BEM - Business Intelligence Expansion Model

(1)

BEM - Business Intelligence Expansion Model

Jefta van Vliet University of Groningen Faculty of Economics and Business

Thesis submitted for the Degree of Master of Business & ICT at the University of Groningen

26-03-2010

(2)

Preface

Before this thesis I researched topics regarding BI such as a BI software selection tool and a BI qualitative improvement model. Furthermore I have done extensive theoretical and practical research in BI. Besides BI I have knowledge in many business and especially IT related aspects. Also some economic and commercial knowledge are present. In the future I would like to work as an IT/business consultancy expert and perhaps one day start my own company.

Topic BEM

The reason I choose the business intelligence expansion model as the topic for this research has several reason. One of these reasons is to expand my knowledge regarding BI. Other reasons are that I am interested in the wide range of aspects BI spans and that this subject seems to be unexplored and challenging. With this research I expect to add knowledge to the field of BI and maturity scans.

Furthermore with this research a model is created that can help BI consultants or BI experts to map the extent of BI and to find possible areas to improve BI within companies. This in turn will help

companies to expand their BI facilities use and gain more advantages from them. Therefore creating a useful and beneficial model for companies that can be used directly in practice is a stimulating thought. This is more stimulating for me then some abstract results, which can however also be very valuable. For this research mainly academic resources where used to verify my hypothesizes regarding the extent of BI. This thesis is not about the qualitative aspects regarding BI. Often it is easier for people to visualize qualitative aspects regarding BI instead of quantitative aspects. This became clear from talking with several BI experts. This does however not mean that quantitative BI is less

important the qualitative BI. They cannot exist without each other. A certain quantity of BI always has a certain quality and visa versa.

Acknowledgement

Finally I like to thank everyone that helped with this Thesis. First and foremost I would like to thank prof. dr. Bert de Brock who provided valuable insights, useful feedback and we had many interesting discussions. I would also like to thank dr. Thomas de Boer for his feedback. Furthermore I like to thank Centric for there help during this Thesis. I would especially like to thank Joop Rass and dr.

Marjan Dijksta from Centric. Also I would like to thank all Centric employees that helped with the workshop. I would like to thank dr. Douwe Postmus who provided valuable advice and support regarding this thesis. I would also like to thank everybody that helped with the testing of BEM in particularly drs. Gerrit van Dijk, drs. Frank Diepmaat and Leon van Gorp. Finally I like to thank my family and friends and all others for their support during this project.

(3)

Abstract

Before this thesis there was no model available to map the extent of Business Intelligence. Therefore there is no factual method to measure to what extent BI is used in companies and therefore no reliable way to improve the extent of BI use in companies. To reliable expand BI within companies it is vital to first know the current situation regarding the extent of BI.

Hence this thesis has created such a model called the BEM – Business Intelligence Expansion Model.

By combining dimensions that play a role in the extent of BI into a model, BEM is created. By linking questions to this model the BEMMT – Business Intelligence Expansion Model Measure Tool is created. When BI-users in a company use BEMMT it is possible to map the extent of BI within companies.

To test the BEMMT it is deployed at the universities of Groningen and Tilburg. The results of this test provided extensive valuable information and verified that the model successfully maps the extent of BI. The test provided detailed information about where BI use can be extended in the universities.

Academic literature about BI and maturity scans is rare. This thesis provides valuable insight in both BEM and maturity scans and helps to move both practices to a new level of knowledge. Furthermore with the use of BEM it is possible to extent the BI use within any company which will make

companies more effective and efficient thus leading to an economic boost.

(4)

1. Introduction

The subject of this thesis will be the Business Intelligence Expansion Model, abbreviated to BEM. The BEM is focused on measuring the extent of BI – business intelligence use within a company. By gaining this knowledge, it is possible to find areas for improving the extent of BI use. Therefore, this thesis is focused on aspects/factors relevant to the extent or breadth of BI use within companies, such as the number of BI users.

Business Intelligence

This thesis subject is about business intelligence often abbreviated to BI. For this thesis the following broad definition of BI will be used: BI is every aspect required to attain and use knowledge extracted from data. In the definition, the term data is used instead of the term information. This is done because information is already put into context, in for example a report to make the data useful. Therefore, it is already a step beyond data and does not include the activities required to make information out of data, such as aiming for data and collecting data. To also include the activities revolved around

transforming data into information, the term data was chosen; even though knowledge cannot be attained from data directly, it is required for the first steps into attaining and using knowledge. The data itself can be very broad like customer data or internal company data.

Aspects/factors

The term aspects which may also be called factors, can for example be activities such as: to what extent evaluation is done or to what extent disseminating the information to the right locations is done.

It can also be resources such as employees or technologies. There are many aspects/factors that play a role in the extent of BI use. When a company uses BI only for a few specific parts of a factor, it will miss out on many of the possibilities the BI practice has to offer. An example of a factor that plays a role in the extent if BI use is level. In this factor BI can be used on three levels, which are: strategic, tactical and operational. If BI is only or mainly used on one of these levels the opportunities and capabilities BI offers on the other levels, are not fully exploited. This means the extent to which BI is used within the organization can be expanded. Therefore the model that will be created in this thesis project will first map the use of BI within the organization via a measurement tool. This will be done in a similar way as other maturity scans do this. However after it is mapped, this model will go one step further as the results of the measurement tool will be very detailed and show where and how to expand the use of BI within the organization. The expansion of BI in a company is however depended on more factors not directly linked to BI, like the financial capabilities of the company, which are not part of this model. Therefore this model has been named Business Intelligence Expansion Model.

(Zwanenburg, 2008, Rodenberg, 2004; Thierauf, 2001; Pearce, 1976; Huber, 1984).

