金融業與網路科技業導入巨量資料系統的關鍵因素之個案研究 - 政大學術集成

全文

(1)國立政治大學科技管理與智慧財產研究所碩士學位論文. 金融業與網路科技業導入巨量資料系統的關鍵因素之個案研究. Case Studies of Key Factors for implementing Big Data system on Financial and Internet Technology Industries. 指導教授: 吳豐祥博士沈錳坤博士研究生:陳冠廷撰.

(2) ABSTRACT With the popularity of internet, smart phones and Internet of Things begin to emerge. According to Institute for Information Industry, there are approximately 49.5% of smart devices in Taiwan, which mean every two people will own at least one smart device. In addition, more devices are connected to the internet. Therefore, tremendous amount of data is created and increased exponentially. With applicable and correct techniques, Big Data can provide valuable insight and business intelligent. Traditional industries are forced to change. For example, Uber is one innovative idea that changes the ways people ride taxi. Riding taxi become more efficient and effective with Uber. This research explores critical factors of Big Data implementation on financial and internet technology industries from three perspectives. This includes key processes of the Big Data implementation, enterprises factors, and the Big Data system. Moreover, literature review was conducted to. In addition, three case studies were interviewed and analyzed based on research framework. Lastly, three research questions are answered. First, what are the key process for financial and internet technology industries implementing the Big Data system? Second, what are the critical factors for financial and internet technology industries implementing the Big Data system? Third, what are the potential benefits after the Big Data implementation? The research findings are primarily categorized into two parts. First, there are three phases of financial institution and internet technology industries implementing the Big Data system. The three phases included defining, brainstorming and implementation phases. The three phases are described below: 1. Defining Phase: Companies will first define their own interpretation of Big Data in order to plan and coordinated their implementation. 2. Brainstorming Phase: Companies averagely spent most of the time in this phase. The implementation team leads must brainstorm to find the best way to enforce and carry out the Big Data project by searching, organizing and surveying internal and externally. 3. Implementation Phase: Companies follow their previous made proposal steps by steps. This research also concluded and found several critical factors during the Big Data implementation. The critical factors included but not limited to: 1. An implementation team regardless the size to carry out the Big Data project 2. Top management’s commitment on implementation 3. Timing on the implementation 4. Big Data system selection 5. Clear goals and objectives Keywords: Big Data, Internet Technology, Financial Industry, Critical Factors. I.

(3) 摘要隨著網際網路的普及，智慧型手機與物聯網開始興起，根據資策會的調查，台灣約有 49.5%的智慧型手機占有率，大約每 2 人就有一人擁有智慧型手機，而物聯網的興起，製造了大量的數據與資料，而這些數據與資料透過不同的處理方式，可帶給企業不同的商業智慧與洞見，而傳統產業因此面臨了巨大的轉變與挑戰，優步(Uber)就是改變傳統計程車產業與物聯網的一個例子，顧客不再需要招手才能搭上計程車，靠著網路、手機 APP 與 GPS 定位系統即可獲知車輛資訊、到達時間與聯絡司機，而優步可掌握乘客資訊、行車路線與顧客服務。不僅僅是計程車產業，亞馬遜的崛起也代表了傳統零售業的轉變，因此如何面對巨量資料對傳統企業都將是一項挑戰。巨量資料的導入與分析可以提供企業掌握消費者行為，也可透過數據分析研發新服務與產品，此研究從三方面來探討金融產業與網路科技業導入巨量資料系統的關鍵因素，分別為導入流程、企業本身與巨量資料系統，另外藉由三家個案公司訪談，並輔以文獻所探討的研究架構來進行驗證”金融業與網路科技業導入巨量資料的流程為何?”、”金融業與網路科技業導入巨量資料時的關鍵因素?”、” 金融業與網路科技業導入巨量資料後有何優勢?” 本研究最後發現，金融業與網際網路業導入巨量資料分成三個階段，首先企業會先詮釋自身對巨量資料之定義，定義自身巨量資料之意義後，企業會開始集體研討導入流程，依照自身對巨量資料的詮釋來集體擬定對企業最好的導入流程，此階段通常會是三階段中耗時最長，也需做許多內外部研究、規劃與管理。最後一階段為實做階段，企業會依照集思廣益後所擬定出的計畫來完成巨量資料的導入。而本研究透過個案訪談也發現七項導入巨量資料之關鍵因素，包含，導入隊伍的組成、高層管理者的支持、導入時機、巨量資料系統的選擇、明確的目標與策略、內外部員工訓練與支援。最後企業運用第三方與開放式資料軟體來處理巨量資料使企業更了解顧客需求與運用於新產品研發。關鍵字 : 金融業、網路科技業、巨量資料、關鍵因素 II.

(4) 致謝 “論文，是研究生涯中送給自己最好的禮物”還依稀記得這是當初與陳世倫學長討論論文架構時學長對我說的話，每一本論文都是花了無數的心血與努力，透過不斷投入的時間與思考才有論文的誕生，同樣，每一本論文的誕生也會透過許多貴人的協助與幫忙，因此有幾位非常重要的人我想在這裡提到。感謝吳豐祥教授在我在科管所求學期間與論文中所給予的教導與教誨，老師從未否定過我們的想法，而是以循序漸進的方式給予我們適時的方向與協助，讓我們可以在自己的想法中盡情發揮，並且磨練我們的思考，不管老師有多忙，總是會抽出時間來跟大家一起開會與討論。感謝沈錳坤教授在此論文給予我的協助與付出，在我人生與求職生涯中碰到挫折時鼓勵我與開導我，在求學與撰寫論文碰到困難時協助我與教導我，二位教授不僅僅是我求學中的導師，更是我人生中的貴人。感謝我的父母在這 25 年給我的一切，在我成長與念書生涯中，讓我豐衣足食，在我失敗時不斷的給予我鼓勵，秉持著只給釣竿不給魚的精神，不斷的磨練我與教導我，希望下輩子可以再當你們的小孩。感謝彭韻倫助教在我撰寫論文與口試時不辭辛勞的幫忙，從幫忙我聯絡教授、訪談對象到口試時幫忙我準備口試事宜，很高興可以在科管所認識妳並和妳成為好朋友。此論文的完成還要感謝個案中的訪談對象，感謝各位公司主管的時間與幫忙，協助我完成論文。最後感謝我女朋友佳玲，謝謝妳給予我的支持與包容，從我決定回台灣繼續就學到現在與我在台灣一起為未來而努力，妳是我堅持下去的動力。. 陳冠廷謹誌. III.

