Running head: BUSINESS INTELLIGENCE SOFTWARE
Business Intelligence Software
Customers’ Understanding, Expectations and Needs
Adis Sabanovic
Thesis for the Master’s degree in Business Administration, Spring 2008
Executive summary
Modern companies operate in incredibly complex and dynamic environments.
This is clearly characterized by constant changes in technology and in various market forces as well as by enormous amounts of data and information that need to be gathered and analyzed every day. Governmental regulations and ongoing competitor pressures, among other external and internal factors, are issues that managers and decision makers in a company must take into a consideration when making decisions. The need for BI systems is growing stronger and businesses in various industries demand such tools that will help them stay on the edge in order to be competitive. Hence the purpose of this paper is to find out what their companies desire when choosing a BI system to work with. What are their needs and what do they expect and understand from this technological system that will hopefully make them work easier and gain their knowledge about the business they operate in.
A web questionnaire is aimed at 67 Swedish companies from various industries and the answers have been summarized and analyzed in different cross tables for comparison reasons. Respondents from the Manufacturing industry were those with the highest response rate. A model called The PET-model of BI implementation was created, as a result of the theoretical findings, and this model is used to finalize the results and the conclusions of this paper.
Key words: BI, Business Intelligence, Business Intelligence Software, Competitive Intelligence, Decision Support Systems
TABLE OF CONTENTS
1 INTRODUCTION ...5
1.1 BACKGROUND... 5
1.2 THE DEFINITION OF BUSINESS INTELLIGENCE... 6
1.2.1 BI wide‐ranging ... 7
1.2.2 Business Intelligence Software ... 8
1.2.3 Why BI software ... 8
1.2.4 Some categories of BI Tools... 9
1.2.5 BI software in organizations... 10
1.2.6 Expectations and needs of a BI software.... 11
1.2.7 Common issues regarding BI usage... 12
1.2.8 Market of BI solutions today ... 13
1.3 DEFINITIONS... 15
1.4 LIMITATIONS... 17
2 METHOD...18
2.1 CHOICE OF METHODOLOGY... 18
2.2 RESEARCH PHILOSOPHY... 18
2.3 CHOICE OF THEORY... 19
3 THEORETICAL FRAMEWORK...21
3.1 BI SOFTWARE CLASSIFICATION... 21
3.1.1 End‐user query, reporting, and analysis ... 21
3.1.2 Advanced analytics... 22
3.2 ANALYTICAL APPLICATIONS... 23
3.2.1 Logical integration... 24
3.2.2 Interactive reports ... 25
3.2.3 Integrated information... 25
3.2.4 Addressing of a Business domain ... 25
3.3 TYPES OF BUSINESS INTELLIGENCE SYSTEMS... 26
3.3.1 Model‐driven BI system ... 26
3.3.2 Data‐driven system ... 26
3.3.3 Communication‐driven system ... 26
3.3.4 Document‐driven system... 27
3.3.5 Knowledge‐driven system... 27
3.3.6 Web‐based system ... 27
3.4 REAL‐TIME BI SYSTEM... 27
3.5 HOW REAL‐TIME BI SYSTEM WORKS... 28
3.5.1 Time importance when working with BI... 28
3.6 THE DIFFERENT USER‐GROUPS OF BI ... 31
3.7 BI PLACEMENT IN THE ORGANIZATION... 32
3.7.1 The special dept. model of intelligence ... 32
3.7.2 The advisory model of intelligence ... 32
3.7.3 The professional model of intelligence ... 33
3.7.4 The top‐down model of intelligence ... 34
3.7.5 The Integrated Intelligence Model... 34
3.7.6 The down‐up model of intelligence ... 35
3.7.7 The departmental model of intelligence... 35
3.8 SOME BI TOOLS ON THE MARKET TODAY... 36
3.9 SUBSOFT – BRIEF PRESENTATION... 39
3.10 THEORY MODEL CREATION... 41
4 EMPIRICAL METHOD ... 45
4.1 RESEARCH STRATEGY... 45
4.2 TIME HORIZON... 45
4.3 DATA COLLECTION METHOD... 46
4.4 POPULATION... 47
4.5 SAMPLE SELECTION... 48
4.6 RESEARCH CONDUCTION... 48
4.7 BI RESEARCH PLAN... 49
4.8 DATA ANALYSIS... 53
4.9 RESEARCH QUESTIONS... 53
4.10 RELIABILITY AND VALIDITY... 54
5 ANALYSIS ... 55
5.1 EMPIRICAL FINDINGS... 55
5.2 CRITIQUE... 62
5.3 ANALYSIS CONCLUSIONS... 63
5.4 SUBSOFT ‐ COMPARED TO THE RESEARCH FINDINGS... 65
6 THESIS CONCLUSION ... 67
6.1 PRACTICAL RELEVANCE... 67
6.2 DISCUSSION... 67
LIST OF REFERENCES ... 69
APPENDICES ... 72
APPENDIX 1, EMAIL (SWEDISH) ... 72
APPENDIX 2, EMAIL (ENGLISH)... 73
APPENDIX 3, QUESTIONNAIRE RESULTS... 74
APPENDIX 4, QUESTIONNAIRE... 82
APPENDIX 5, INDUSTRY ‐ ANSWERS... 92
LIST OF TABLES
TABLE 1-WORLDWIDE BUSINESS INTELLIGENCE TOOLS REVENUE BY SEGMENT,2004–2006...14
TABLE 2–PETMODEL’S PURCHASE FOUNDATIONS IN THE MANUFACTURING INDUSTRY...56
TABLE 3-PETMODEL’S PURCHASE FOUNDATIONS IN ALL INDUSTRIES COMBINED...59
TABLE 4-PETMODEL’S EMPLOYMENT FOUNDATIONS IN ALL INDUSTRIES COMBINED...59
TABLE 5-PETMODEL’S TASK FOUNDATIONS IN ALL INDUSTRIES COMBINED...61
LIST OF FIGURES FIGURE 1-WORLDWIDE BUSINESS INTELLIGENCE TOOLS REVENUE SHARE BY REGION,2006...15
FIGURE 2–CLASSIFICATIONS OF BI SOFTWARE...1
FIGURE 3-REAL-TIME BIPROCESSING COMPONENTS...1