(7)

Why BEM

In a short discussion with prof. dr. E.O. de Brock we came to the idea to create a scan for Business Intelligence. In a previous project I already worked on a similar project. The model of this similar project is called Business Intelligence Improvement Model (BIM). I enjoyed working on BIM and found it very interesting and learned a lot from it. Therefore I liked the idea to work on a similar project but with a different perspective. The main reason for choosing this subject is that it seemed challenging, besides this I could also learn from it and find it interesting. Besides this, very little information about these types of scans have been researched and made publicly available, especially in the area of BI. Therefore this paper will add knowledge to the field of maturity scans but even more to the field of BI. This paper aims to create a model that can be used in any company to increase their extent of BI use and can be used as a framework for further similar models.

Other models?

As there currently is no known model focused on extending the BI-use via dimensions this will be a valuable and essential model for many companies and BI-experts. Furthermore it is the only fully publicly available model of its type at the moment regarding the mapping and increase of the extent of use of BI. For many companies their extent of BI is in dire need of attention as often there are only a small number of BI users or only a few BI possibilities are exploited. Therefore this thesis topic is considered very important and it is remarkable no earlier attempts have been made to create such a model. This is most likely due to the complexity and broad range of the subject. Even so this should not be a valid excuse to not delve deeper in this subject.

Achieving BEM

In order to achieve BEM, as a model that maps the extent of BI use and shows possibilities to increase the extent of BI use, it is vital that literature is found regarding the extent of BI. Form this literature the factors/aspects that play a role in the extent of BI can be extracted.

BIM vs. BEM

BIM is focused on improving the actual BI-system use in a company. For example to improve data mining features or user friendliness of the BI-system. The subject of this thesis, BEM, will however look at the extension/expansion of actual BI use in a company. For example to what extent is structured query language – SQL, which is by the broad definition of this thesis seen as part of BI, used by the departments of a company and could this be expanded. The focus of BIM is on the quality of the BI use, what we call the depth of BI use. The focus of this thesis project (BEM) however is on the extent of BI use, what we call the breadth of BI use. Figure 1 shown on the next page illustrates this.

(8)

Figure 1: BIM VS BEM

BEM and BIM

BEM can be used in cooperation with BIM, also see Figure 1. The company could choose to first expand the use of BI and then the quality of it. Once BI is readily available in the company more valuable information can be gained to improve the quality of BI.

The company could also decide it is vital to first improve the quality of the BI practice before expanding its use. The quality of the BI practice may be so low that it is vital to first improve it before expansion is possible. Finally it is also possible to use BIM and BEM concurrently. For example the department marketing of the company may already be score high in the BEM, but still can gain a lot from applying BIM. While other parts of the company, for example procurement, are still low on the BEM scale and can start by improving BI via BEM. Many more examples are possible, nevertheless each of these approaches to improve BI use has advantages and disadvantages and the right approach depends on the situation. However the models reinforce each other making them even more useful.

(9)

Document Guide

This introduction of the Thesis will be followed by a chapter about Centric. Chapter three will examine the problem regarding this thesis subject further. Chapter four will discuss the applied research methodology. All the BEM dimensions that are extracted from the literature will be discussed in chapter 5. Chapter six will summarize the model in overviews. Chapter seven will discus the maturity of BI extent. In chapter eight the model will be made measurable by translating it into questions and answers. Chapter nine will provide information about how to apply the model in practice. Information about the validation of the model can be found in chapter ten. Chapter eleven will discus the most important aspects of the thesis. The next chapter, chapter twelve will give a summary of the thesis and in chapter thirteen the conclusion will be given. Chapter fourteen is dedicated to reflecting in back on the thesis and problems that occurred. After this the research limitations and further study possibilities will be discussed in chapter fifteen.

(10)

2. Centric

The first phase of the project was done in cooperation with Centric with the main goal to test the initial version of the model. Besides this reason there where more reasons why Centric was chosen like:

• Centric knowledge base (many consultants in BI with practical knowledge).

• Centric is very enthusiastic about the thesis subject

• Centric offers interesting career possibilities

• Possibility to expand my BI knowledge

• Friendly and helpful employees

• Convenient location and workplace at home

Centric locations

According to the information from the Centric website (Centric, 2009), Centric is a company with about 5.500 employees. It has many offices in the

Netherlands for example in Groningen,

Amersfoort, Diemen, Emmen, Haarlem, Ijsselstein, Amsterdam, Eindhoven, Rotterdam, Weesp and Zwolle. Centric also has offices in Belgium, Germany, Norway, UK and Switzerland. Besides this Centric is also active in Romania, Luxemburg and Sweden. However this project will be done for the office in Groningen. A picture of this office is shown in Figure 2 to the right. This office is located in the “Noordelijk Trade Center” in the

Netherlands together with other companies. Figure 2: Centric Groningen

Centric organization

Centric consist of three main components which are: Finance, Centric Holding and Oranjewoud.

Finance is focused on financial services, like finance, accounting and security aspects. The Centric Holding is focused on ICT-Services such as Consultancy, software engineering E-business and more.

Oranjewoud is focused on advice and engineering services like: Secondment or Space and Mobility.

An overview of this is given in figure 3 on the next page.

(11)

Figure 3: Organization structure Centric

The department where this project is part of is called SE which stands for Software Engineering. SE in turn is part of Centric Holding, also see figure 4. A part of SE is PS which stands for professional services. The business intelligence activities of Centric belong to PS. Therefore this project also belongs to PS. PS in the Netherlands is split into three regions, namely NorthEast, West and mid- south, also see the figure below. This project is done for the establishment in Groningen and therefore is part of the NorthEast region of PS. The establishment in Groningen has 60 employees of which about 14 are active in BI.