(5) 2015 年 7 月. TABLE OF CONTENTS ABSTRACT ..................................................................................................................................... I 摘要................................................................................................................................................. II 致謝............................................................................................................................................... III TABLE OF CONTENTS .............................................................................................................. IV LIST OF FIGURES ...................................................................................................................... VI LIST OF TABLES ....................................................................................................................... VII Chapter 1. Introduction ................................................................................................................... 1 1.1 Motivation and Background .................................................................................................. 1 1.1.1 Big Data, Big Value........................................................................................................ 2 1.1.2 Implementing Big Data System for Enterprises ............................................................. 3 1.2 Research Questions and Objectives ...................................................................................... 4 1.3 Research Process ................................................................................................................... 5 Chapter 2. Literature Review .......................................................................................................... 7 2.1 Big Data................................................................................................................................. 7 2.1.1 Definition of the Big Data .............................................................................................. 8 2.1.2 Big Data Challenges in Management ........................................................................... 10 2.1.3 Approach to Big Data ................................................................................................... 12 2.2 ERP Implementation Process .............................................................................................. 15 2.2.1 History of ERP and Definition ..................................................................................... 15 2.2.2 ERP Implementation Motivation and Benefits ............................................................. 17 2.2.3 ERP Implementation Definition and Process ............................................................... 18 2.3 Critical Factors for ERP Implementation ............................................................................ 28 Chapter 3. Research Method and Design ...................................................................................... 34 3.1 Research Framework ........................................................................................................... 34 3.2 Research Method ................................................................................................................. 36 3.3 Research Target ................................................................................................................... 36 3.4 Data Collection Method ...................................................................................................... 37 3.4.1 Interview Design........................................................................................................... 37 IV.

(6) 3.4.2 Question Design ........................................................................................................... 38 Chapter 4. Case Study ................................................................................................................... 39 4.1 Case Study of Company A .................................................................................................. 39 4.1.1 Background: Financial Company A ............................................................................. 39 4.1.2 Implementation and Process ......................................................................................... 41 4.1.3 Enterprise Requirements............................................................................................... 44 4.1.4 Benefits after Implementation ...................................................................................... 46 4.1.5 Big Data System Performance (Company A)............................................................... 47 4.2 Case Study of Company B .................................................................................................. 48 4.2.1 Background: Financial Company B ............................................................................. 48 4.2.2 Implementation and Process ......................................................................................... 50 4.2.3 Enterprise Requirements............................................................................................... 51 4.2.4 Benefits after Implementation ...................................................................................... 52 4.2.5 Big Data System Performance (Company B) ............................................................... 53 4.3 Case Study of Company C .................................................................................................. 56 4.3.1 Background: Company C ............................................................................................. 56 4.3.2 Implementation and Process ......................................................................................... 57 4.3.3 Enterprise Requirements............................................................................................... 59 4.3.4 Benefits after Implementation ...................................................................................... 60 Chapter 5. Research Findings and Discussion ............................................................................. 63 5.1 Implementation Process ...................................................................................................... 64 5.2 Enterprise Requirements ..................................................................................................... 66 5.3 Performance of the Big Data System .................................................................................. 67 5.4 Relationship Discussion between Implementation process, Enterprise requirements and Big Data system ............................................................................................................................... 68 Chapter 6. Summary and Recommendation ................................................................................. 73 6.1 Research Summary.............................................................................................................. 73 6.2 Research Limitation and Recommendation ........................................................................ 74 References ..................................................................................................................................... 76. V.

(7) LIST OF FIGURES Figure 1. Research Process ............................................................................................................. 6 Figure 2. Total Amount of Data Generated on Earth...................................................................... 8 Figure 3. IBM Big Data Adoption Pattern .................................................................................... 13 Figure 4. PwC Big Data Approach ............................................................................................... 14 Figure 5. The ERP Lifecycle Framework ..................................................................................... 20 Figure 6. Theoretical Framework for ERP Implementation Management ................................... 22 Figure 7. Five Major Stages for ERP Implementation Process .................................................... 24 Figure 8. Five Major Phases for ERP Implementation Process .................................................... 26 Figure 9. Three Phases of Implementation ................................................................................... 31 Figure 10. Bank A Governance Structure ..................................................................................... 40 Figure 11. A Three- Tier Data Warehousing Architecture ........................................................... 43 Figure 12. SinoPac Structure ........................................................................................................ 49 Figure 13. IBM PureSystem Advantages...................................................................................... 54 Figure 14. Hype Cycle for Emerging Technologies ..................................................................... 58 Figure 15. Company C School-Career Development Map (104, 2011) ....................................... 61 Figure 16. Company C Career Wikipedia (104, 2011) ................................................................. 61 Figure 17. Three Phases of Big Data Implementation .................................................................. 66. VI.

(8) LIST OF TABLES Table 1. ERP Historical Account .................................................................................................. 15 Table 2. ERP Definition ................................................................................................................ 16 Table 3. ERP Implementation Life Cycle ..................................................................................... 27 Table 4. Compared & Contrasted table for ERP implementation ................................................ 28 Table 5. ERP Implementation Framework with Critical Factors ................................................. 29 Table 6. Critical Factors Organized .............................................................................................. 31 Table 7. Data Warehouse Specifications ...................................................................................... 47 Table 8. IBM PureData System Advantages ................................................................................ 55 Table 9. Company A, B & C Comparison Based on the Research Framework ........................... 63 Table 10. Company A&B Comparison Based on Research Framework ...................................... 68 Table 11. Company B & Company C Comparison Based on Research Framework ................... 70. VII.

(9) Chapter 1. Introduction 1.1 Motivation and Background Moore’s Law states that processor speed, or overall process power for computer will be doubled every two years. This same logic can be applied to technology and engineering as well. In the 21st century, internet, software and hardware are evolved tremendously rapid. According to Institute for Information Industry, there are approximately 49.5% of smart devices in Taiwan, which mean every two people will own at least one smart device. This report indicates that the internet has become essential in the 21st century. In addition, everyone and everything can be, also will be connect to the internet. From that, many of the traditional industries are forced to change themselves tremendously, and new innovative businesses have emerged accelerated. For example, People usually go to traditional supermarket for grocery shopping. However, ever since Amazon was founded in 1994, the behaviors people shop for consumer goods has become incredibly different. With Amazon’s two days delivery and enormous variety of product lines, people purchase their consumer goods from Amazon instead of traditional retailers. There are additional examples such as Uber, Yelp and Airbnb, which they all have one common advantage. They are all connected to the internet. Banking is one traditional industry that has not yet been evolved dramatically, but faces huge challenges. Because of the internet and smart devices are broadly available, most of the people can access and transact their bank accounts through the internet. Smart devices allow people to handle their daily financial needs without going to an actual bank. “This is the way banking will be done from this day forward, without exception. We’re never going back to a world without internet banking access, mobile phones, social media and multi-touch” (King, 1.