FIGURE 4–LATENCY IN BUSINESS INTELLIGENCE DECISION MAKING (HACKERTHORN,2003)...1
FIGURE 5–REAL-TIME BI;ACTION TIME VS.IT COSTS (WHITE,2003) ...30
FIGURE 6–DIFFERENT BI USER NEEDS IN THE HIERARCHY (SOLBERG SØILEN,2008) ...1
FIGURE 7–THE SPECIAL DEPARTMENT MODEL OF INTELLIGENCE (SOLBERG SØILEN,2008) ...1
FIGURE 8–THE ADVISORY MODEL OF INTELLIGENCE (SOLBERG SØILEN,2008)...1
FIGURE 9–THE PROFESSIONAL MODEL OF INTELLIGENCE (SOLBERG SØILEN,2008)...1
FIGURE 10–THE TOP-DOWN MODEL OF INTELLIGENCE (SOLBERG SØILEN,2008)...1
FIGURE 11–THE INTEGRATED INTELLIGENCE MODEL (SOLBERG SØILEN,2008)...1
FIGURE 12–THE DOWN-UP MODEL OF INTELLIGENCE (SOLBERG SØILEN,2008)...1
FIGURE 13–THE DEPARTMENTAL MODEL OF INTELLIGENCE (SOLBERG SØILEN,2008) ...1
FIGURE 14–SUBSOFT MODEL OF INTERNAL AND EXTERNAL FACTORS (SUBSOFT,2008)...1
FIGURE 15–THE PET MODEL OF BI IMPLEMENTATION...1
FIGURE 16–BI RESEARCH PLAN...52
FIGURE 17–SURVEY RESPONDENTS REPRESENTED FROM DIFFERENT INDUSTRIES. ...55
FIGURE 18–PETMODEL AFTER THE ANALYSIS...1
1 Introduction
This Chapter describes the definition of Business Intelligence, why we choose to use it and also anticipated problems when using it as well as the purpose of this thesis. For the start, a short historical background about the term Business Intelligence is presented.
1.1 Background
In the modern world of today the access to information is greater than ever before.
In many cases the information flow is overwhelming and it sometimes leads to valuable information losses. Company leaders and other decision makers are trying to overcome this problem by investing in various sophisticated computerized solutions, also known as Business Intelligence Systems. But it is not only in the modern world that Business Intelligence Systems have been appreciated for their great capabilities of creating a better understanding of one’s working environment.
In 1958, the term Business Intelligence is used for the first time in an article called A business intelligence system by Hans Peter Luhn. Luhn was describing how to automate the process of collecting and sorting information from documents using current photo-printing technology. He was saving information on magnetic tapes and driving it through a process of auto encoding and auto abstracting programs to later sort it in different pattern storages. Processed information would then be put into a comparison area and sorted into three main categories: who needs to
know, who knows what, and what is known (Luhn, 1958). Already in 1958, Luhn had discovered the importance of information processing and that all greater information flows contains even greater value for the one who has the ability to turn it into knowledge. However, in his article Luhn admits that the type of equipment used for processing information, in late 1950’s, was in early stage of development and that a great deal of research has yet to be done to perfect the information processing technique.
Ever since Luhn introduced us to Business Intelligence terminology, the importance of knowing how to turn information into knowledge has grown tremendously, especially among today’s modern business leaders and other decision makers around the world.
1.2 The definition of Business Intelligence
Today, after many facelifts and makeovers of BI there are quite many definitions.
In many cases the same definition will be used for other terms such as;
Competitive Intelligence (CI) or Decision Support Systems (DSS).
A more recent definition of the term was coined by The Data Warehousing Institute (TDWI), a provider of education and training in the data warehousing and BI industry; and is as follows(Loshin, 2003, p. 6):
The process, technologies, and tools needed to turn data into information, information into knowledge, and knowledge into plans that drive profitable business action. Business intelligence encompasses data warehousing, business analytic tools, and content/knowledge management.
The characterization of the term Business Intelligence or BI, as it will be referred to throughout this paper, is basically still the same as in the early 1960’s but the significance and understanding of BI has changed as technology has improved, organizations have decentralized and the complexity levels of new information have increased.
1.2.1 BI wide-ranging
The popularity of the term “Business Intelligence” has grown rapidly in the last decade. As mentioned earlier the definition of a BI software is yet somewhat open-ended and may differ from author to author. BI gives the impression of being a multifaceted term that can refer to processes, techniques or tools to support the making of faster and better decisions (Pirttimäki & Hannula, 2003). Expectations of what a BI software is supposed to perform, or accomplish, is even more differently understood by the users. In many cases, corporations are already using some kind of BI tools or solutions but have chosen to call them differently, e.g.