Figure 4: Professional services regions

NorthEast

Software Engineering

West mid-south

Professional services Finance

Financial services Finance Accounting

Security

Centric

ICT-Service Consultancy

IT solutions Software engineering

E-business Systems integration Managed ICT services

Training

Advice- and engineering services Space & Mobility Environment & Safety

Sport & Techniques Secondment

Centric Holding Oranjewoud

(12)

Centric Activities

Centric Holding has split her core activities in three main types, namely: managed ICT services, IT solutions and software engineering. Under managed ICT services activities such as ICT management and outsourcing, ICT infrastructure and migration, training and advice (consultancy) can be found. IT solutions consist of activities like: standard solutions for specific generic branch processes, integration of solutions in existing or new infrastructures and delivery and hosting of infrastructures. Finally software engineering is focused on: custom made software, quality care, process improvement and testing of software and more. Centric is active in many braches ranging from finance and government to travel and medical care. An overview of Centric activities is shown in figure 5 below

Figure 5: Centric Holding activities

IT Solutions Managed

ICT Services

Software Engineering

Total solutions in the area of office automation:

• ICT Management and outsourcing

• ICT infrastructure & Migration

• Flexible support

• Training

•^Advice

Centric Holding

• Standard solutions for branch specific and generic company processes

• Integration of solutions on existing or new

infrastructures

• Delivery and hosting of infrastructures

• Customer specific custom software for front- and back offices

• Development, integration, renovation, migration and management of applications

• Quality care, process improvement and testing of software

• Message exchange and security

• Embedded excellence

(13)

Mission and culture

Centric’s mission is to be a knowledgeable partner for solutions in ICT. Stability and continuity towards customers are critical in this for Centric. Centric states to be special in the simplest way possible. Centric works in a way that is focused on projects. According to Centric this leads to insightful responsibilities, clear deadlines, quality and budgeting. Centric likes to do business in an open way. Centric claims their people have character and are not afraid to think outside the box to find the best solutions. Centric tries to keep there promises and deliver on time (Centric, 2009). The success of Centric especially the last years is clearly visible in figure 6 and 7 below. On a side note Centric did acquire some new companies in this period of time like Altro Consult and SP Solutions.

Figure 6: Net turnover Figure 7: Profit after tax

Net turnover

0 100 200 300 400 500 600 700 800 900

1998 19 99 20

00 20 01 20

02 20 03 20

04 20 05 20

06 20 07

Amount (in million euros)

Net turnover

Profit after tax

0 10 20 30 40 50 60

1998 19 99 20

00 20 01 20

02 20 03 20

04 20 05 20

06 20 07

Amount (in million euros)

Profit after tax

(14)

3. Problem description

There are several issues leading to this subject. The main issue is: many companies use BI in some form, yet they mostly specialize in one or a few applications of BI and miss out on the many possibilities offered by BI. Some examples of this will be given below:

Example 1

Difference between departments or even holdings of a company in the extent of BI-use can leave many BI features unused. However, it is not always so that the department with the lowest BI extent gains the highest benefits by increasing their extent of BI use or that the BI use has to be equally high in every department throughout the company. Department A for example has a high extent of BI even though improvements are still possible and this department also gains a lot by using BI. In contrast department B has a low BI extent but also gains little from using BI. Compared to department B department A is the wisest option to improve, even when its level of extent is higher, the benefits gained by improving department A are much higher then with department B.

Example 2

If BI is used for a certain goal, advantages that may be gained through other goals may be neglected.

For example, when gaining knowledge is the main goal of the BI system, it will not or be less optimized for decision making or decision support. Therefore advantages that may be gained by focusing on decision support or decision making are lost. Therefore to gain most advantages of the BI facilities the goals should all be exploited as intensively as possible and not just one or a few.

(Zwanenburg, 2008).

Example 3

The form in which the company sees BI is also very important in the way the company uses BI. For example if the company sees BI as a product it is focused on the outcomes and less on the process that comes before this. This process nevertheless is vital to get quality outcomes (products) from BI. This means that by limiting their focus on the outcome (products) they are actually limiting their own BI facility possibilities. Each form offers benefits and some even counter the disadvantages of other forms. Therefore if each form of BI gets sufficient resources the outcomes, processes and other forms will all increase the value and use of the BI system and become more valuable for the company (Zwanenburg, 2008).

(15)

Example 4

The set of activities regarding BI is very important to determine the usefulness of the BI facility.

Examples of activities that are part of the set of activities regarding BI are gathering data or analyzing data. Many companies gather a lot of data that they never use or they do not gather the data they should gather, but instead unwillingly accumulate useless data. Even so, they may spend great efforts on analyzing the data or other activities. All the effort in these activities is wasted however, as the data is useless. In these cases, the set of activities are not aligned correctly. For the correct use of BI all the BI activities, ranging from measuring the right data to evaluation of the BI process should all be aligned (Zwanenburg, 2008). Besides that these activities should be aligned, they should also be of an as high as possible extent of BI so that these activities can provide the right results and possibilities for the company.

Example 5

The source from which BI information is extracted can also limit the extent of BI use. If only internal sources are used, many external possibilities such as the internet or television that might provide valuable knowledge are lost(Zwanenburg, 2008).

The points made in the examples can be summarized as: even while the use of BI is becoming more popular in companies, it is very likely that there are still many possibilities to increase the extent of BI use in most companies throughout several aspects/factors. This will allow companies to gain many more benefits from their BI-facilities.

3.1 Research objective

With this in mind, it would be useful to have a model that can be used to extent the amount of BI use.

To realize this, the current situation regarding the extent of BI use in a company needs to be mapped before improvement can be started. This leads to the following research objective:

“To create a Business Intelligence Expansion Model that maps the extent of BI use and opportunities for BI expansion within organizations.”

This objective leads to several sub questions. Some sub questions:

• What literature is available that can be used to help create the BEM?

• What dimensions regarding the expansion of BI are there?

• Can these dimensions be translated to a model?

• Can this model be translated to a maturity model?

(16)

• How can the model be measured?