(10) 2012). Therefore, financial industries need to refocus the ways they communicate with customers and investors. Furthermore, financial industries need to be prepared for the internet era. 1.1.1 Big Data, Big Value The words “Big Data” are emphasized extremely frequent in the past few years. Big Data is not a new technology or strategy. In fact, Big Data does not simply mean the size of the data, but more importantly the strategies and the engineering applied within. For instance, text-mining and machine learning are part of the technical techniques in Big Data. Moreover, Most of the enterprises are able to store and collect huge amounts of data because the cost of hardware declined periodically. In addition, cloud computing technology is improved enormously for the consumption of computing resources. Therefore, a lot of Big Data applications are available such in retailers, medicals, telecommunications and governments. Big Data includes structured, semi-structured, and unstructured information from demographic and psychographic information about consumers to product reviews and commentary; blogs; content on social media sites; and data streamed 24/7 from mobile devices, sensors, and technical devices. The digital era is pushing the financial institutions on many fronts, such as customer data, market expectations, and operational efficiencies (PWC, 2013). Therefore, processing increasingly large volumes of data in a timely manner has become a major challenge for financial institutions. Financial institutions should focus on using Big Data to get the right information to identify the right markets and customers at the right time, which enable institution to make the right strategic decisions. In fact, 62% of companies believe that Big Data has significant potential to create competitive advantage (PWC, 2013). Moreover, the Big Data market is at 5.1 billion in 2012 and is expected to grow to 32.1 billion by 2015 (Alspach, 2012).. 2.

(11) 1.1.2 Implementing Big Data System for Enterprises Financial institutions must find a new way to analyze and interpret their customers. More importantly, they have to be ready for the internet era and the Big Data boom. Therefore, implementing a Big Data system is essential for the financial institutions. However, many problems and difficulties will occur during the implementation. For instance, in 1990, Enterprises Resource Planning is first introduced by the Gartner Group, and the ERP system is defined as seamless integration of all information through a company. The system provides integration includes, but not limit to financial information, human resource information, supply chain information, and customer information (Davenport, 1998). There is no surprised that the system solves a lot hassles for companies. For instance, many managers struggle with incompatible information systems, and inconsistent operating practices. Therefore, ERP system provides easy solutions and consistent operating practices across all departments effortlessly. Besides the fact that ERP systems offer various benefits, many problems and difficulties can occur during the implementation because every company is distinctive. ERP implementation can be categorized in three processes, which are pre-implementation (Setting-Up), implementation and post-implementation (Motwani, Subramanian, Gopalakrishna, 2005). In addition, each category has several key components to be considered before ERP implementation. For example, during the pre-implementation, company should have clear understanding of strategic goals for ERP. Moreover, the commitment from top-level managers is also important during the pre-implementation stage. After the first stage, companies have to choose which ERP package selection that best fits with their current business models. According to the article “Critical factors for successful ERP implementation” by Motwani, IT Leveragability and knowledge capability are important during the implementation stage. Furthermore, company 3.

(12) should assemble an extraordinary team to lead the implementation and monitor on the progress. Lastly, companies usually evaluate their ERP implementation by monitoring their ERP system’s implementation. In the post-implementation stage, companies can be more adaptable to the change of programs and adjust their system by continuously monitoring to derive the maximum benefits from ERP. By taking a glance back to the ERP era, companies understand ERP provides beneficial contributions to the company. However, many problems and difficulties can be occurred during the implementation. Companies need to consider several components include, but not limit to clear business objective, comprehension of the nature of changes and understanding of the project risk (Mandal, Gunasekaran, 2003). In addition, strong leadership and constant watch to budget are essential as well (Wagle, 1998). From the previous descriptions and reasons, the implementation for the Big Data system can be both beneficial and rewarding in compare with ERP system. However, it can have several challenges also similar to ERP. Therefore, this research explores on questions such as what processes are needed or developed for financial institutions implementing the Big Data system. 1.2 Research Questions and Objectives According to the research background and motivation above, this research will focus on financial institution that implementing the Big Data systems. More importantly, the implementation processes and the Big Data systems itself along will be discussed. Furthermore, the key factors and requirements for implementation regarding enterprise are also studied in this research. Therefore, several research questions will be studied and discussed in below. 1. What are the key processes for financial institutions to implement Big Data system? 4.

(13) 2. What are the essential critical factors for financial institutions implementing the Big Data system? 3. What are the preliminary benefits for post-implementation on Big Data system? 1.3 Research Process There are several reasons that compiled the research processes in this study, which will be conducted and followed throughout the sections. First, this paper explains the motivations and the possible research target for the study. After collecting literatures and details, this study proposed the potential research objectives and questions. Second, in order to introduce potential research framework, literature reviews will be conducted. Research questions and framework are possible to modify based on the literature finding and evidence. After thorough and complete research framework, individual case study and proper interviews will be conducted subsequently. Finally, this research will organize and analyze all the materials from the case study, interview and literature reviews to give appropriate research finding and opinions. Figure 1 indicates the research process:. 5.

(14) Research Scope & Targets. Research Questions & Objectives. Related Literatures Review. Construct Research Framework. Interview Financial Institution. Case Study & Analyze. Proper Finding & Opinions. Figure 1. Research Process. 6.

(15) Chapter 2. Literature Review Implementing Big Data system can be quite comparable with implementing Enterprise Resource Planning. It is controversial to say that the implementing processes for the ERP and the Big Data are similar. However, the fact is that there are more literatures in ERP implementation than Big Data system. In this section, literature review will first interpret what Big Data is. Then, this research will focus on conducting both broad and rooted ERP implementation process literatures. Lastly, several of the critical issues or factors on implementing ERP will be discussed in this section as well.. 2.1 Big Data Cloud computing and Internet of Things (IOT) has become two main leading directions for the most of technology companies in the year of 2015. Due to the accelerated internet speed and economic data storage, cloud computing and the development for smart devices has emerged exponentially. According to the article “Sensing as a Service and Big Data” by Zalavsky, Perera and Georgakopoulos from Research School of Computer Science, the Australian National University, Internet of Things will comprise billions of devices which can sense, communicate, compute and potentially actuate. Therefore, data streams coming from these devices will definitely challenge the traditional approach to management. How big is the data in the past and present? In 2010, the total amount of the data on earth exceeded one zettabyte (see Figure 2) (Zikopoulos, 2012). By the end of 2011, the number grew up to 1.8ZB (Reed, Gannon, Larus, 2012). Moreover, according to Gartner Research, wirelessly networked devices will form a new Web, and only the terabyte of data it generates can be collected, analyzed and interpreted. Ultimately, companies and industries will be destined to prepare for the Big Data wave. 7.

(16) Figure 2. Total Amount of Data Generated on Earth Source: Zikopoulos, 2012. 2.1.1 Definition of the Big Data From the introduction section above in this research, raw data can be categorized in structured, semi-structured and unstructured, but how much data is enough to be called Big Data? In this section, the definition of the Big Data will be discussed thoroughly and intensively. “Information is the oil of the 21st century, and analytics is the combustion engine,” the SVP of Gartner’s research, Mr. Sondergaard said. (Gartner, 2010). Big Data has been generally recognized, but people still have different opinions on its definition. Below is a definitions list organized by this research. 1. According to Apache Hadoop in 2010, Big Data is defined as “datasets which could not be captured, managed, and processed by general computers within an acceptable scope.” 8.