Management Information Systems (MIS), Decision Support Systems (DSS), Executive Information Systems (EIS), et cetera. (Pagels-Fick, 2000) It is also common that companies, unknowingly, use small parts of a complete BI system, e.g. CRM- Customer Relation Management (CRM) and Knowledge Management which focuses exclusively on customers and knowledge while a complete BI system primarily deals with information (Solberg Søilen, 2005).
1.2.2 Business Intelligence Software
Business Intelligence (BI) software is used as an effective reporting and analyzing tool to better understand a company’s organizational surrounding and environment and it gives managers basic data for decision. There are some main and very basic objectives that a BI tool must accomplish. These are to generate better information than your rivals do, to analyze that information and make sound choices, to make those choices quickly and to convert strategic choices into decisive actions (Vine, 2000).
1.2.3 Why BI software
There are many reasons for why a company should use business intelligence or decision support systems. Eckerson (2004) has, in his research, found that BI systems do not only help decision makers to make better and more efficient decision but that BI also helps the entire organization to improve Return on Investment (ROI) profitability, gain customer/supplier, as well as employee, satisfaction, et cetera. He also points out that if one BI system is implemented throughout the entire company, there is a single version of truth which helps the company to avoid misunderstandings and gets everyone going in the same direction (Eckerson, 2004).
Loshin (2003) points out how Customer Relationship Management is improved and how certain risks are decreased by analyzing supplier/consumer activity and reliability, providing insight into how to rationalize the supply chain. BI can also help the companies to evaluate organizational costs and to improve logistics management, lowering the operational costs and decreasing the investments
required to make sales. Another areas of usage for BI is evaluation of customer lifetime value, short term-profitability expectations and using this knowledge to distinguish between profitable and non-profitable customers to increase profitability (Loshin, 2003).
1.2.4 Some categories of BI Tools
Most companies today use a set of different BI tools, instead of focusing on only one. The reason for that may be simple; different users prefer different types of BI tools. The tools may differ in reporting, ad hoc queries, OLAP, et cetera. BI tool vendors are doing their best to meet all those requirements allowing organizations to standardize on using one single tool and on one single vendor (DM Review and SourceMedia, Inc., 2005). Below, a list of some major categories of BI tools is presented:
• Production Reporting Tools: Used by professional developers to create standard reports for groups, departments or the enterprise.
• End-User Query and Reporting Tools: Used by end users to create reports for themselves or others and require no programming.
• OLAP Tools: Enable end users to "slice and dice" data dimensionally to explore data from different perspectives and time periods.
• Dashboard/Scorecard Tools: Enable end users to view critical performance data at a glance using graphical icons and drill down to analyze detailed data and reports if desired.
• Data Mining Tools: Enable statisticians or business analysts to create statistical models of business activity.
• Planning and Modeling Tools: Enable analysts and end-users to create business plans and simulations against BI data. Planning tools supply dashboards and scorecards with targets and thresholds for metrics.
(DM Review and SourceMedia, Inc., 2005)
1.2.5 BI software in organizations
When a company’s business information is isolated in different BI tools the information risks to disappear and to never be used again. Many companies are therefore trying to tie the information together to create one overall strategy (Rådmark, 2007). Different suppliers of BI solutions are offering a too wide range of products as decision makers are only requiring one product that will give them a better overall picture of the company’s activities and the surrounding environment.
In the early days, BI software’s focus was on the technical solutions and on the business analysis process that would provide the decision makers with information needed. Nowadays a BI-software must focus on making the information available for more people (workers) in the organization and making it more usable (Rådmark, 2007). Most companies are using different systems that control the information torrents, but only those who use BI can exploit the crucial information from different sources and decide what information to use. The leaders or decision makers are more interested in that specific information than in what technology is used exploiting it. It is about the management information rather than technology, because when the technical side is in focus the attention is
rather on the applications instead of creating what is best for decision making through small and complex BI solutions.(Lindström, 2007)
Traditionally, BI takes place high up in the organizations’ hierarchy but in today’s organizations there is a strong demand for BI solutions that gives all decision makers access to relevant information, regardless of the level in the organization.
One problem, Rådmark points out, is that if there are several solutions in the organization, there is a lack in the common strategy and responsibility distribution. Since BI, in its best form, should cut through the whole organization, or the bigger parts of it, it is not possible to place responsibility on one certain function. Therefore, the problem that many organizations face today is that BI tools are requiring a change in the organizational structure to create the best possible environment to not isolate vital business information but rather to spread and distribute it throughout the whole organization (Rådmark, 2007). The possibility of bringing fast information and making in transparent is very important. It is not only economically effective but also a competitive advantage to be able to analyze information faster and more effectively than your competitors (Lindström, 2007).
