, 53(3-4), 504-521.

Narasimhan, B., & Nichols, R. (2011). State of cloud applications and platforms: The cloud adopters' view. Computer, 44(3), 24-28.

Natoli, J. (2008). SOA and SaaS ... What's the Difference?, Intel Software Network (Vol. 2011).

O2. (2009). O2 Litmus. Retrieved February 22, 2011, from http://www.o2litmus.co.uk/

OPTA. (2011a). Large number going digital in the Netherlands: OPTA.

OPTA. (2011b). Market Monitor 2010: OPTA.

OPTA. (2011c). New figure for communication markets, second half of 2010: OPTA.

Orange. (2011). Orange Partner. Retrieved March 17, 2011, from http://www.orangepartner.com/site/enuk/home/p_home.jsp

Page | 105

Osterwalder, A., & Pigneur, Y. (2003). Modeling Value Propositions in E-Business, Pittsburgh, PA.

Porter, M. E. (1985). Competitive Advantage: Creating and Sustaining Superior Performance. New York: Free Press.

Rake-Revelant, J., Holschke, O., Offermann, P., & Bub, U. (2010). Platform-as-a-Service for business customers: A path for future revenue streams beyond the bitpipe.

Reynolds, E., & Bess, C. (2009). Clearing up the Cloud: Adoption Strategies for Cloud Computing. Cutter IT Journal.

Rich, E. (2002). Modeling the dynamics of organizational knowledge. University at Albany, SUNY, Albany, NY.

Richardson, G. P., & Pugh, A. L. (1981). Introduction to System Dynamics Modeling with Dynamo. Cambridge, MA: MIT Press.

Salesforce.com. (2008). BT Drives Tailored Approach to SME Customer Management. Press Releases Retrieved June 30, 2011, from http://www.salesforce.com/uk/company/news-press/press- releases/2008/10/081002.jsp

Salesforce.com. (2010). Salesforce.com Signs Agreement with NTT Communications to Establish Data Center in Tokyo.

Scarpati, J. (2011). Telecom cloud services must differentiate with security, provisioning. from http://searchtelecom.techtarget.com/news/2240037429/Telecom-cloud-services-must- differentiate-with-security-provisioning

Seddon, P. B., & P., L. G. (2003). Strategy and Business Models: What's the Difference? Paper presented at the Pacific Asia Conference on Information Systems.

Shafer, S. M., Smith, H. J., & Linder, J. C. (2005). The power of business models. Business Horizons, 48(3), 199-207.

Stabell, C. B., & Fjeldstad, Ø. D. (1998). Configuring value for competitive advantage: On chains, shops, and networks. Strategic Management Journal, 19(5), 413-437.

Subashini, S., & Kavitha, V. (2011). A survey on security issues in service delivery models of cloud computing. Journal of Network and Computer Applications, 34(1), 1-11.

Sultan, N. A. (2010). Reaching for the "cloud": How SMEs can manage. International Journal of Information Management, In Press, Corrected Proof.

T-Mobile. (2010). Mobile broadband future [Mobile service innovation course presentation]. Delft.

Tesch, R. (1989). Computer software and qualitative analysis: A reassessment. New Brunswick, NJ: Transaction Books.

Verschuren, P., & Doorewaard, H. (2005). Designing a Research Project. Utrecht: Lemma.

Visscher, R. (2011). Cloud computing is a hackers dream. TU Delta Retrieved July 24, 2011, from http://www.delta.tudelft.nl/nl/archief/artikel/opinion-please-cloud-computing-is-a-hackers- dream/23447

Page | 106

References

Vodafone. (2010). Betavine. Retrieved March 17, 2011, from http://www.betavine.net/home/main/home.html

Wang, C. (2009). Cloud Security Front and Center, Forrester Blogs (Vol. 2011): Forrester.

Watson, R. (2011). Google App Engine, Windows Azure, and Force.com: What's your PaaS Style? MIdvale, Utah: Burton Group.

Weaver, P. (2011). Multinationals trust telcos with the cloud (Vol. 2011): telecoms.com.

Weill, P., & Vitale, M. R. (2001). Place to Space: Migrating to e-business Models. Boston: Harvard Business School Press.