(17) 2. NIST defines Big Data as “Big Data shall mean the data of which the data volume, acquisition speed, or data representation limits the capacity of using traditional relational methods to conduct effective analysis or the data which may be effectively processed with important horizontal zoom technologies.” 3. According to Oracle, Big Data is the foundation of value from traditional relational database driven business decision making. In addition, there are many new sources of unstructured data, which included blogs, social media, sensor network, and image data. They are all vary in size, format and other factors. 4. McKinsey & Company defines Big Data which could not be acquired, stored, and managed by classic database software. The volume of a dataset is not the only criterion for Big Data. However, the increasingly growing data scale and its management issues are the two key features in the future. From the above list, Big Data can be identified in several characteristics. According to Laney (2001), a VP researcher at Gartner, stated in his original work that Big Data consisted of 3V framework. First, Big Data has tremendous amount of “volume”. There are terabytes to exabytes of existing data to process. In addition, the data is tremendously enormous that traditional storage approach and data processing method are not capable to handle. Second, data “velocity” is another characteristic involved in Big Data. It is explained as the data in motion, which streaming data, and data collection must be rapidly and timely in order to maximize the value of Big Data. According to Laney (2001), E-commerce has increased point-of-interaction (POI) speed, which is increasingly perceived as a competitive differentiator. POI performance differentiates web site response, inventory availability analysis, transaction execution, order tracking update and product delivery. Therefore, data velocity management is more than a. 9.

(18) physical bandwidth and protocol issue. Furthermore, Variety indicates the several forms of data, such as structured, semi-structured, and unstructured. Data sources can come from social media, YouTube, Twitter and smart wearables. Apparently, everyone produces enormous and various amount of data. Lastly, IBM (2012) stated that veracity should be the 4th V in Big Data because there is uncertainty of data. This uncertainty is due to the data inconsistency or incompleteness. 2.1.2 Big Data Challenges in Management Big Data has many beneficial value, but also it creates several inevitable complications regarding the both technical and managerial. From the managerial point of view, the first challenge is how decision is made with Big Data. According to Harvard Business Review “ Big Data: The Management Revolution” by McAfee and Brynjolfsson (2012), although a number of senior executives are willing to override their own institution when the data don’t agree with it, they believe most of the people rely too much on experience and institution, not enough on data. In fact, they constructed a 5-point scale research to capture the overall extent to which a company was data-driven. Their research shows 32% of their respondents rated their companies at or below 3 on the scale. Therefore, it is difficult for company to transit from traditional decision-making to data-driven. According to McAfee and Brynjolfsson (2012), in order for companies to achieve the full benefit of Big Data, there are five management challenges that companies should carefully reviewed and considered. The five management challenges are organized as below: 1. Leadership: “Big Data’s power does not erase the need for vision or human insight.” (McAfee and Brynjolfsson 2012). Leadership teams should set clear goals, define what success like, and ask the right question. For example,. 10.

(19) executives should ask question such as “what do the data say?” or “what kind of analyses were conducted?” Leadership teams should be able to spot and understand the opportunity from data. In addition, they will be able to provide compelling reasons while changing the way their organizations make decisions. 2. Talent Management: There are several crucial skills and techniques in high demand for company, who wants to capture the value of Big Data. Statistic is basic, and the ability to organize and clean large dataset in variety of formats, are far more crucial. According to the Harvard Business Review “Data Scientist- the sexiest job of the 21st century” by Davenport and Patil (2012), the most basic, universal skill is the ability to write code. More importantly, data scientists need to be able to communicate in language that all their stakeholders understand. Last but not least, data scientists must demonstrate the ability involved in storytelling with data, whether verbally, visually or both. 3. Technology: Although most of the Big Data softwares are not extremely expensive, and they are usually open source. For example, Hadoop is a commonly used open-source framework, which contents the tools for analyzing the data. However, most of the IT departments have to acquire a new set of skills in order to use Hadoop, and integrate all the relevant internal or external data. 4. Decision making: An exceptional leader must create an organization, which can effectively put information and the relevant decision rights in the same location. In addition, the organization should be flexible enough to maximize crossfunctional cooperation.. 11.

(20) 5. Company Culture: Most of the companies pretend to be more data-driven than they actually are. For example, many executives will accept the report with a lot of data that support the decision they had already made by instinct or experience. The data report is made with only to find the number that would justify the decision. It’s obviously easy to mistake correlation for causation and to find the misleading pattern in the data (McAfee and Brynjolfsson 2012). Therefore, organizations need to ask “What do we know?” instead of “What do you think?”, and company culture challenges are definitely enormous. 2.1.3 Approach to Big Data Although there is no concrete and specific implementation process for Big Data, three literature reviews do have suggested approach on how organizations should adopt the Big Data. These framework and approach are provided by IBM, PWC and Harvard Business Review. 1. According to IBM Institute for Business Value (2014), their research shows four main stages for organization adopting the Big Data. Four main stages are included educate, explore, engage and execute. In the education stage, organizations should find out what they have, and establish the ability to access it. Therefore, it’s important for organizations focus on knowledge gathering and market observation. Second, organizations should be able to find, visualize and understand the Big Data in the explore stage. In addition, they should develop strategy and roadmap based on business needs and challenges. Furthermore, in order to engage, organizations begin piloting the Big Data initiative to validate value and requirements. Lastly, organizations should be able to execute the Big Data. 12.

(21) initiatives, and continually apply advanced analytics. See Figure 3 for the IBM four stages chart:. Figure 3. IBM Big Data Adoption Pattern Source: IBM, 2014. 2. Pricewaterhouse Coopers (PwC) provides five well-defined approaches to Big Data (Fig. 4). First step is to consider if the Big Data is the right tool by identifying potential Big Data opportunities. In addition, organizations should research then determine if Big Data has been used previously to solve the problem in question. Later on, organizations should develop Big Data organizational structure and confirm that the Big Data goals align with firm goals. Third, organizations will establish business case evaluation such as determine the cost to establish a Big Data organization, also outline benefits and risks to the organizations. Next, organization will enable opportunity with Big Data by doing pilot, assess, and operationalize. Organizations will enable opportunity with Big Data by running pilot, and assess the potential business value. In this stage, organizations will make decision on seize or not for the opportunity. Lastly, 13.

(22) organization will keep improving and evaluating their Big Data capability. (PwC, 2013).. Figure 4. PwC Big Data Approach Source: PwC, 2013. 3. Harvard Business Review (2012) provided a more straightforward process on organizations approach to Big Data. The process is consisted with four phrases. The first phrase is suggested that organizations should pick a business unit to be the testing ground. This unit should include a quant-friendly leader with a team of data scientists. Second, organizations will identify five business opportunities with each key functions based on Big Data. Each of which could be prototyped within five weeks with team no more than five people. Third, a process of innovation will be implemented, which is included experimentation, measurement, sharing and replication. Lastly, organizations should open up some of their data set and analytical challenges to interested parties across the internet. 14.