1.2.6 Expectations and needs of a BI software
A research conducted by BetterManagement (division of a SAS institute Inc.
which does researches about business management issues around the world) in 2005 showed that only nine percent of BI software users were always provided with all the necessary information from the BI software to make effective business decisions and that only 45 percent of the users did sometimes get all the
information they needed (Miller, Bräutigam, & Gerlach, 2006). These numbers indicate that many corporate leaders have high expectations on a BI software before purchasing it but in very few cases the decision makers will actually completely rely on the information extracted from the software. What was instead demanded, or needed, by the companies, according to the survey, were the following statements:
1. Improved quality of information available to them.
2. Access to relevant information in easy to use reporting interfaces for ad hoc reporting.
3. Assistance with interpreting and drawing conclusions from the information.
4. Access to relevant information in standard reports.
5. An overview of which data is available for analysis.
6. A formal assessment of their information needs.
7. Training on how to use BI tools.
(Miller, Bräutigam, & Gerlach, 2006)
1.2.7 Common issues regarding BI usage
Companies that have started data warehousing projects or have purchased large- scale data mining software suites often have very high expectations but also many disappointments related to failure in the way that data is conceived, designed, architected, managed and implemented. The vague understanding of what BI methods and products can do frequently results in a lack of a proper value proposition on behalf to the business sponsor (Loshin, 2003). Also the scope of
the project is not always fully understood which causes delays in delivery to the decision maker. Another issue that companies face when using BI is insufficient technical training of the users. This prevents company’s developers and analysts from using software products to the full capacity and from doing what the vendors claim they do. Poor understanding of technology infrastructure also leads to disadvantages such as poor planning and scheduling which often leads to lack of trustworthiness in the results due to poor data quality. Some BI software users also lacks a clear statement of success criteria, along with a lack of ways to measure program success and this is inevitably leading to a perception of failure (Loshin, 2003).
1.2.8 Market of BI solutions today
According to a report from Datamonitor (leading provider of online database and analysis services for key industry sectors) the market for business analysis is increasing tremendously fast. The report shows that the value of the BI market will increase from four billion dollars in 2006 to an estimated eight billion in 2012. This means an annual raise is about 12,5%. The battle between small independent BI suppliers is losing attention while the focus is now on the big giants that are constantly buying smaller BI suppliers (Wallström, 2007).
According to Gartner, the world’s leading information technology research and advisory company, the market of BI solutions is basically shared between three mega-suppliers, Oracle, SAP and Microsoft who together own about 20% of the global market.
Fusions between these mega suppliers and smaller ones are occurring constantly.
Today’s BI tools have a broader area of usage and BI is not only about reports and analytics but also about real time dashboards and scorecards, predictive models, workflows, visualizations and searches. Basically, the market for BI tools consists of two segments, Query, reporting and analysis (QRA) and advanced analytics (Adv. An). (A more detailed explanation of the two segments will follow in the theory chapter). In 2006, as shown in Table 1, the BI tools market grew 11.5%
and reached $6.25 billion in worldwide license and maintenance revenue. During that time there was no significant consolidations in the BI tools market (Vesset &
McDonough, 2007). A huge number of mergers and acquisitions occurred between larger BI tools vendors and smaller software vendors.
Table 1 - Worldwide Business Intelligence Tools Revenue by Segment, 2004–2006
Revenue ($M) Share (%) Growth (%)
2004 2005 2006 2004 2005 2006 ’04-‘05 ‘05-‘06
QRA 4,004.9 4,487.6 5,008.5 79.5 80.0 80.1 12.1 11.6
Adv. An. 1,031.9 1,118.6 1,244.6 20.5 20.0 19.9 8.4 11.3
Total 5,036.7 5,606.2 6,253.0 100.0 100.0 100.0 11.3 11.5
(Vesset & McDonough, 2007)
In Figure 1 the geographic allotment of the BI tools market is shown. The Americas region has the largest segment of the market, followed by Europe, the Middle East, and Africa (EMEA) and Asia/Pacific.
Figure 1 - Worldwide Business Intelligence Tools Revenue Share by Region, 2006
(Vesset & McDonough, 2007) (IDC, June 2007) 1.3 Definitions
Business Intelligence Software, -System, -Application – program that makes decision making more efficient and easier through different processes like information gathering, analysis, spreading of information and communication within a company.
Business Process – a complete series of activities in a company or an authority.
Dashboard/Scorecard - a dashboard or scorecard is a graphical display that compares performance against predefined goals. A dashboard records actual
performance or behavior, like an automobile dashboard, while a scorecard measures that performance against objectives or goals. A dashboard tells how you are doing, a scorecard, how well. (Eckerson, 2005)
Embedded Business Intelligence – a business analysis that is built into different business programs and that does not exist as separate program.
Neural Network – an interconnected group of artificial neurons that uses a mathematical or computational model for information processing
OLAP – On-line Analytical Processing – a technique for searching gathered data from databases while they are online. OLAP is used for sales analysis and decision making. OLAP can be used as an alternative to data warehousing and data marts.
Portal - a web system that provides the functions and features to authenticate and identify the users and provide them with an easy, intuitive, personalized and user- customizable web-interface for facilitating access to information and services that are of primary relevance and interests to the users.
Real-Time BI - an organization’s ability to react to business needs and changing business circumstances within a single day. (White, 2003)
TCP/IP - the Internet protocol suite (commonly TCP/IP) is a set of communications protocols on which the Internet and most commercial networks run.
SOA - Service Oriented Architecture is a search engine technology and the main integration component in an information system
Open Hub services - service used in analytical applications to distribute specific data
1.4 Limitations
Due to the time that was provided for writing this thesis (6 weeks), working conditions were a little bit tough. Hence it shall be acknowledged that time was a limit.
When contacting respondents for the survey, besides time, money was also a limit. More efficient ways could have been used when collecting respondents if the right amount of money was invested into certain databases on the Internet.
In this thesis a model was created as a result of existing theories. The model is called The PET model of BI implementation and consists of nine different foundations divided into three layers. Better research conditions might reveal other interesting facts that can change the appearance of the model, improve it or in worst case scenario completely reject it.