Weinhardt, C., Anandasivam, A., Blau, B., & Stöer, J. (2009). Business models in the service world. IT Professional, 11(2), 28-33.

White, E. (2011). The Role for Telcos in Cloud Computing, Cloud Expo: Press Release: Cloud Computing Journal.

Wilson, C. (2011). Microsoft Calls on Telcos for Office 365. News Analysis, from http://www.lightreading.com/document.asp?doc_id=209443&f_src=lightreading_gnews

Xignite. (2011). Xignite - On Demand Financial Market Data. Retrieved February 22, 2011, from http://www.xignite.com/

Page | 107

This page intentionally left blank

Page | 108

Appendix

APPENDIX A - Telco Market Overview in the Cloud

Current Telco activities in cloud computing as cloud providers1.

Telcos IaaS SaaS PaaS Belgacom Computing, backup Enable! MS CRM, MS Exchange, N.a. Videoconferencing, MS Sharepoint Computing, storage (in E-mail archiving, Unified N.a. Switzerland) Communications and Collaborations (UCC) services and video communication AT&T Synaptic: computing and Enablement infrastructure that N.a. storage as a service platform allows ISVs to bring their software to market via an on- demand, cloud based delivery model Telecom Italia Ospit@Virtuale: Dyanmic Infomobility, Document Virtual Hosting Platform Management and ERP inclusion providing full IaaS stack, in hospitals combined with VPN to provide hybrid cloud offering TeliaSonera N.a. Initially offer cloud-based N.a. conferencing and services to integrate video, telephony, chat and file sharing, and plan to expand the cooperation with more offerings further on British Telecom Virtual Data Center Services On demand application N.a. including server, storage, offerings such as CRM, HR, security and networking accounting, eSignature, services communication & collaboration (OneVoice) Vodafone N.a. N.a. Betavine: mobile application development platform Verizone Computing-as-a-Service, Secure Communication and N.a. Cloud Storage targeting collaboration solutions focusing complete Enterprise sector on SMBs Orange Flexible Computing: virtual IT Selected business app including Orange Partner: infrastructure, Fastbeat 360: Desktop as a Service, Microsoft mobile application Content Delivery Cloud service BPOS, Announced Enterprise development platform for up/downscaling and App Store distribution of digital content Deutsche Telekom/ Zimory Dynamic Services for ERP Database and T-Systems applications (SAP), CRM, middleware Communication and environment collaboration and archiving and e-commerce

1 Source: Forrester “Market Overview of Cloud IT Service from Major Telcos”, 2009

Page | 109

APPENDIX B - Telecom Market in the Netherlands

Telecommunication market in the Netherlands is grouped into mobile telephony, fixed telephony, broadband, and television (OPTA, 2011c). Within our discussion we will only consider mobile telephony and broadband markets, because these two markets are the ones that provide Internet services. Internet services are the backbone of any cloud computing services. Players in mobile telephony market as what we call Mobile Network Operators (MNOs), provide internet through their mobile networks. While players in broadband market as what we call Broadband operators , provide internet connection through fixed line.

Fixed line broadband internet connections are delivered mainly through three main networks, they are: Cable, DSL, and Fibre Optics. In the Netherlands DSL was the most popular network with 3.6 million connections, while cable was number two with 2.5 million connections. (OPTA, 2011b). While the wireless Internet connectivity is mainly delivered through 3G network technology at this moment. In the future, MNOs are expected to deliver faster wireless Internet connectivity through 4G network technology such as LTE (T-Mobile, 2010).

Moreover, Figure 51 below show the overview and the market share of MNOs and Broadband operators in the Netherlands respectively. Figure 51 (left) shows that KPN is the MNO that has the largest market share in the Netherlands, followed by T-Mobile and Vodafone. Despite these three main players, part of the market is served by a so called Mobile Virtual Network Operators (MVNOs). These MVNOs use the excess mobile network capacity of the MNOs to deliver their service. Thus MVNOs do not have their own mobile network infrastructure.