(23) 2.2 ERP Implementation Process In contrast with the organizational approach to Big Data, there are tremendous more concrete literature reviews in ERP implementation process, which will provide a more precise and definite perspective in implementation. In this section, the definition of ERP and history will be discussed first, then the research will get in depth with various ERP implementation process from pre-implementation to post-implementation. 2.2.1 History of ERP and Definition According to article “Enterprise resource planning (ERP) - A brief history” In the 1960s, material requirements planning (MRP) is the backbone of MRP II and ERP. The early MRP system provides a method for planning and scheduling material for complex manufactured products. In the 1970s, when the primary competitive shift towards marketing, which require in the adoption of target-market strategies with emphasis on broader production integration and planning. MRP became established as the fundamental parts in the production management and control. Until early 1990, Gartner Group provided a more definite term for the enterprise resource planning (ERP), which included criteria for evaluating the extent that software was actually integrated both across and within the various functional silos (Wylie, 1990). In conclude, there are four phases of the ERP history (Table 1). Table 1. ERP Historical Account The 1960’s. Most of the software packages were designed to handle inventory based on traditional inventory concepts.. The 1970’s. The focus shifted to MRP systems, which translated the master schedule built for the end items into time-phased net requirements for sub-assemblies, components and raw. 15.

(24) materials planning and procurement. The 1980’s. The concept of MRP-II systems evolved, as an extension of MRP to shop floor and distribution management activities.. The early 1990’s. MRP-II was further extended to cover areas like engineering, finance, human resource and project management. Hence, the term ERP (enterprise resource planning) was coined.. Source: de Sousa, J. M. E. (2004). ERP has variety of definitions, which can trace back as early as 1990s by the Gartner’s group. According to article “ Harnessing ERP systems with Knowledge Management Capabilities” by Fadlalla from George Washington University, Garter defined ERP as a business strategy and a set of industry-domain-specific applications that build customer and shareholder value by enabling and optimizing enterprise and inter-enterprise collaborative operational and financial processes. See table 2 below for the organized definition: Table 2. ERP Definition Author. Definition. Blackstone and Cox (2005). Framework for organizing, defining, and standardizing the business processes necessary to effectively plan and control an organization so the organization can use its internal knowledge to seek external advantage. Ptak and Schragenheim (1999). ERP is defined as an integration system which combined tools, technology, and application.. Kale (2000). ERP is an industry term for the multi-module application software that allows a company manages the set of activities and transactions necessary to manage the business process for moving a product from the input stage, along the value chain, to the final customer.. 16.

(25) Gould (1997). Enterprise resource planning (ERP) is to planning with manufacturing execution system (MES) is to scheduling. ERP is also a software system for planning production, and consolidated performance information about all other aspects of the global enterprise, including finance, customer service, sales and marketing.. Fu & Hsiung and Tai (2004). ERP combine both business processes in the organization and IT into one integrated solution, which MRP and MRP II were not able to provide.. Bingi and Sharma and Godla. An ERP system is an integrated software solution that spans the range of business process that enables companies to gain a holistic view of the business enterprise.. 2.2.2 ERP Implementation Motivation and Benefits Above details thoroughly lists out the history and the definition of ERP. In this section, Motivations and potential benefits will be discussed for organizations implementing ERP. ERP is the most important development in the corporate use of technology in the 1990s (Davenport, 1998). According to Ross and Vitale (1998), there are six common motivations for company to adopt ERP, included the need for a common platform, process improvement, data visibility, operating cost reductions, increased customer responsiveness, and improved strategic decision making. In Deloitte’s research (1998), it concluded that a solution of technological problems and a vehicle for solving operational problem are two main reasons for ERP adoption. Therefore, ERP adoption can lead to potential mixed benefits and risks for the organizations regarding financial performance or operational. One research from Appleton (1997) estimates that half of ERP implementation failed to meet expectations. In addition, 70% of the ERP implementations fail to achieve their estimated benefits (Al-Mashari, 2000). However, One survey shows that. 17.

(26) organizations acquire benefits such as an increase in supplier’s and customer’s satisfaction and an increase in productivity, but the level of the ROI is rather low (Themistocleous et al., 2001). 2.2.3 ERP Implementation Definition and Process ERP systems are integrated to create a seamless software application link between all the process and application (Hendrickson, 2000). In addition, it supports many aspects of an organization’s information needs (Davenport, 2000). In this section, the term “ERP Implementation” will be defined first. Then, several ERP implementation process will be studied thoroughly and also serve as potential models for Big Data implementation. Walsham (1995) mentioned that the term implementation “is sometimes used to mean technical implementation, namely ensuring that system development is completed and that the system functions adequately in a technical sense. At other times, it is used to refer to the human and social aspects of implementation, such as that the system is used frequently by organizational members or that it is considered valuable to them in their personal work activities or coordination with others.” On the other hand, Krammergaard and Moller (2000) defined that ERP implementation is different according to consultants, vendors and organization's view. Implementation is often used as a term to describe a well-defined spanning from the choice of the systems through the configuration and the training until going live, where the system is becoming operative. ERP Implementations process can be varied across the organizations. There is no standardized model or process for organizations to adopt since every organization is distinctive. Therefore, several suggested ERP implementation process will be reviewed and discussed regarding the steps and appropriateness.. 18.

(27) 1. According to Esteves and Pastor (1999), their research develops a six phases ERP adoption cycle with four dimensions (Figure 5). The phases included adoption decision, acquisition, implementation, maintenance, evolution and retirement phase. In addition with the four dimensions, this included product, process, people and change management. This framework is useful for identifying the origins and impacts of change, and thus provides a way of identifying and characterizing research issues in ERP system. In the ERP cycle, adoption phase is where manager question the need for a new ERP system. The definition of the system requirements, goals, impacts and benefits will often be analyzed in this phase. In the acquisition phase, organizations will do the product selection, which best fits the requirement and the strategy of the organizations. In addition, organizations often analyze the ROI in this phase. Later, implementation phase consists of customization, parameterization, and adaptation of the ERP package acquired by the organizations. Consultants usually are needed in this phase regarding the implementation methodologies, know-how and training. Next, organizations should expect the use of product return benefit and minimizes disruption. Organizations must be aware of the functionality, usability and adequacy of the ERP and business process in order to do maintenance after implementation. Furthermore, evolution phase are consisted by the integration of more capabilities into the ERP system. Organizations will search and analyze for potential new benefits or external collaboration with other partners. Lastly, managers decided if they will substitute the inadequate ERP system with other information system to fit the business needs of organizations.. 19.