Not all questions that were put in the questionnaire were analyzed either. In the creation process of the questionnaire, some questions that could be related to the theory provided in this thesis were not asked. For example in the theory chapter, Real-Time BI system is described. But due to the limitations of time no analysis was made upon this subject.
2 Method
In this chapter a presentation of how the research for this thesis has been conducted is given. It will give a better understanding of how the theoretical and empirical reasoning has contributed to the purpose. It will also describe the different methods used in this study.
2.1 Choice of methodology
The aim of this paper is to present a basis upon which the reader can gain an understanding of how companies in various industries in Sweden relate to BI.
Hence the central point of this thesis is to provide an argument for and analysis of what is expected from a complete BI Software Solution.
Companies’ relation to BI Software Solutions will be measured and mean values will be calculated. The results will be presented in cross tables as well as in a model, which is based on the existing theories, to illustrate an overall picture of the companies’ relation to BI. Flowcharts and diagrams are also used to present the results. Hence a deductive approach, discussed more in the research philosophy part is applied.
2.2 Research philosophy
The research problem of this thesis is built on existing theories, what means that the research approach is of a deductive nature. The opposite, an inductive
approach is not suitable for this research since the research is based on already existing theories.
When working with an observable social reality and when the result of a research can be made to draw law-like generalizations, a research is produced with quantitative and deductive nature with a positivistic approach (Saunders, Lewis, &
Thornhill, 2007). It is also likely that the existing theory about BI in this paper will be further developed and tested by further research. According to Saunders et al. (2007) this is also an indication that the research philosophy in this thesis is of a positivistic nature. The thesis aims to observe and study the companies’ points of view and their relation to BI. For that purpose a questionnaire is applied. A questionnaire is a kind of study that fits with a positivistic research approach and from the questionnaire the quantifiable data can be examined and analyzed.
2.3 Choice of theory
Books that were used for this thesis all come from the library of Kristianstad University. Most of the electronic articles were either downloaded from various journals on the Internet or other online databases such as Emerald and The Data Warehouse Institute or from the Kristianstad University’s First Class Email/Course client. There are many high-tech explanations of BI found on the Internet, in various books and in articles but the first thought when collecting information about BI and writing about it, for this thesis, was to build a appropriate and relevant theoretical ground to present an introduction of the subject on a very low technical level so that the readers will easily understand
what BI is and how it works. The idea was also to maintain the “easy-to- understand” level throughout the thesis. Although in some parts it is essential to use complex terms and idioms necessary for the explanation of a matter.
3 Theoretical framework
This chapter will provide the reader with information about how BI software functions and how we classify the different BI analytics tools available on the market today and what such applications consist of. In this chapter theories about the organizational structure and the placement of BI inside the organizations will also be discussed.
3.1 BI software classification
BI tools are a part of the broader market called business analytics, which is illustrated in Figure 2. The market for BI tools includes both standalone packaged software and embedded BI tools provided by database management software vendors (Vesset & McDonough, 2007). The BI tools market itself is divided into two market segments, Query reporting, analysis and Advanced analytics, and these are the two areas of BI tool applications that this thesis is focusing on. In Figure 2, these areas are the two dash-boarded rectangles.
3.1.1 End-user query, reporting, and analysis
Query, Reporting, Analysis (QRA) software includes ad hoc query and multidimensional analysis tools as well as dashboards, scorecards and production reporting tools. These tools are designed specifically to support ad hoc data access and for report building by either IT or business users and do not include any other applications or tools that may be used for report building what so ever
(Vesset & McDonough, 2007). Yet they are justified as multidimensional analysis tools that include both online analytical processing (OLAP) servers and client- side analysis tools that provide a data management environment that is used for modeling business problems and analyzing business data. Packaged data marts are also included in this function. These data marts are preconfigured software used for combining data transformation, management, and access in one single package and are usually presenting the results in various business models (Vesset
& McDonough, 2007).
3.1.2 Advanced analytics
The main occupation of advanced analytics software is data mining and statistics.
Technologies that are used are neural networks, rule induction, and clustering, among others, in order to discover relationships in data and then make hidden, not apparent or complex predictions for reporting and multidimensional analysis (Vesset & McDonough, 2007). In this sector there are technical, econometrical and other mathematical operations that provides libraries with statistical algorithms so that the data can be processed and analyzed. Most common functions are frequencies, cross-tabulations and chi square but there can also be some other specialized and sophisticated functions focusing on the functional area such as the industrial design, clinical trial testing, exploratory data analysis, and high-volume and real-time statistical analysis (Vesset & McDonough, 2007).
(Vesset & McDonough, 2007) 3.2 Analytical applications
An analytical application, just as Business Intelligence itself, is very difficult to define and many professional programmers and users of BI tools will have their own definition when explains the tool, the technology or the architecture. In this thesis though, the author has found one definition is used that will hopefully satisfy most of the analytical application industry’s “pundits”: (Eckerson, 2005, p.
5):
An analytic application consists of a series of logically integrated, interactive reports, including dashboards and scorecards, that enable a wide range of users to access, analyze, and act on integrated information in the context of the business processes and tasks that they manage in a given domain, such as sales, service, or operations.