• How far can the advice go that the model gives?

Extent of BI use

The extent of BI use has been mentioned a few times before. When this term is mentioned it is about the breath of BI use within a company or companies. The extent of BI use or breath of BI use can be related to many things such as in what form or forms a company sees their BI, they can see it for example as a process, tool or from a more philosophically perspective or as a combination of them.

More examples are the companies BI-ambition and the BI activities which are all related to the extent of BI use.

Definition dimension

What previous was described as the aspects or factors that play a role in the extent of BI can also be seen as dimensions. From a mathematics and physics perspective, dimensions of an object or space are seen as the minimum number of coordinates to define each point within it. These dimensions are orthogonal from each other, meaning that each dimension in the BEM can be measured independently from each other. To determine the extent of BI use within a company, their current situation needs to be known. This means they are at a certain point regarding the extent of BI use. To determine this point, the coordinates that determine the extent of BI use are required. These coordinates are the dimensions that determine the extent of BI and how they score in a company. This means that the dimensions of BEM together give insight in the extent of BI use within a company. By measuring the performance via a score of an organization on these dimensions, insight will be gained regarding the extent of BI use.

3.2 Requirements of BEM

In order to increase the chance of success of the thesis several requirements will be followed during the realization of this project. Theses requirements are:

• BEM must apply to any company

• BEM must be usable in any situation

• BEM must be measurable

• BEM must be focused on a practical and theoretical perspective

• BEM must add to the understanding of the BI-field

• BEM must be usable by BI-users

• BEM must be interpretable by BI-experts

• BEM must be detailed but not overwhelming

(17)

• BEM must be translatable to online software.

Following these requirements will allow BEM to become a useful model for the BI-field and an example for other fields. BEM must be usable in any company within any situation. This increases the usability of BEM as each company regardless of their level of BI extent can use BEM. Also BEM is usable as a measurement tool to measure the BI-state of the company. To gain any useful information from BEM, it is vital that the model is measurable. This will be attained by translating the

aspects/factors of the model to questions which can be measured. Furthermore, the model needs to be focused on both practical and theoretical ideas which will offer real knowledge about the BI state of the company. To reach this practical and theoretical perspective, both literature and BI-expertise from experts is used. This research will add to the understanding of maturity scans and to the knowledge of BI. The model and especially the question tool are meant to be used by any BI-user within a company.

Nevertheless, only BI-experts are expected to make trustworthy conclusions based on the results of the model. BEM can be detailed but this should not make it too complex to use. In order to achieve a reduced complexity, clustering of dimensions is used to combine similar dimensions together. Finally it must be possible to create an online software tool to automate the BEM procedure.

(18)

4. Methodology

The goal of this research is “To create a dimensional Business Intelligence Expansion Model that maps the extent of BI use within organizations.”. This research is mainly based on a literature research with interviews to verify and extract information. The methodology can be given in several steps which are given below

Overview of the methodology

In this section the steps taken to conduct this research will be summarized. The global steps taken to do this research are:

1. Screen the literature for possible dimensions related to the extent of BI use

2. Critically analyze if the dimensions are really related to map the use of BI and not for example related to the causes explaining the extent of use of BI.

3. Initiate detailed searches of literature to find the sub-dimensions, sub-sub-dimensions and values to map the entire chain regarding each dimension.

4. When the dimensions, sub-dimensions, sub-sub-dimensions and values have been found combine them into a model

5. Test the model with BI-experts 6. Translate the model into a framework 7. Test the model in practice

Screen the literature

The first step of this project is to orientate and familiarize oneself into the literature regarding BI and in particularly regarding the extent of BI. It helps if knowledge about BI is available at forehand. If this is not the case this task might proof time consuming but not impossible. It is required that a firm understanding of BI and the extent of BI is available to the creator of the model to successfully conduct the other steps of this project. From the literature the dimensions that play a role in mapping the extent of BI must be collected. Once the dimensions have been collected and their relation to the extent of BI can be sufficiently proven or described they are collected together and written down.

They can then be added to a concept model for example in a table or overview.

Critically analyze

This concept model is critically analyzed by not only the author himself but also by several other independent researchers such as the project examinator and Centric. Some dimensions for example, may seem relevant to the extent of BI, but do not directly measure it and therefore have no place in this model. Examples of this are dimensions related to the causes of the BI-extent such as culture and

(19)

resistance and are not related to mapping the extent of BI use. Therefore only the dimensions that directly map the extent of BI use are used during this research.

Detailed searches of literature

Once the dimensions are found, literature regarding them should be collected and analyzed to create a greater understanding of them. This will lead to sub-dimensions and deeper levels.

Combine them into a model

Once the details (sub-dimensions etcetera) regarding the dimensions are subtracted from the literature they should be combined with the dimensions into a model. This model consists of several levels with the top level being dimensions followed by sub-dimensions etcetera.

Test the model with BI-experts

To test whether the model is correct, complete and usable, the model is tested in a workshop session with BI-experts who can comment on each dimension and suggest new dimensions, sub-dimensions or deeper levels. This step validates the model and confirms the completeness of it.

Translate the model into a framework

Once the model is tested it needs to be translated to questions to be measurable. Therefore the values in the model are translated into questions. This way information about higher levels can be retraced by using the values to determine the BI extend regarding all dimensions of the model. Once the values are translated into questions they are collected into a framework that can be used to measure the extent of BI on a detailed level.

Test the model in practice

The final step is to test the model in practice by offering the framework to a company and let their BI users fill it in. Then translate the results of the test into the thesis.

.