(28) Figure 5. The ERP Lifecycle Framework Source: Esteves & Pastor, 1999. 2. "A well-planned and well-executed ERP implementation in conjunction with a good change management program can create a dramatic turnaround for the company.” (Davenport, 2000). According to Kettingers and Grover (1995), they proposed a framework, which consist seven constructs for ERP implementation management (Figure 6). These constructs included strategic initiatives, learning capacity, cultural readiness, information technology, network relationships, change management practice, and process management practice. At first, the process will begin with strategic initiative, which the organizations react to a need to leverage the potential opportunity. For example, company’s inability to provide adequate customer service (Earl, 1994). Second, learning capacity is to provide positive outcome through effective adaptation to environmental change, also improved efficiency in the process learning. In the culture readiness, open 20.

(29) communication and information sharing can promote a common culture and innovative behavior for business process change (Kilman, et al. 1986). In this constructs, organizational culture facilitate the integration of individual learning with organizational learning, which ensure higher rate of successful implementation for ERP (Guha et al., 1997). Likewise, information technology leveragability and knowledge-sharing capability are important because evidence suggests that IT led project often fail to capture the business and human dimension in the process. Furthermore, research indicated that under most circumstances cooperative, interpersonal and group behavior result in superior performance during the process (Johnson, 1989). Lastly, change management process and process management practice are important in company transformation.. 21.

(30) Figure 6. Theoretical Framework for ERP Implementation Management Source: Kettingers & Grover, 1995. 3. According to Madsen and Ehie (2005), they provided five major phases for ERP implementation process (Figure 7). In the project preparation phase, enterprises analyze the driving motive for ERP implementation. Then, they should create a comprehensive planning process involving people handling leadership roles, establishing budget targets, and determining the project plan. Also, the analysis will be conducted for the system selection before extensive education and training. Third, enterprises will focus on developing the technical foundation while testing each process design on a conference room level. Moreover, the entire process design integration will be tested under full data load and extreme situation. 22.

(31) Likewise, the people intended to use the system or influenced by it will go through the education and training. Lastly, the live and support phase will focus on continuous expansion of the system to enjoy new competitive advantage.. 23.

(32) Figure 7. Five Major Stages for ERP Implementation Process Source: Madsen & Ehie, 2005. 24.

(33) 4. According to Winkelmann and Klose (2008), there are five phases during the ERP implementation (Fig. 8). These phases are included project initialization, as-is analysis, to-be design, realization prototyping, and operate. In the project initialization phase, enterprises usually design the project plan, and train the core team. During the as-is analysis, enterprises examine the functional range of the asis IT solution and business documents. In addition, SWOT analysis is usually included, and conducted in the context of the earlier software evaluation. Third, definition of the interface and concept of variant configurator are included in the to-be design phase. In the realization prototyping phase, enterprises will execute several functional and integration tests. Also, enterprises will begin the creation of business documents forms and start the user training. At the last stage, enterprises should operate the ERP by doing data migration from test system, tuning the technical or behavioral optimization, and adjust to user requirements.. 25.

(34) Figure 8. Five Major Phases for ERP Implementation Process Source: Winkelmann & Klose, 2008. 5. The UK National Computing Centre (1998) first introduced a structured approach on 1960 for system development life cycle (Table 3). This development life cycle was later in compared or contrasted with the ERP implementation life cycle provided by SAP (Table 4). From the compared and contrasted table, there are lots of similarities components between the system development cycle and ERP implementation cycle. One study (Ahituv et al., 2002) defined phases of the ERP life cycle based on an integration of the traditional SDLC, prototyping, and application package development models. As shown in Table 4, both life cycle. 26.

(35) focuses on planning and analysis in the first two stages, and support in the last stage. Both life cycle share common stages and a common set of activities. Table 3. ERP Implementation Life Cycle Life Cycle Stage. Activities. Feasibility Study. - examine current system - identify problems - identify solution - evaluate costs/ benefits. Systems Investigation And Analysis. - analyze old/new system - analyze processes - analyze functionality - analyze users. System Design. - overview functional elements - detail inputs/outputs - detail screen/interface design - hardware requirements - operating instructions - testing/changeover procedures. Implementation. - build/program system - test system and fix errors - install hardware - train users - change over new system. Source: UK National Computing Centre, 1998. 27.

(36) Table 4. Compared & Contrasted table for ERP implementation. Source: SAP, 2006. 2.3 Critical Factors for ERP Implementation After thorough reviews on several ERP implementation model and process, critical factors regarding the ERP implementation will be discussed in this section. What factors facilitate the ERP projects? More importantly, what critical issues need to be considered during each stage of the implementation? These are the two questions, which will be focused in this section. 1. In Motwani and Subramanian’s research (2005), they re-construct a framework for ERP implementation with detailed critical factors within each stage (Table. 5). They concluded that a cautious evolutionary, bureaucratic implementation process backed with change management, network relationship, and cultural readiness can 28.

(37) lead to successful ERP implementation. Moreover, their research clearly indicated that vision and top management commitment are fundamental for ERP implementation. In addition, evaluation and proper monitoring are also critical to the post-implementation. Table 5. ERP Implementation Framework with Critical Factors Pre-Implementation (Setting Up) ● Clear Understanding of strategic goals for ERP ● Commitment by top management ● Cultural & structural changes/ readiness. Implementation ● Excellent project management ● ERP package selection that best fits with current business procedure ● Open information & communication policy ● Exhaustive analysis of current business process ● Importance of data accuracy ● IT leveragability & Knowledge capability ● A great implementation team ● Focuses performance measures. Post-Implementation (Evaluation) ● Post implementation audit ● Documentation and advertising ERP success ● Benchmarking. Source: Motwani & Subramanian, 2005. 2. Implementing ERP can be volatile and hazardous to enterprises internal management. This effect increases the fail possibilities and difficulties. Many ERP system face tremendous because of employee resistance (Stratman & Roth, 1999), according to Aladwani (1998), there are overall three phases for ERP implementations with many critical factors within these phases (Fig. 9). In the 29.

(38) first knowledge formulation phase, enterprises should effectively managing change to identify and evaluate the attitudes of individual users and influential group. Enterprises must minimum the resistance by understand employee-raised facts, beliefs, and values (Hulman, 1979). Second, communication is the important and effective in the strategy implementation phase (Al-Mashari & Zairi, 2000). Enterprises and employee should have knowledge about what the system can deliver to the organization and its worker can build anticipation for the system. Lastly, enterprises must monitor and evaluate change management strategies for ERP implementation to ensure that desired business outcomes were achieved (AlMashari & Zairi, 2000). In conclude, this reviews provided three suggestions for ERP implementation. a. Study the structure and needs of the users and the causes of potential resistance among them. b. Deal with the situation by using the appropriate strategies and techniques in order to introduce ERP successfully. c. Evaluate the status of change management efforts. 30.