Business analytics software
Performance Management tools and applications
Financial Performance and strategy mgmt.
applications (budgeting, planning, consolidation, profitability,
mgmt./ABC, scorecards)
CRM analytical applications
(sales, customer service, contact center, marketing,
website analytics, price optimization)
BI tools:
Supply chain and service operations analytic
applications
Workforce analytic application
Analytic spatial information management tools
Data warehouse platform (data warehouse mgmt. and generation)
Query reporting, analysis (includes dashboards)
Advanced analytics (includes data mining and
statistics)
Figure 2 – Classifications of BI software
Generally an analytical application consists of elements whose purpose is to build up a business logic that will take the user through a series of interactive reports where it will be possible to access, analyze, and take necessary action to optimize the activities in a specific business domain. Analytical applications are, therefore, not about randomly created reports that a user can upload from an “inbox” or from a “my reports” folder, but rather about the interactive and dynamic play where the user is given the possibility to utilize something that is highly valuable for his/her company’s endurance (Eckerson, 2005).
3.2.1 Logical integration
The first part of a BI analytical application is called logical integration and is about stepping the user through different series of interactive reports and views of dimensional data, which will lead to the important point of action or to the request for more information. Different users have different knowledge or know-how when it comes to usage of analytical applications so therefore the navigational logic is important when a user wants to navigate through different reports on the
“reports page” to effectively analyze data and make decisions. Interactive dashboards and scorecards are used to inform the user what metrics or data to examine so another important logic of a BI analytical tool is also offering of recommendations (Eckerson, 2005). This is about giving the user, novice or professional, the best possible overview of the data and to make sure that important information is not missed or neglected.
3.2.2 Interactive reports
The interactive reports key is about giving the user opportunity to interactively search through the reports for additional information by simply “drilling” from a top-level view to a lower level. Reports should be unfixed and possible to change into tables, charts, or other transactional data. Some technologies worth mentioning that are used for delivering such interactive reports are OLAP cubes, parameterized reports, linked static reports, advanced visualization techniques, dashboard/scorecards, numeric searches, et cetera (Eckerson, 2005).
3.2.3 Integrated information
Various data and information from different sources should be put in analytical applications and then stored in one single warehouse where all data is processed and analyzed once again. Large companies, like Continental Airlines, have many different analytical applications running against one single enterprise data warehouse where all data, for example tracking flight process, fraud detection, or revenues management, are put through one large analytical procedure. Integrating the information will help managers to avoid problems when seeking one consistent version of the enterprise information (Eckerson, 2005).
3.2.4 Addressing of a Business domain
Different business areas (domains) such as sales, service, or manufacturing, have different information requirements and analytical applications are defined by those specific requirements. A sales analytical application may monitor a production line performance or other sales representatives and regions or it can
examine the sales and contact history, et cetera. It is the interconnection of these domains that must be used and placed within a logical model because several business areas actually represent the same company (Eckerson, 2005).
3.3 Types of Business Intelligence Systems
3.3.1 Model-driven BI system
In a model-driven BI system, the information / intelligence is mostly presented thorough a series of different models. The user can access and modify financial, optimization and/or simulation models of various kinds (Hedgebeth, 2007). The most basic function of the model-driven BI system is the provision of quantitative models.
3.3.2 Data-driven system
In data-driven systems the most basic functional level occupies search tools that access simple file systems (Hedgebeth, 2007). Here the user has access to and can modify real-time internal and external data.
3.3.3 Communication-driven system
In communication-driven systems, different networking technologies drive decision based collaboration activities. Examples of these are video conferencing, groupware and computer bulletin board systems (BBS) (Hedgebeth, 2007).
3.3.4 Document-driven system
Via computer storage and processing, a document-driven retrieval is made. Here, via a search engine, the user may access documents, policies, images, sound, scanned documents et cetera. (Hedgebeth, 2007).
3.3.5 Knowledge-driven system
In knowledge-driven systems, trained and professional users with knowledge are used to solve various problems.
3.3.6 Web-based system
Intelligence from a web-based system is presented via a web browser and TCP/IP (Internet protocol suite) (Hedgebeth, 2007).
3.4 Real-Time BI system
Another BI system that is not mentioned under the previous heading but nonetheless deserves special attention is called The Real-Time Business intelligence system. This system is about organization’s ability to react in time and become more alert and more responsive to various changing business conditions (White, 2003). In order to make effective decisions, accurate business intelligence is required. The problem with accurate intelligence is that it always takes time to collect and deliver it to the right users and it also takes time for the users to act on this information. As shown in Figure 4 the delay between a business event occurring, and action being taken, is an issue of decisive importance when the
value of the information is to be determined. The technology used to deploy a Real-Time BI application must, first of all, aim to reduce a user’s reaction time if the information value is to be as high as possible (White, 2003).
3.5 How Real-Time BI system works
Basically a Real-Time BI system consists of two operational components (Figure 3). One is for data-integration and the other one is for decision-making. The data integration component captures business events from operational systems and then integrates them into the low-latency store. The decision making component, on the other hand, supports real-time performance management and other significant real-time analysis and reports (White, 2003).
3.5.1 Time importance when working with BI
As illustrated in Figure 4, a business event’s road to become an action consists of three latency periods, data latency, analysis latency, and decision latency (Hackerthorn, 2003). The result of the three latencies is called action time or action distance and the central objective of a real-time BI system is to reduce the
Operational applications
Real-time data integration component Real-time decision
making component
Operational data
Events
Low latency store Reports, Alerts
& Messages
Figure 3 - Real-Time BI Processing components
action time as much as possible to respond to a business happening. If the problem is in data latency or the analysis latency the time gap can be reduced by improving the technology used. On the other hand if the problem is decision latency, then the latency depends on the user. Therefore, the information that is provided to the user must be improved to solve the decision latency problem.
