Framework to design a business intelligence solution

Published: Last Edited:

This essay has been submitted by a student. This is not an example of the work written by our professional essay writers.

Abstract. In the present research we propose a framework to design a business intelligence solution based on the integration of business and technological domains. The main contributions of the framework are: 1) enterprise architecture that combines the approach of the Zachman Framework and the Balanced Scorecard, 2) mapping the System Model layer with the tools of the Pentaho BI Suite and 3) an indicator for the control management that facilitates to measure the performance of the strategy from the compensation of the indicators defined in a Balanced Scorecard, based on compensatory fuzzy logic.

Keywords: Business Intelligence, Enterprise Architecture, Compensatory Fuzzy Logic, Balanced Scorecard.


The business intelligence (BI) have the capability to take the information flow, that the organization gathers every day and transforms it into active information that allows to improve the decision making process in order to assure the success of an enterprise. Today the companies invest a lot of money to buy the best applications that allow them to develop BI solutions. Nevertheless, these applications do not guarantee, that all the necessary information for the decision making process is available. Most of the existing solutions are concentrated on the technological capabilities, and designed to answer the question of, how to achieve the solution. They are not answering what would be the necessary information that the solution must support, in accordance with the real needs. This problem occurs because of the lack of alignment between business and technological domains.

A survey of 385 finance and IT executives, by CFO Research Services (Fig. 1), asked them to identify the drivers for poor information quality (IQ). Nearly half of them pointed (45 percent) the non-integration of IT systems and the variability of business processes as an acute problem that constrains management's ability to work effectively and focus on high-value activities. Approximately the same number agrees that finance and business units alike spend too much time developing supplemental reports and analysis. Other disappointing and productivity sapping by products of poor information quality include that "multiple versions of the truth," misguides incentive programs, and leads to unrealistic plans and budgets.

In fact, 61 percent of respondents say they could still do a better job of just making sure the financial information they generate accurately reflects the performance of their businesses.

The business impact of this poor IQ, say respondents, includes widespread decision-making problems that are often tied to inaccurate, untimely, and irrelevant information [1].

The main goal of this paper was to develop a framework that facilitates the design of BI solutions based on the integration of the both: business and technology domains, as a tool that contributes to the improvement of the availability of necessary information in the decision making process.

Business Intelligence

The BI term is not used uniformly by all authors. From its creation in the 90s by Gartner Group, the persons have interpreted it from very narrow up to very wide perspectives, due to the fact that there does not exist any strict scientific definition of the term.

Microsoft Corporation raises that it is the "Aptitude to take the flow of information that every organization gathers every day and to transform it into active information that allows obtaining the success"[2].

On the other hand Gartner Group rises that it is "An interactive process for exploring and analyzing structured, domain-specific information (often stored in data warehouses) to discern business trends or patterns, thereby deriving insights and drawing conclusions. The BI process includes communicating, discovering and effecting change. Domains include customers, suppliers, products, services and competitors"[3].

Marrero Atunez brings up that it is "A set of systems, strategies and informatics tools, whose functionalities are oriented to the support to the decision making process in an organization. Its enhance the availability and timely analysis of key data for the performance of the organization and its proper implementation and use requires a comprehensive vision and strategic operation of the entity, as well as mastering processes and information flows that characterize it"[4].

It is possible to affirm that in spite of a big concepts variety related to the BI term, that exists in the literature, in all of them is included the aptitude to extract information for the decision making process. The principal differences are focused in that, some authors see the BI only from a technological perspective, while others including the authors of the present paper; see it from as an integration of both: business and technological domains. Thus a facilitation of the decision making process in the different hierarchic levels of the organizations is to be made.

Enterprise Architecture

Enterprise Architecture (EA) is a framework or "blueprint" for how the organization achieves the current and future business objectives. It examines the key business, information, application, and technology strategies and their impact on business functions. Each of these strategies is a separate architectural discipline and Enterprise Architecture is the glue that integrates each of these disciplines into a cohesive framework [5].

A study realized in the year 2007 led by the Society for Information Management's (SIM) Enterprise Architecture Working Group (EAWG) was aiming to understand better the state of the practices of the AE in the organizations and to evaluate the state of the capacities of IT of the organizations in means to develop an AE, had the following principal results (Error: Reference source not found) [6].

In a general way, we can summarize that because of the changes of the environment, the manager tries to discover the data, which supports having the general perspective of the business, and he has to be enabled to understand how the parts of the organization are interrelated with each other's.

Design of BI solutions

In accordance with the expressed previously, a BI solution is a finished platform that helps to compile automatically the flow of information, generated in a company, provided, there is a computer network, which has to be capable to transform this data into information to improve the decision making process.

On the other hand, it is important to understand that the BI tools alone facilitate "how" it is possible to achieve a solution from the computer point of view, but they do not assure, "what" is the information that is really needed. Often occurs that only a part of the whole the information needed to make decisions is missing. The reason for that is either lack of knowledge of the person requesting it, or because the information is not supported from the organization's database(Fig. 2)[7].