(39) Figure 9. Three Phases of Implementation Source: Aladwani, 1998. Table 6. Critical Factors Organized Dimension Technology. Critical Issue. Critical Factor. ➢ Adequacy for specification ➢ User’s maturity for the application of new technology ➢ Evaluation and integration for legacy system. 1. Technical specification 2. Customization 3. User’s maturity against the application of new technology 4. Evaluation and integration of legacy system. 31. Author 1. (LeonardBarton, 1988) 2. (Holland and light, 1999; Somer and Nelson, 2001) 3. (LeonardBarton, 1988; ColeGomoiski,199 8) 4. (Holland and Light, 1999; Sprott, 2000).

(40) Delivery system. ➢ Role of the MIS department in organization ➢ Process adaptation ➢ Harmonious implementatio n ➢ System establishment ➢ Project management ➢ Employee education and training ➢ External partner support ➢ Internal staff involvement. 1. IT infrastructure 2. Enterprise vision and strategic goals 3. Information transparency business process redesign 4. Upper management support for implementatio n 5. Change management 6. project team ability 7. Employee education and training 8. software vendor support 9. The ability of consultancy 10. Participation and coordination of internal staff. 32. 1. (LeonardBarton, 1988; Carton et al., 1994; Broadbent et. al., 1999; Holland and Light, 1999) 2. ( Holland and Light, 1999; Welti, 1999) 3. (Hammer and Champy, 1993; Bingi et al., 1999; Somers and Nelson, 2001) 4. (Bingi et. al., 1999; Laughlin, 1999; Soer and Nelson, 2001) 5. (Appleton, 1997; Laughlin, 1999) 6. (Somer and Nelson, 2001) 7. (LeonardBarton, 1988; Davenport, 1998; Bingi et al., 1999; Somer and Nelson, 2001) 8. (Bingi et al., 1999) 9. (Bingi et al., 1999) 10. (Grover et al., 1995; Bingi et al., 1999; Kumar and Hillegersberg, 2000).

(41) Performance criteria. ➢ Performance evaluation. 1. Impact of performance criteria 2. Assessment of system performance. 1. (LeonardBarton, 1988) 2. (LeonardBarton, 1988). This research aims to discover the appropriate and fundamental processes for organizations, which plan to implement the Big Data system. More importantly, this research explores the potential factors for enterprises during the implementation processes. In the above literature reviews, there are few models provided by IBM and PwC on how to approach the Big Data. Many ERP implementation processes are discussed as well in order to have more concrete understanding of the implementation. Therefore, a research framework can be established based on the literature review above. The research framework is combined and rearranged with few model from above, which included model from IBM Institute for Business Value and Motwani’s re-construct model.. 33.

(42) Chapter 3. Research Method and Design 3.1 Research Framework According to literature review above, the research framework is established in three groups regarding the Big Data and implementation processes. In addition, these groups consist of sub-sections with various detail factors. In the implementation process group, it aims to discover how organizations implement the Big Data system. This group is designed and modified based on Big Data Adoption Pattern provided by IBM Institute of Business Value. This group has three stages, and contains sub-sections with several factors, which are strongly associated. The second group is about the factors for enterprises. The purpose of this group is to find out what factors or requirements are essential for enterprises in order to implement Big Data system. The factors listed in the second group are all standard and frequent factors in ERP implementation literature review. Finally, there are several factors to be examined for the Big Data system. This research will examine the 4V requirements of the Big Data system, which the enterprise implemented. Then, this research will seek for the system advantages that make organizations more competitive than others. Overall, the possible outcome benefits that are delivered and created by the Big Data analytics to the customer will also be discussed.. 34.

(43) Implementation Process 1. Explore & Preparation a. Decision planning b. Driving motivation 2. Engage & Implementation a. Analyze functionality 3. Execute & Maintenance a. User training b. IT support c. Modifying d. Improve & expand. Enterprise Requirements. 1. Pre-Implementation a. Clear understanding of strategic goals b. Commitment by top management c. Culture & Structural change d. Timing 2. Implementation a. Project management b. System selection that best fits current business process c. IT leveragability & Knowledge d. Internal/ External support 3. Post- Implementation a. Employee education and training b. Evaluate & further improvement. 35. Big Data System Performance 1. System Advantages 2. Potential benefits to the customers.

(44) 3.2 Research Method The purpose of this research is to recognize how organization implementing the Big Data system, and factors that organizations are needed to be able to implement. Most of the other studies focus on the Big Data applications, technical difficulties and solutions. This research focuses on the implementation processes and factors. There are not many studies regarding Big Data implementation. Therefore, case study method will be conducted in this research (Yin, 1984). According to Benbasat (1987), case methodology is useful when focus on contemporary events is needed. In addition, research phenomena not supported by a strong theoretical base may suitable for case research. In this research, two case studies will be interviewed and conducted. These case studies are organizations size over 1000 employees. According to Benbasat (1987), a single case is useful at the outset of theory generation and late theory testing. On the other hand, multiple-case designs are desirable when the intent of the research is description, theory building, or theory testing. Therefore, multiple-case method is used for this research.. 3.3 Research Target This research wants to discover the implementing process and factors for the financial institutions, hence the research targets should have at least one Big Data system. Since Big Data systems are fairly sophisticated and expensive. Small financial institutions will have tremendous difficulties and lack of resource to implement the Big Data system. Therefore, the financial institutions should be medium to large in size with at least over 1000 employees. Due to the complication and affection of the Big Data system to the financial institutions and the customers, the institutions should involve variety groups to ensure the implementation. For example, the suitable financial institutions for interview should have at least a retail banking group or 36.

(45) wholesale banking group. Moreover, several divisions are required in order to implement and monitor the Big Data system performance. These divisions in finance institutions are information technology division, process service division, risk management division, finance division, performance management division and corporate planning division. All of the divisions are internally or externally related to the Big Data implementation and its performance to the customers. In general, financial institutions are adequately conservative regarding of strategy and internal information. Therefore, the interview target is extremely difficult to find. However, there will be one organization that is traditional financial institution, and one is not. Finally, all of the personal information regarding of interview target will be strictly confidential.. 3.4 Data Collection Method According to Yin’s identifications (Yin, 1984), there are several sources to support the research finding. This research will focus on case study methods, which also include direct observation, documentation, archival records, and interviews. The goal is to obtain rich amount of data on specific research issue. 3.4.1 Interview Design The interview is designed to ensure the research has enough and qualified data for further interpretation and analysis. Several conditions must be met to ensure the quality of the interview. First, the interviewees must be at least manager ranking of the company, and experienced over 3 years. This condition is to ensure the interviewees have profound knowledge and experienced in the field. Second, the questions will be sent to interviewees before the interview is conducted, so interviewees can be well prepared for the interview. Beside the direct observation technique, the interview will be voice recording to ensure data completeness. Lastly, the interviewer will ask additional questions or customize the existing questionnaire depend on the interviewees situation. 37.