Another solution could also be an automatization of some BI processes that will automatically take action on behalf of the user (White, 2003). Hackerthorn (2003) describes how decision latency may be reduced by applying three requirements to the system; alerting, information, and guidance. He finds that the system should be configured in a way which alerts the user if some unusual business situation occurs. Secondly the system should be able to show situational-specific business information so that the user quickly gets an understanding of the business environment he is working in. Thirdly, the user should be guided by the system that suggests the most suitable action for the specific situation.
Action taken Information
delivered Data stored
Business event
Value
Time
Action time or action distance
Figure 4 – Latency in Business Intelligence decision making (Hackerthorn, 2003)
Another important success aspect in realizing the benefits when working with Real-Time BI is recognizing that the Return on Investment (ROI) depends on two factors, the time that it takes to reduce an action and the organizations ability to modify its business practice. Figure 5 illustrates that there is a point (exploration threshold) beyond which reducing the action time any further has no value to the business. The smaller the action time required, the bigger the Information Technology (IT) costs are (White, 2003). Figure 5 combined with Figure 4 show us that a shorter action time gives higher value to the intelligence but it also increases the costs for the investment in required Information Technology. First after a certain time (at the break-even threshold) the costs for the Information Technology will become so low that ROI becomes positive.
Figure 5 – Real-Time BI; Action time vs. IT costs (White, 2003)
$
Action time
Incremental IT costs
Business benefits Break-even
threshold Exploration
threshold
3.6 The different user-groups of BI
BI is a relatively new organizational function, and most companies will have no or little experience in its implementation.
(Solberg Søilen, 2008)
Different users necessitate different intelligence and a BI tool’s main priority should be to provide the right user with the right intelligence. In Figure 6, on the bottom axis, different user groups can be identified with the specific intelligence presentation requirements (on the vertical axis). Executives tend to have little or no time to read long reports and are therefore only interested in fast figures or the
“executive summaries”. These can be presented in Scorecards or Dashboards shown as Key Performance Indicators (KPI). Analyst or Senior managers on the other hand like to work with advanced online analytical processes and explore different way of making analysis. Written reports are in the interest of department managers. These are interested in reading and analyzing compiled text reports such as, sales analysis, budgets et cetera. that will give them a good basis for making correct decisions. Workers on lower levels in the organization work with invoices, shipping, logistics et cetera.
Business Metrics
Performance Production Times Customer Churn
Sales Totals Lead Analysis Click through Relations Budgets
Invoices Shipping Documents Pick List
Executives KPI’s Scorecards and Dashboards
Analysts,
Senior managers Department
managers Employee
partners Production reports Management
reports OLAP exploration
Figure 6 – Different BI user needs in the hierarchy (Solberg Søilen, 2008)
(X)
Figure 7 – The special department model of intelligence (Solberg Søilen, 2008)
X
Figure 8 – The advisory model of intelligence (Solberg Søilen, 2008)
3.7 BI placement in the organization
3.7.1 The special department model of intelligence
In the special department model the intelligence function is placed in a separate department as a part of e.g. marketing department. The idea is to have a special intelligence department bedded inside an already established department where it would be easy make analysis and draw conclusions by already established employees. A problem that can occur with this model is mostly isolation because special intelligence departments often close themselves in. This creates misunderstandings and develops self initiated projects that often are not sure to be useful for the company. Communication between top managers and the departmental team must be very well established in order to make correct and needed analysis (Solberg Søilen, 2008).
3.7.2 The advisory model of intelligence
The advisory model places a senior advisor to the CEO and to top management. The senior advisor is then responsible for two functions in the intelligence cycle, formulating the questions to be answered, and delivering the results. All though the information gathering and the analysis making must not necessarily be performed by the advisor himself. Some advisors simply do not have the time or
X
Figure 9 – The professional model of intelligence (Solberg Søilen, 2008)
the skills to search for necessary information so, therefore, a separate department or even an outsider could be used according to the needs. In this model, the senior advisor has access to a certain number of workers who he believes have the certain knowledge that is required to perform the intelligence work. But the problem often lays in the quality of intelligence that is carried out. Reports and analysis tend to be less accurate, relevant or effective than if they were done by professionals (Solberg Søilen, 2008).
3.7.3 The professional model of intelligence
In the professional model, special personnel have gone through sufficient training in form of university studies and some practical training in intelligence work and are therefore used by the organization for their specific professional knowledge. Their main priority is intelligence work. In this way the organization can use professional workers and benefit from a broader supply of services offered like field-work, in-the-terrain intelligence gathering or desk jobs with intelligence analysis, reporting and presentations. In most cases the professional model is a question of resources. Mid-size and smaller companies find this model expensive and in very few cases they will have the funds available to make investments in this kind of skilled intelligence personnel (Solberg Søilen, 2008).
X
Figure 10 – The top-down model of intelligence (Solberg Søilen, 2008)
X
X X X X
Figure 11 – The integrated intelligence model (Solberg Søilen, 2008)
3.7.4 The top-down model of intelligence
The idea of this model is to gather and communicate intelligence and knowledge from the top of the organization. Top management processes all intelligence and spread it out on the need-to-know basis downwards throughout the organization. Companies that use this model are often small or midsize and have low-tech production based organizations with low-skilled workers. Top managers are often the ones with most knowledge and best qualifications in the company and are classified as the most important persons for the running of the firm. The problem that can occur with this model is that the top management stops listening to what other workers have to say (Solberg Søilen, 2008).