Figure 51 Dutch mobile telephony market (left) and broadband market (right) as of 2010 Q3 (OPTA, 2011b)

Figure 51 (b) shows that KPN is the only telecom company in the Netherlands that serves both of mobile telephony market and broadband market. KPN still holds the largest market share in the provision of broadband service. KPN’s main competitor in this market is Ziggo and UPC that come second and third respectively in their share of the Dutch market.

In our discussion, to limit our scope we only consider the market leaders of mobile telephony market and broadband market in the Netherlands, namely KPN, T-Mobile, Vodafone, UPC and Ziggo.

Page | 110

Appendix

APPENDIX C – Interview Protocol

INTERVIEW PROTOCOL – CLIENTS

Name of the organization : Industry sector : Respondent’s name : Position/role : Date and time :

PHASE 0 - INTRODUCTION

Good morning/afternoon Mr/Ms … . My name is Aryo Primagati. Before I start, I would like to thank you for your participation in this research. This interview will last approximately one hour.

The general idea of this interview is to get insight on how large enterprises (as clients) perceive value that is offered from PaaS offerings, what kind of business model and business applications that your company (want to) have by utilizing the cloud technology (particularly Platform-as-a-Service), and what related factors that influence the enterprise’s adoption of the PaaS technology are.

Recording Confidentiality

I would like to ask your permission to make recording during our conversation. The main reason is because I can get full details about this interview which will help me as well in the analysis phase. Other than that, it will also be more comfortable to us since I can focus on our conversation by not writing too much about the details of our conversation.

The contents of this interview will be processed anonymously. The outcome of this research will be shared with you. The outcome will include an overview of the interview results and the views on the related subject from other companies (anonymous) across different industries, which can be used as a guidance for future reference in adopting cloud computing (particularly PaaS).

PHASE 1 – MODEL VALIDATION

Some lists are handed out to help the respondent to answer some questions.

General Question 1. Have your organization used any kind of cloud-based services? (YES/NO) a. Could you explain more about it? [if yes] b. Which cloud providers are you using? [if yes]

Page | 111

Now, I would like to shift our focus into Platform-as-a-Service. I would like to talk about your company’s business model in utilizing PaaS, from a client perspective. You could tell me your experience or your expectation regarding the use of PaaS in your company.

2. Does your company consider to use PaaS? If not, why?

Service Domain 1. What is the underlying reason that triggers your company to develop a cloud service by utilizing PaaS? a. What is the underlying reason that you consider to use PaaS in the future?  give List-A 2. What kind of cloud application(s) do you build on PaaS now? a. What kind of cloud application(s) do you consider to build on PaaS in the future? 3. Which end-users would you target with your cloud application(s)? a. Which end-users would you target with your cloud application(s) in the future? 4. What added value do you propose towards your end-users by the provision of the cloud application? 2 5. What benefits have you got from utilizing PaaS? a. What benefits do you seek from utilizing PaaS in the future?  give List-B 6. What are the unique characteristics of the service/application that can only be realized by utilizing PaaS? a. What is the level of criticality of work done by the cloud application service? b. What is the level of data sensitivity handled by the cloud application service? c. How is the utilization pattern of the cloud application service? (explain utilization pattern) d. Others?

Technology Domain 7. How the characteristics of the cloud application that you want to develop implies to the technical requirement in terms of security, QoS, and integration with your on premise IT system? 8. What is the most important technical issues in your service/application?

Organizational Domain

9. Which PaaS providers does your company use now? a. Which PaaS providers would you consider to use in the future?  give List-D 10. Could you please name the factors/criteria do you take into account to select the PaaS providers? a. Which criteria would you consider to be important in selecting PaaS providers in the future? b. How would you consider branding (Trustworthiness of PaaS provider) to be an important factor?

2 Questions in grey are the second priority

Page | 112

Appendix

c. How would you consider customer lock-in mechanism by PaaS provider to be an important factor? d. How would you consider the security profile offered by PaaS provider to be an important factor? e. How would you consider Quality of Service offered by PaaS provider to be an important factor? f. How would you consider the integration issue with your existing IT system to be an important factor? g. Are there any other factors that you consider to be important? (e.g. Marketplace, UI?) 11. Apart from the PaaS provider, are there any other actors/players that you consider to be important in the value network of the business model with regards to the provision of the cloud application service? Why? 12. How does the involvement of the actors influenced by the other aspects from the other domains?