(20)

5 BEM dimensions

In this chapter the Business intelligence Expansion Model will be given which is based on the dimensions regarding the extent of BI. Therefore before the Business intelligence Expansion Model can be given the dimensions of the model will be explained in this chapter. An overview of all dimensions, sub-dimensions and values can be found in appendix D. The model is focused on the BI- users within a company. BI-users are users that in some way interact with the BI system. This

excludes users that only read static reports and not themselves interact with the BI-facilities. These BI- users therefore can be an extensive and varied group of people like business analysts, salesman, marketeers etcetera. Therefore an attempt is made to make the model as simple as possible. This has been done by detailing the model in a way that abstract and difficult to understand terms are omitted from the lower levels of the model. Furthermore the terms are explained and many examples are given to make them easier to understand. Besides this in the framework to interview a company, it is also possible to add descriptions to make the model even more accessible. Besides this in most cases if people have not heard of the term it is unlikely they use it. To reduce the risk of unknown or unclear terms even further before the BEM is used in practice, an information/introduction session can be given. In this session terms can be explained and questions should be answered. In this way the use of BEM is supported in an advanced way. The BI-users are therefore a target of the BEM, yet the BI- experts that analyze the results are also a target. BEM must provide sufficient information to give an advice and show where improvements are possible regarding the extent of BI. This results in valuable information that the BI-expert can use to give advice, based on the companies situation.

5.1 Levels of detail

As mentioned before BEM will have several levels of detail. The level of detail of the model can go very deep. However, the goal of the model is to create an overview of the extent of BI use within a company. If this view is too detailed, it may become too complex. On the other hand, if the model is not complex and detailed enough, it may provide very little useful information and be limited to just a global overview of the extent of BI use throughout a company. Although this is useful information, it does not allow a company to actually improve their extent of BI-use. This is because they do not know what the opportunities are to increase the extent of BI use. Therefore, this project will aim for the middle way; BEM should provide enough information but not become too detailed and overly complex. For this reason four levels of detail will be used. These levels are:

First level - Dimension Second level - Sub-dimension

(21)

Third level - Sub-sub-dimension Fourth level - Values

As we can see above and in figure 8 at the first level there are the dimensions. These dimensions are abstract and by itself do not provide ample information to improve the extent of BI. For example the dimension BI users by itself provides abstract information, but the value “percentage of BI users within a company” provides more detailed, clear and usable information to improve the extend of BI.

Dimensions provide a global overview of the extent of BI use per dimensions. The average of these dimensions together provides an overview of the total extent of BI use throughout a company. The second level contains the sub-dimensions that are considered a part of the dimensions. They are more detailed and provide a categorization within dimensions. On the third level there are sub-sub-

dimensions which are even more detailed then sub-dimensions and are also used for categorization and allows for deeper levels of details. On the lowest level there are the values. Each dimensions, sub- dimension or sub-sub-dimensions exist of values. Even so, per company and situation only one of these values can be valid. Questions can be formed to determine these values to measure the extent of BI use per situation and per company. Not all dimensions need to have sub-dimensions or sub-sub- dimensions, but they always have values otherwise they cannot be measured.

(22)

Figure 8: Levels of detail BEM

(23)

5.2 Given values

BEM itself is focused on dimensions where it, from a BI perspective, can play a role in. There are many more dimensions that play a role in the extent of BI use, but are not BI-related or values that cannot be changed from a BI-perspective or with a model. These dimensions should still be measured but are not part of BEM. These dimensions are called given values because they cannot be changed based on the outcome of the BEM model. Two examples of such dimensions are company size and company activities.

Company size

Bäcklund(2004) researched the effects of company size on ICT usage in Sweden. An overview is given in figure 9 below. This figure shows that there is a difference in ICT use between smaller and larger companies. Therefore the company size plays a role in the extent of BI use. Note worthy about figure 9 is that the data about email in companies with a size of 200-499 is missing. Bäcklund (2004) does not give the reason why this information is missing in this particular article. Therefore it could be any reason, like that it was perhaps not measured or perhaps the number of measures where to low to add. Even though this information is missing, the figure still provides valuable information regarding company size.

Figure 9: Company size (Bäcklund, 2004)

(24)

Company activities

If a company is only active in one country it means the use of BI is focused on that country. However if a company is international active the extent of BI use is not focused on one country. As the article of Porter et al (2001) shows much can be gained by being active international. Nevertheless to manage a company successfully the article also shows different and more information is required. Information for example about different approaches to take and to make the right decisions. More information will be available as the company is active in a much bigger playing field compared to companies that are active in one country. This means the role of BI to manage information will become even more important and extensive. Therefore company activities plays a role in the extent of BI use.

Why given values

Both Company size and Company activities play a role in the extent of BI as the previous text shows.

However the fact that the company is active internationally or nationally, or the size of the company are both values that should not be changed based on the outcome of the model. For example the model can measure that only 10 people work in a company and it is active nationally. Therefore the model would advice to increase the number of employees and become internationally active to increase the extent of BI use.

However it is not always possible to expand a company and it does also not always lead to an extent of BI use. For example a company may become larger but percentual the use of BI may remain the same or even become lower. Furthermore it is a business related issue to expand or not and less related to business intelligence. This means that these dimensions are valuable to measure to see for example if a small company is using BI in a similar way other small companies use BI, or if an international company uses BI in a way it should be expected from such companies etcetera. However it is not part of BEM as it not directly based on a business intelligence perspective and does not fit into the model which is meant to map and extent the use of BI.

The same reasoning also counts for the sector in which the company is active and more similar dimensions that are given values and should not be changed based on only a model. These given values can be measured and are often values that do not quickly change such as the sector in which the company is active. Therefore they can be measured easily and then compared to each other. For example to compare the sector “Industrial Goods” BI use to the sector “Services”. Or perhaps even more interesting is the ability to compare the results of for example 1 financial company to the results of other financial companies. Finally the previous means that BI is influenced by the given values but BI does not influence the given values and therefore they cannot be a direct part of this model that aims to improve the extend of BI.