3.7.5 The Integrated Intelligence Model
In integrated intelligence model, intelligence activities are run on a basis where every employee, on every level in the organization, contributes with intelligence. In this way the whole organization’s experience and effort are collected. When working according to this model, the intelligence seems to be less secretive and less dangerous, which of course is an advantage when building trust and creating an atmosphere where everyone is feeling that information they share is important. This model is very often practiced in Japan (Solberg Søilen, 2008).
X X X X
Figure 12 – The down-up model of intelligence (Solberg Søilen, 2008)
X
Figure 13 – The departmental model of intelligence (Solberg Søilen, 2008)
3.7.6 The down-up model of intelligence
In companies where employees on the lower levels in the organization have access to especially important information, a down-up model can be applied. It is common that competitive sales and marketing driven organizations use this model when they work with intelligence gathering and intelligence processing. When workers, often sales people or other field and out of office workers, have a direct contact with the customers it is crucial that they bring their knowledge and intelligence back to the company where top managers and CEO’s can deal with it and support decision making. It is common that companies that use this model also reward their low level workers in form of incentives, such as higher salary, so that information they bring home is more valuable and more effective. Some companies even use Intelligence Reward Systems where information is divided into different classes and workers can systematically be rewarded via the company’s intranet system, for the information they bring home. The more value the information has, the more money on the paycheck (Solberg Søilen, 2008).
3.7.7 The departmental model of intelligence
In the departmental model of intelligence, a company dedicaties a whole department for only Business Intelligence operations. Companies have full time intelligence officers and analysts
working in the BI department. It is common though that these also work in other departments or as a supporting function to top management. The idea is to increase the focus on intelligence gathering as much as possible. It could be a question of security and development of new technologies as well as on the rising the level of information quality, effectiveness and perfection (Solberg Søilen, 2008).
3.8 Some BI Tools on the market today
Oracle Enterprise BI Server is a tool designed by the Oracle Corporation which is the world’s largest enterprise software company. This tool includes dashboards, ad hoc queries, intelligent interaction capabilities, enterprise and production reporting, financial reporting, OLAP analysis, data mining, and other Web Service-based applications (Oracle, 2008).
Business Objects Enterprise is formerly designed by Business Objects, now SAP Company and is built on a service-oriented architecture. Some key features are: Auditing, BI content Search, Information Portals, Web based Queries and graphical design tools (Business Objects, 2008).
SAP NetWeaver BI is designed by SAP Company and is installed in the organization’s network and can be accessed by most users. It consists of data warehousing, OLAP, Business planning, Queries, Reports, Analysis, Open Hub services, Information broadcasting et cetera (SAP, 2008).
SAS Enterprise BI Server is a server, designed yet again, by SAP Company and allows organizations to quickly access and derive the information they need to
make better decisions. Some key features are Targeted fit-to-task Web and desktop reporting interfaces, Multiple self-service query interfaces, OLAP, graphic data presentation options, integration with Microsoft Office, a dynamic desktop interface for guided analysis and model development et cetera (SAS Enterprise, 2008).
TM/1 & Executive Viewer is an advanced analysis and reporting tool designed by IBM Cognos and is a working through real-time Web-based access to information from OLAP. Some other key figures are Ad hoc analysis and dynamic graphical reports (IBM Cognos, Cognos TM1, 2008).
BizzScore Suite is a tool designed for nonprofit organization by EFM Software.
The tool is build upon four components Bizzscore (management dashboards, extensive analytics, built-in action management and messaging), Bizzdefiner (strategy based formulation of the performance management blue-print and KPI's from mission to measure), Bizzdata (for integrating a variety of data-sources and scheduling its import), and Bizzquality (for input of “soft” data such as customer satisfaction and employee motivation using web based questionnaires) (Bizzscore, 2008).
WebFocus is designed by Information Builders and is a one single platform for enterprise business intelligence. It contains integration tools such as: web services including data and application adapters, Real-time transformation, and Process- driven BI. WebFocus also includes; dashboards and scorecards, queries and analysis, reporting, portals, and information delivery (Information Builders, 2008).
Excel, Performance Point, Analysis Server are all designed by Microsoft Corporation. Excel can be used as ananalytical tool working in a spreadsheet environment for exploratory data analysis. Researches show that Excel is the most used spreadsheet tool in the world today (Kelly, 2008). Performance Point and Analysis Server are designed to monitor, analyze and plan for the organization so that the changing business conditions are met at all time. Some key figures are scorecards and dashboards, analysis, planning, budgeting et cetera (BI, 2008).
QlikView is an analysis tool designed by QlikTech. Some key features are:
Analysis (online and offline), dashboards and scorecards, reports, alerts, and zero footprint DHTML-client that gives user the access to full web-based analysis without any installation requirements on the client machine (QlikTech, 2008).
Microstrategy is designed by the Microstrategy Company, and is a real-time business monitoring tool. Besides real-time functions, it contains analysis, reporting and other intelligence-integration functions such as heterogeneous joining of data and data marts (Microstrategy Inc, 2008).
Hyperion System was formerly designed by Hyperion Company, but was later bought by Oracle. This system consists of several functions such as financial managing including planning, strategy and quality management. It also includes performance scorecards and dashboards (Oracle, 2008).
Actuate is a reporting application designed by Actuate Company. This tool focuses mainly on reporting designs and modifications as well as users’ analysis of information. It uses electronic spreadsheets, electronic reports via an open