Financial Domain 13. How does your company’s IT budget influence the selection of PaaS provider? 14. How does revenue scheme (if applicable) influence the selection of PaaS provider?

Barrier and Risk 15. What barriers do you perceive in adopting PaaS? 16. What risks do you perceive in utilizing PaaS?

Role of Telecom 17. How would you think Telecom company can be involved in the value networks of your PaaS business model? What role would you think they could play?

PHASE 2 - CLOSING 1. Do you like to know more about the research project? 2. Is there anything else you would like to tell me? 3. Would it be possible to contact you latter if I have further questions?

PHASE 3 - CONCLUSION

First of all, allow me to say thank you very much for your time and cooperation in this research. I will process the information gathered through this interview, and I will be keeping in touch with you regarding the result of this research.

Page | 113

INTERVIEW PROTOCOL – VENDORS

Name of the organization : Industry sector : Respondent’s name : Position/role : Date and time :

PHASE 0 - INTRODUCTION

Good morning/afternoon Mr./Ms … . My name is Aryo Primagati. Before I start, I would like to thank you for your participation in this research. This research will last approximately one hour.

The general idea of this interview is to have more insight on what the intended and proposed value of PaaS offerings are, particularly from the vendors’ perspective. Moreover this interview is also aimed to understand the PaaS market condition in the Netherlands.

Recording Confidentiality

I would to ask your permission to make recording during our conversation. The main reason is because I can get full details about this interview which will help me as well in the analysis phase. Other than that, it will also be more comfortable to us since I can focus on our conversation by not writing too much about the details of our conversation.

The content of this interview will be processed anonymously. The outcome of this research will be shared with you. The outcome will include an overview of the interview results and the views on the related subject from other companies (anonymous) across different industries, which can be used as a guidance for future reference in adopting cloud computing (particularly PaaS).

PHASE 1 – VENDOR’S PAAS BUSINESS MODEL

Some lists are handed out to help the respondent to answer some questions.

PaaS Market 1. Which clients/customers use your PaaS offering currently? 2. Which clients/customers would you target with your PaaS offering in the future? Is there any particular reason why? [give list?] 3. What is your main strategy to sell your PaaS offering? a. What is your distinguishing selling point of your PaaS offering compared to the other PaaS vendors? 4. What is your views and strategy regarding the public vs. private cloud?3

Proposed Value 5. How do you think clients/customers would change their Business Model by using PaaS? 6. Which main resource and capabilities do you offer to your PaaS clients/customers?

3 Questions in grey are the second priority

Page | 114

Appendix

7. How do you convey trustworthiness towards your (potential) clients? 8. What level of customer lock-in do you have in your PaaS offering?  give List-C a. Why do you choose to apply that customer lock-in strategy? 9. How would you deal with security issues that becomes one of the main issues in PaaS adoption? 10. How would you maintain your PaaS offering Quality of Service, in terms of up time and availability? 11. How would you differentiate your offering with competitors in terms of pricing?

Marketplace 12. Do you have a marketplace in your PaaS offering? 13. How would the existence of marketplace influence the adoption of PaaS? 14. How would you think the importance of attracting third party application developers to adopt the platform and publish new applications in the marketplace? a. What is your strategy to achieve that? b. What kind of price incentives do you have?

Telecom Role/Complementary actors 15. How would you think Telecom companies can be involved in the value networks of your PaaS business model? What role would you think they could play? 16. Who are other actors you consider to be important in PaaS ecosystem? Why?

Barrier and Risk 17. What factors do you think are preventing the utilization or adoption of PaaS by enterprise organizations? 18. What risks do you think large enterprise clients/customers would face in adopting PaaS?

PHASE 2 - CLOSING

1. Do you like to know more about the research project? 2. Is there anything else you would like to tell me? 3. Would it be possible to contact you latter if I have further questions?

PHASE 3 - CONCLUSION

First of all, allow me to say thank you very much for your time and cooperation in this research. I will process the information gathered through this interview, and I will be keeping in touch with you regarding the result of this research.