(25)

5.3 Dimensions

In the remainder of this chapter the research question “What dimensions regarding the expansion of BI are there?“ will be answered.The answer to this question is that BEM consists of the following dimensions regarding the expansion of BI.

• Global internal variables

• BI-definition

• Data form

• BI users

• BI usage forms

• Alignment

Theses dimensions will be described in the following paragraphs.

5.4 Global internal variables

The global internal variables are dimensions that come from within the company. There can also be global external dimensions such as company type, sector and location, but these dimensions are more cause related then they are to mapping the extent of BI use and therefore are not part of the BEM.

The dimension global internal company variables, includes the sub-dimensions:

• Establishment/building

• Division

• Department

• Project/work group

These sub-dimensions all provide internal information about the company. The extent of BI directly depends on how it is used throughout the company. There can be a difference in the extent of BI use within the company. There could for example be a difference in the extent of BI use between the department R&D who use advanced BI to support innovation, and the department production who do not use BI at all. There can also be differences in the extent of BI use between

establishments/buildings, divisions and project/work groups. The values of these dimensions depend on the situation and vary widely. Besides this they do not provide a direct overview about the extent of BI, but can be used as analytic tools to for example discover how each department or establishment uses BI.

(26)

5.5 BI-definition

This dimension is inspired by the framework of BI definitions described in Zwanenburg (2008). In the introduction it was already made clear that for this project a broad definition of BI is used,

incorporating many aspect regarding the use of information within companies. This definition is: BI is every aspect required to attain and use knowledge extracted from data.

Depending on how a company defines their BI use this has a great impact on the way BI is used. For example when a company is only or mostly focused on gathering information a lot of potential from other activities such as analyzing or using information will be lost. Zwanenburg(2008) also describes the dimensions data type, data source and data time which have been deemed more suitable as sub- dimensions under the dimension data form, which will be discussed in the next sub-chapter. Of the framework of Zwanenburg(2008), except BI aim, the following sub-dimensions are part of the BEM:

• Form

• Goal

• Level

• Set of activities

• Drive

• BI aim

5.5.1 Form

According to Zwanenburg (2008) and other literature, form is an elementary dimension in determining the definition of BI. If a company for example sees BI in a very specific form or only exploits a few forms the advantages from other forms might be suppressed or not exploited at all. This can have a negative impact on the extent of BI use. Therefore the extent of BI use depends on the form in which the company sees and uses BI. Therefore this dimension is added as sub-dimension to this model.

Zwanenburg (2008), Philips and Vriens (1999), Gilad et al. (1986), Rouibah and Ould-ali (2002), Zeng et al. (2006), Rouibah and Ould-ali (2002), Thierauf (2001), Herring (1988) define the following sub-sub-dimension in a way that they are part of the sub-dimension form in BEM:

• Process

• Product

• Organizational function

• Approach

• Set of technologies

• Tool

• Discipline

• Philosophy

(27)

Zwanenburg (2008) extracts process as a value from the article of Gilad and Gilad (1986:53) and sees product as the outcome of a process based on the literature of Thierauf (2001:4). Herring (1988:5) supports both process and products as values for this model, see the quote below:

“Intelligence is hard to define because it is both a product and a process. The product is actionable information; the process is the systematic means of acquiring, evaluating, and producing the actionable information.”

According to Zwanenburg (2008) the literature also acknowledges the forms:

• an organizational function (e.g. Philips and Vriens, 1999; Gilad et al., 1986)

• an approach (Rouibah and Ould-ali, 2002; Zeng et al. 2006)

• a set of technologies (e.g. Negash, 2004; Van Beek, 2004).

Zwanenburg (2008) also acknowledges that the forms Tool, Discipline and Philosophy exist but are very rare. However in practice many people such as IT experts and software companies see BI as a tool. Also Gilad, B. and Gilad (1988) demonstrate that BI is often seen as a tool for just a certain domain, such as competitive intelligence, while in fact BI spans a much larger domain. Besides this there are consultants and implementing parties that see BI as a discipline. As an example see Figure 10 below that shows the disciplined form of an ICT company called Lancet regarding BI Implementation (Mattran, 2009). This figure displays a clear discipline as how Lancet executed BI-projects. Steps as envision, plan, develop and deploy that are supported by project, quality and risk management are part the discipline project execution. The other four disciplines in their disciplined view of BI are:

• “Business integration - The goal is to foster collaboration between end-user organizations and IS and to build an invincible bond of trust between information consumers and the capabilities offered by the BI program.

• Program management - drives and defines the activities of the other four Disciplines.

• Operations & Service level management - all the activities involved with running the components that the project teams develop.

• Architecture & technology - An architecture is a style, a repeated pattern, a set of rules and guidelines, and it might imply a usage of certain materials.“

(28)

Figure 10: Lancet 5 disciplines of business intelligence (Mattran, 2009)

There is also literature that sees BI as a philosophy. For example a quote from Kudyba et al (2001):

“The BI Spectrum: Data Extraction & report Writing, OLAP, Intranets and the Internet The above process has evolved into a philosophy referred to as “business intelligence.”

Therefore all three forms: tool, discipline and philosophy will be added to BEM.

5.5.2 Goal

According to Zwanenburg (2008) the sub-dimension goal can have different aggregation levels.

Nevertheless he identifies the following four sub-sub-dimensions for the sub-dimension goal:

• Decision support

• Decision making

• Detect signs/signals

• Acquiring knowledge

This is mainly based on a quote of Herring (1988:5):

(29)

“[BI systems] must serve four basic functions: (1) support the strategic decision-making process of the corporation; (2) provide early warning on opportunities and threats; (3) provide competitor

assessment and tracking; and (4) support the strategic planning and strategy processes.”