(46) 3.4.2 Question Design The questions design in this research can be categorized into five parts. The first part is designed to discover the whole processes of the implementation. Questions regarding the strategy, concept and plan before the implementation will be ask first. Then, the focus shift to the enterprise itself, and the questions relate to how the enterprises react to the implementation internally. Interviewer asks questions such as does the current employees in the enterprise be able to handle the implementation. In addition, this research wants to find out how each department cooperates internally and externally to adapt with the implementation. Furthermore, this research also wants to find out how top managers shift the decision from traditional to datadriven. The third part of the questionnaire involve with the both technical and business difficulties on implementation. This part of the questions wants to recognize how enterprises overcome the obstacle, and how system compatible with their current strategy. Furthermore, this research would like to find out the result after the implementation. Questions like what benefits the company after implementation? Any new product or service are revealed because of the implementation will be asked on interview. Lastly, this research will examine advantages and key components of the target’s Big Data system.. 38.

(47) Chapter 4. Case Study This research will conduct and review with three case studies, two are financial industries, and one is not. Because of the confidentiality issue, name of the companies in this research will remain anonymous. Each case study will be divided into five parts. The first part will introduce the background of the case study. The second part will discuss the aspects before the implementation and the implementation process. Furthermore, the requirements for the enterprises to implement the Big Data system will be discussed in the third part. In addition, the fourth part will examine the benefit after the implementation. Finally, the advantages of the Big Data system in each case study will be reviewed and examined in the fifth part.. 4.1 Case Study of Company A 4.1.1 Background: Financial Company A Financial company A (Bank A) is established on Aug. 1991 by Mr. Thomas T.L. Wu (Taishin, 2013). The bank has been constantly expanding their business items and operating network, actively exercising the intermediary role for the supply and demand of funds in the society. The headquarter, located in Taipei, has variety divisions in human resource, finance, corporate planning, legal, compliance, information technology, administrative, risk, performance management, and process. In addition, the bank’s major business items include deposit reception, loan extension, export/import foreign exchange, foreign-currency deposits, discount of negotiable instruments, currency conversion, guarantee, surrogate collection/payment, custody, trust, credit card, trading in derivatives, brokerage of short-term bills, securities dealer, certification and underwriting, factoring, securities investment and underwriting, offshore. 39.

(48) banking, issuance of financial bonds, wealth management, and sale of gold bullion, gold and silver coins (Taishin 2013). The Bank A governance chart is below:. Figure 10. Bank A Governance Structure Source: Taishin (2013). 40.

(49) Two interviewees are interviewed in company A, one person is the CRM manager, another is an IT manager of the bank. The interview last for around 40 minutes, and the interview questions are all asked. Additional questions are asked to obtain specific and thorough information. Several questions are answered from both CRM and IT perspectives. Therefore, CRM manager usually answered non-technical questions, and IT manager filled in with information from technological perspectives. Overall, Company A provided splendid and valuable insight in Big Data implementation. 4.1.2 Implementation and Process Before the interview began, CRM manager first defined their Big Data system. To company A, Big Data is definitely not a new concept. Their Big Data system is the data warehouse (DW) they had built on 1998. According to article “Data Warehouse process management” (Vassiliadis, Panos, et al., 2001), Data warehouses (DW) integrate data from multiple heterogeneous information sources and transform them into a multidimensional representation for decision support applications. Company A has tremendous number of client account and information. Therefore, they must develop their own data warehouse in order to cluster or even analyze them together. They spent 16 years to build and advance their own DW. The IT manager later emphasized why they must develop this DW. He said when they encountered the three difficulties, the size of the data, accuracy of the data, and the velocity of data, they knew that they must develop a DW in order to solve these problems. According to the IT manager, Company A did not ask any third company for a total solution on implementing the Big Data system. Company A refused to reveals, which brand of hardware or software they used. In addition, IT manager thinks it’s more important to focus on the framework of the DW, instead of the brand. According to him, DW has 3-Tier of data 41.

(50) warehouse architecture, which is bottom tier, middle tier, and top tier (Fig. 11). In the bottom tier, data warehouse server fetch only relevant information based on data mining. For example, customer profile information provided by external consultants. Moreover, data is feed into bottom tier by some backend tools and utilities. The backend tools provided functions such as data extraction, data cleaning and data transformation. The middle tier represents a multidimensional data from data warehouse or data mart. It is usually implemented in two models, the ROLAP Model and MOLAP Model. The ROLAP model presents data in relational tables. However, MOLAP model present data in array based structures, which means map directly to data cube array structure. Lastly, the top tier usually presents as front-end tools. These front-end tools included, but not limited to reporting tools, analysis tools and data mining tools. Therefore, the IT manager mentioned that it is important to understand the DW framework, and put less attention on the brand. Also, he mentioned that their DW is well completed all tiers with consulting help from IBM and DataAge.. 42.

(51) Figure 11. A Three- Tier Data Warehousing Architecture Source: California State University Northridge Company A had a project team ready before and during the Big Data implementation. This project team is formed with around 20 people, and included IT engineers and front-end business consultants. Also, the team has more front-end business consultants than IT engineers. CRM manager clarify that business consultants understand more about the demands or needs. At the beginning of the process, IT engineers and business consultants will discuss the criteria for the implementation. They will discuss what are the needs and the requirements for the customers and clients. Later, IT engineers will create a standard development process, which fulfill the needs of front-end business consultant and the implementation.. 43.

(52) 4.1.3 Enterprise Requirements In this section, the interview questions are focused on how Company A reacts before and after the implementation. In addition, this research also interested in what difficulties Company A has and how they overcome those difficulties. According to CRM manager, top management directors and CEO all have very clear objective and consent before Big Data implementation. The goal is to understand their customer and clients more efficiently and effectively. Also, Company A wants to target more accurately on their customers. Therefore, they have prepared for the Big Data implementation around 8 years ago when they realized the increase amount of the data and the value of analyzing them. As for now, they have around 100 people across the business division to handle data analytics. At the beginning, these people were chosen from the internal business division. Most of them are nontechnical PM and business consultants. CRM manager also emphasized why data analytics does not belong to IT division. IT divisions focus more on troubleshooting the hardware and DW maintenance. In addition, they must follow the Software Engineering Standards. Therefore, if Company A shifted the data analytics to them, the whole process will be inefficient. There are two ways how Company A manages to find the right person for data analytics. First, Company A picks exceptional business consultants and non-technical PM internally. Most of them usually have tremendous interests in data analytics. Therefore, they usually need to learn the fundamental data analytic tools such as R programming and statistics. Moreover, Company A provided incentive for employees who are willing to learn data analytics. Incentives included but not limited to promotion and salary raise. Second, industry-university cooperative research project is also another way for Company A finding the right person for data analytics. Company. 44.