Page | 115

LIST-A: Possible Reasons/Goal to Use PaaS

A. To change business process rapidly B. To develop new/improved service/product more quickly C. To develop products/service that were not feasible before D. To enable processes that are not otherwise cost effective or feasible E. To establish uniform processes in different region F. To exchange data more efficiently with outside organizations G. To improve analytical capabilities H. To improve communication and collaboration between individuals I. To provide IT where infrastructure resources is limited J. To provide new ways to engage and interact with customers/suppliers K. To provide platform for standardized efficient business process L. To reduce energy consumption M. To reduce upfront IT capital expenditures N. To reduce cost of IT maintenance

LIST-B: PaaS Benefits

1. Finance a. To reduce (eliminate) capital cost of IT, and replace with operational cost b. To get time-boxed projects done at a low total cost c. To lower TCO 2. Technology a. To get access to state-of-the-art technology b. To get applications and services that are available only in the cloud i. Scalability ii. Internet-based accessibility 3. Organization a. To reduce maintenance and support 4. Service a. To focus on service innovation b. Faster time to market

LIST-C: Customer Lock-In Levels

1. Programming languages that cannot be used on other platforms 2. Frameworks and containers providing infrastructure capabilities with proprietary application programming interface 3. Data, when customers cannot easily retrieve their data in a format that preserves semantics, policy, and context 4. Hosting, when a deployed application can only run on a single vendor’s infrastructure

Page | 116

Appendix

LIST-D: PaaS Vendors

PaaS Providers/ Description Vendors Google App Engine enables you to build web applications on the same scalable systems that power Google applications. Partnering with Salesforce.com to give Google App Engine developers using Google App Engine native access to the capabilities of Salesforce.com platform including multi-tenancy, authentication, scalability, database, and workflow so that their software runs in a business environment. Platform that enables developers to create and deliver any kind of business application, entirely on-demand without software. Adopting their own proprietary SFDC Force.com standard called Apex. Currently with a partnership with Google App Engine. Through this partnership Salesforce.com developers get native access to Google's Bigtable distributed-storage system and the ability to talk to apps in Python. Platform hosted in Microsoft data centers. Supports a consistent development Microsoft Azure experience via integration with Visual Studio including .NET applications. Also supports third party tools such as Eclipse, Ruby, PHP, and Phyton. Offer a solution that focus on the enterprise IT organization. Offer tool to build Web- based mashup applications. The types of applications deemed appropriate include “situational applications” and those that support dynamic business processes. Cordys Cordys also offers an online marketplace to sell applications to end customers, and it supports ISVs with a program that covers partner on boarding, training and support, as well as assistance with pricing models. Platform that targeting replacement of traditional development with a point-and-click platform for building applications. ISV market is not a current focus. Hosts the Caspio platform in a commercial data center facility based in the US. The main emphasis for Caspio is a high-productivity hosted environment for the development and deployment of two-tier, database-centric Web applications. Platform that manages organizations Rails application deployments, from EngineYard small/medium applications running on 3 slices to large applications that automate business operations. Ruby on Rails Platform with instant deployment. When developers create or import an app with Heroku, it’s already live on the web. No configuration or deployment Heroku necessary. Heroku is built on Amazon EC2, thus the apps have access to unlimited computing power. On-demand platform for creating and delivering business applications to manage data, streamline collaborative processes and provide actionable analysis. This robust LongJump PaaS solution reduces the time and costs of developing and delivering secure, database-driven applications, while also providing extensive features round security access, data analysis and visualization, and process automation – all on the web. Platform that position itself as offering a process-centric solution, suitable for situational applications and those with unpredictable volumes or spiky usage. OrangeScape Focuses firmly on reducing the time for its customers to develop applications, as well as the flexibility of deployment. Possible to deploy its service on more than one cloud infrastructure provider. Easy to program by means of visual user interface. Platform that focus on the creation of customizable solutions. Supports nonprogrammers via a drag-and-drop development model. Differentiate itself from integration with third-party services like FedEx. Have an appstore that offers access WorkXpress to its own CRM and ERP offerings. Provide choice of where the user’s applications is hosted, thus supporting deployment where data center location is important. Helping ISCs with market positioning. Support a range of multi-tenancy choice.

Page | 117