Other authors like Thierauf (2001:4), Huber (1984), Den Hamer (2005:3) and Rouibah et al

(2002:134) support one or two of the sub-sub-dimensions instead of several like Herring (1988:5) and Zwanenburg (2008). Not setting goals regarding BI or pursuing a limited set of goals reduces the BI exploitation possibilities. Therefore the sub-dimension goal is a part of BEM (Zwanenburg, 2008).

5.5.3 Level

BI can be used on several levels within a company. The sub-dimension level consists of the following sub-sub-dimensions:

• Strategic

• Tactical

• Operational

Quoted from Zwanenburg (2008):

“Many authors think the output of the BI process should contribute on a strategic level. Some believe BI plays an important role in making information technologies useful for higher level management.”

It becomes clear some authors limit their scope of BI to the strategic level. However many

opportunities that BI offers on the other levels, such as operational efficiency are lost. An example of operational efficiency is running a machine on the best possible maintenance schedule based on the information gathered in the BI system.

Others such as Pearse (1976:123) Strague et al (1986) and Quinn (2006) are focused on strategic, tactical and operational aspect of the BI level. Figure 11, shows the levels as a pyramid in an organization. Where in the figure the middle level is called analytical the word tactical is also often used to describe the middle level between strategic and operational.

(30)

Figure 11 Level BI (Quinn, 2006)

In recent literature the focus has shifted on how to successfully combine these levels in one BI facility and to allow them to work together and reinforce each other (Imhoff, 2006; White, 2006). This is a challenge as the goals and information requirements per level vary as figure 12 below shows.

Figure 12: Comparison strategic, tactical and operational levels (Imhoff, 2006)

It is likely that a company that uses BI is focused on one level or does not use all the levels efficiently.

Not using the other levels or limited use of the levels restricts the possibilities of BI and the extent of BI use.

(31)

5.5.4 Set of activities

Zwanenburg (2008) defines six required activities (Rodenberg, 2004; Thierauf, 2001; Pearce, 1976;

Huber, 1984) regarding the fulfillment of the BI goals. These are:

1. Aiming for data - Identify the required data.

2. Gathering data - Seeking and collecting data, which is often done by an ETL tool that extracts, transforms and loads the data into a data warehouse from the original sources.

3. Analyzing data - Regarding the extent of BI use it is essential that the data is analyzed and for example not only gathered. Pirrtimäki et al. (2006:83) claims that all perceptions regarding the definitions of BI: “include the idea of analysis of data and information”.

4. Disseminating information - Get the information where it is required, via email, overviews, dashboards, reports, internal/external information networks etcetera.

5. Using knowledge. - This step is vital. If companies have the information to act accordingly but fail to act accordingly all previous steps have been in vain.

6. Evaluation - According to Zwanenburg (2008) most authors state that evaluating the decision making process is part of the BI process. However evaluation should go well beyond this. It should envelop each of the activities named before. As each activity dependents on the other activity it is vital that they are balanced. Therefore evaluating each activity is also vital.

To actually be able to adjust BI activities to changing environments and to improve and extent the use of BI, these six activities can be seen as a life cycle as shown in figure 13, on the next page.

(32)

Figure 13: BI activities lifecycle

Regarding the extent of BI it is vital that each of the activities is in alignment. If for example a company spends a lot of effort analyzing the data with advanced data mining capabilities, but if the company unknowingly: aims for useless or incorrect data, gathers a very limited amount of data, is not disseminating the information to the right sources or the information is not used, then all efforts in the analyzing phase are ineffective. Therefore regarding the extent of BI use the better the activities are in alignment combined with a fitting effort from the company the greater the extent of BI use will be. If one activity however suffers from a reduced quality or attention, it is likely other activities will also suffer. Therefore these activities play a role in the extent of BI use and are part off BEM.

(33)

5.5.5 Drive

The sub-dimension reactiveness drive and research drive are both focused on the drive which companies use regarding BI.

Reactiveness drive

For the sub-dimension reactiveness drive Zwanenburg (2008) describes the following possible drives:

• Reactive

• Proactive

According to Zwanenburg (2008) only reactive and proactive stands are often named in literature.

Reactive can be seen as reacting to an event after it happened. This could be to take measures to solve or reduce the chance of problems or testing hypothesis in which a certain stand is taken. If no event has happened yet and BI is used to find information, this is called a pro active stand. An example of this is to take measures solving problems that might occur but actually have not occurred yet.

To be able to be proactive a company needs at least the capabilities to be reactive. Therefore there is a relation between these two sub-sub-dimensions.

Research drive

For the sub-dimension research drive Zwanenburg (2008) describes the following possible drives:

• Explorative

• Descriptive

• Prescriptive

Even though the literature is not extensive about BI and explorative, descriptive and prescriptive uses it is very likely that in practice BI is used for these purposes. It is possible BI is used simply to gain information about certain topics (explorative) or to get a clear picture of a concept (descriptive) or BI can be used to set rules or find/give directions (prescriptive) etcetera. If a company for example does not use the explorative possibilities of BI it is limiting itself in the possible extent of BI use. Therefore these drives play a role in the extent of BI use and are added to BEM. If an employee of a company is using BI for descriptive purposes than in some phase before this he or she has already been

explorative. To be prescriptive both explorative and descriptive phases would be required. Therefore these values are related as displayed in figure 14, on the next page.

(34)

Figure 14: Phases research drive

5.5.6 BI-ambition

Van Beek (2006) speaks of four levels of BI ambition which are:

1. To understand 2. To coordinate 3. To improve 4. To innovate

The highest level of BI-ambition is to innovate. On this level products and services are easily expanded (Huber, 1984). On the improve level of BI-ambition the company processes will be reformed and rearranged allowing optimization of the entire system. On the level of BI-ambition called coordinate, an organization coordinates tasks more efficient and effective. On the lowest level of BI-ambition called to understand, a company gains insight in their own business and relations between processes such as their value chain and turnover times (van Beek, 2006). The extent of BI use depends on the level of BI-ambition in a company. If a company uses BI to innovate, it is using BI much more extensive than a company that uses BI just to understand their own business (Galbraith, 1977).

(35)

5.6 Data form

The dimension data form is focused on the form the data is available to the users. The data form is important to the extent of BI use. If the data form is inappropriate for the BI users, this directly affects the usability of the data and therefore the extent of BI. Examples are that the users are presented with information they cannot read or understand or that the information is received to late and thus after when it was needed.

Data form has the following sub-dimensions in BEM:

• Language barriers

• Data presentation

• Data availability

• Data accessibility

• Data time

• Data type

• Data source

5.6.1 Language barriers

Some examples of language barriers can be found in healthcare (Bowen, 2001). Nevertheless this phenomenon is not limited to healthcare only as it also plays a role in other companies. For example a location in the Netherlands stores some information in Dutch which also (unknowingly) provided useful information for establishments in France. However they cannot read it because they do not understand Dutch. This is still only the internal source of data. When looking at the external source of data for example the internet, language plays an even bigger role. Just think of your own personal experience, how many times did you end up at websites with a language you do not understand, when looking for information. This is not surprising as the information in figure 15 and 16 on the next page show:

(36)

Figure 15: Top languages internet users (Miniwats Marketing Group, 2009)

Figure 16: Languages on the internet (Miniwats Marketing Group, 2009)

Most people do not speak all of the top 10 languages and especially not all the other languages on the internet. Therefore a lot of information in these languages is not usable, limiting the extent of use of the external sources such as the internet and thus also limiting the extent of BI use. Using translation software or other tools within the company could improve the use of this otherwise unreadable information and the extend of BI use. Therefore language is added as sub dimension to BEM.

(37)

5.6.2 Data presentation

Chung et al (2005) and Hao et al (2000) show the importance of data presentation in the use of BI. In practice the useful visual presentation is challenging as the following quote from Hao shows:

“There are several challenges to providing useful visualization for business intelligence applications.

First, these applications typically involve the navigation of large volumes of data. Quite often, users can get lost, confused, and overwhelmed with displays that contain too much information. Second, the data is usually of high dimensionality, and visualizing it often involves a series of inter-related displays. Third, different visual metaphors may be useful for different types of data and for different applications.”

Therefore the presentation of data can hinder the usability of BI and therefore the extent of BI use.

This means it is part of the BEM as sub-dimension. An example of an overly complex and

overwhelming data presentation is given below. Unclear labels, too many dimensions, too much data, too many relations, etcetera all add to create an overly complex presentation of the data(Soukup et al, 2002).

Figure 17: Complex visualization 5.6.3 Data findability

Data findability means that the data the users require is available and findable from the BI facilities. It can be that the BI facilities are loaded with information and perhaps even the information the users require but: because of the extensiveness of the data they cannot find it, or because of limitations in the tools they use they cannot find it, or because it is available in an unknown place, or fully hidden, etcetera. This means data findability can be seen from different perspectives. Firstly it can be the case

(38)

that the required data simply is not available in the BI facilities available to the user. Secondly it can be that the data is in the BI facilities, but for some reason(s) is still not findable for the users. Finally it can also be the case that the data is findable for the users to some extent. Despite this, in all cases the findability of the data plays a role in the extent of use of BI. Therefore data findability is added to BEM.

5.6.4 Data accessibility

Negash et al (2003) states the following:

“5. Who uses BI?

Business intelligence is used by managers throughout the firm. At senior managerial levels, it is the input to strategic and tactical decisions. At lower managerial levels, it helps individuals to do their day-to-day job.”

As it helps individuals to do their day to day job this means BI can be usable for everybody and therefore access to BI facilities should be as extended as possible. For example a machine maintenance employee could, instead of monitoring 1 local machine at the time, monitor many machines at the same time from different places with the use of BI. Naturally only when the necessary information is available will BI become useful. Like the machine maintenance employee there could be many users within the company that may gain by using the BI facilities, but simply do not have access to it. Or there can be users that use the BI facilities but their limited access rights limit their possibilities. In some cases the required investments to make BI useful for some may simply not be worth the advantages. Besides this there is also the aspect of openness of information which has been an issue for debate. Some claim data should be as open and accessible as possible as it creates transparency.

Notwithstanding this, many still disagree with this (Weitzner, 2004). Therefore data accessibility determines the extent of BI use and will be added to the BEM.

5.6.5 Data time

Zwanenburg (2008) describes three types of times in which data can be found which are: past, present and future. Data from the past can be balance overviews from the last years. Another type of time is present. This can be overviews like the current balance or an overview of a current production run.

Future data are predictions based on information from the present and past such as sales forecast etcetera (Negash et al, 2003; Azvine et al, 2005). If a company only uses data in 1 or 2 data times, for example past and present it misses opportunities offered by data based on the future. In this case their extent of BI use is limited and can be expanded. This means that Data time plays a role in the extent of BI and is therefore added to the BEM.

(39)

5.6.6 Data type

Zwanenburg (2008) names structured data, semi-structured data and unstructured data as types of data (Negash, 2004). Structured data are organized in semantic chunks (entities) which are linked together by relations or classes in a group. Entities from the same class have the same description (attributes) (Wood, 2009). In data warehouses and databases data are often structured in some form, for example as a star structure as displayed in figure 18 below.

Figure 18: Database star structure

Semi-structured data are based on an attempt to combine the two worlds of documents and databases.

An example of this is XML (Extensible Markup Language). Semi-structured data are organized in semantic entities and similar entities are grouped together. Order of the attributes is not necessarily important and not all attributes may get a value. The size and type of the same attributes in a group may also differ. An example of semi-structured data is given next (Wood, 2009).

<?xml version="1.0" encoding="ISO-8859-1" ?>

<email>

BEM - Business Intelligence Expansion Model