A management system based on the strategy is a tool that helps the manager discovering what is really important in order to achieve the company's goals. The knowledge in which moment the right amount and quality of information is achieved in order to provide the right focus enabling the success of the enterprise.

Finally it is important to know that the enterprise architecture is not a simple description of a collection of documents and plans; it is a model of how the different parts of an organization are related. Without a model of enterprise architecture, the executives, the agents and the technologists are essentially "blind runners", making decisions based on their personal perception of the company. It is often the case, that this picture is not shared by the staff rest of the organization. An enterprise architecture model is a tool which helps the executives to think about the organization as a homogeneous system. The architecture captures a wide information variety, establishes relations between the technological domains and the business domains, and stores all the information joined on a single repository. This is a good starting point for the design of any business intelligence solutions. Thus the manager can make decisions, identify problems and analyze information of the business by means of the efficient integration of the technological domain and business domain.

Proposed Framework

The definition of the framework to facilitate the design of a BI solution is based on the elements expressed previously, as well as on the integration of Zachman Framework, the Balanced Scorecard, and the Pentaho BI Suite.

The proposed framework is based on four phases, which associate different procedures, rules and tools facilitating the framework's fulfillment (Error: Reference source not found).

Phase 1: Preparation

The main goal of this phase is the preparation of the necessary conditions for the change, which the company must confront to assume a project of design and development of a BI solution.

Phase 2: Enterprise Architecture

The aim of this phase is to define an EA for designing the BI solution. The methods and the tools to be used for its fulfillment are defined here too. The proposed framework is based structurally on the matrix proposed by the Zachman Framework[8], considering their first three layers, which are defined as: the strategy model, business model and system model. We considered the necessity of including an interface with the Balanced Scorecard (BSC) between the strategy model and the business model, targeting the translation to operative terms of the strategy of the company, since in Zachman Framework it is not contemplated. The general idea will be to construct a pyramid (architecture), which is incorporating in each of its layers the necessary specifications, to support the top level.

From this structure and according to the development of a BI solution, was deployed how to reach the elements that appear in each of the cells, specifying inputs, outputs and tools to use it (Error: Reference source not found).

Fulfillment rules

Based on the structure defined in the architecture above, which must be firstly completed by layers, provide that each layer represents a top level, with regard to the one that follows it. Nevertheless a big dependency among the elements in the columns is sometimes is strongly sized. In Error: Reference source not found are represented the dependencies among each cell and the order to be completed it.

Phase 3: Mapping the enterprise architecture with Pentaho BI Suite

The purpose of this phase is to map the results of enterprise architecture, specifically in the layer of the system model, with the tools of Pentaho BI Suite.

For the development of the application, the analysts of the system must be based fundamentally on the System Model layer. In

we present the mapping between the System Model layer with the tools in Pentaho BI Suite. It provides the necessary information for the implementation of a solution, that truly meets the information needs of the managers, required for the decision making process. The main responsible for this phase is the technological staff.

Phase 4: Monitoring

This phase has as a goal to design an indicator of management control that allows making a measurement, according the behavior of the indicators of the BSC. This phase facilitates the continuous process in the organization and gives the latter the possibility to adapt itself rapidly to the changes of the environment.

Indicator to measure the performance of the strategic plan

One of the points of major importance in the development of any strategic control tool is the development of strategic plans, through feedbacks, where the indicators capable of monitoring the management control, play an important role. Being it, one of the principal difficulties found in the literature used for this research.

In this sense, we proposed an indicator to measure of management control using compensatory fuzzy logic; it is capable to evaluate based on the behavior of the indicators defined in the BSC.

Step 1: Built the matrix GI(m,k)

To prepare a matrix where all the indicators and the strategic goals appear defined for the organization. The prepared matrix, must be presented to a group of chosen experts who have to answer the question: How true is it, that the indicator "k" is an important element in the measurement of the fulfillment of the strategic goal "m"? The scale to be used would be a continuous scale between 0 and 1; where 0 would be the most false value and 1 the most truthful one.

Step 2: Calculation of value of truth, that the strategic goal "m" is being measured by the indicators defined in the BSC (VGIm)

With the previous information we can answer the question: How true it is, that the strategic goal "m" is being measured by the indicators defined in the BSC?

A strategic goal is being measured if and only if exist indicators that measure it.

This can be expressed using compensatory fuzzy logic as:



VGIm: Value of truth that the strategic goal "m" is being measured by the indicators defined in the BSC.

VGI(m,k): Matrix with the truth value of the expert consensus that the presence of the strategic goal "m" is measuring by the indicator "k".

The people in charge of designing the BSC should be ensuring that the value obtained in VGI (m) for each goal, has a value greater than 0.5.The ideal value would be given by: maximizing VGI (m) and minimize the number of indicators defined in the BSC.

Step 3: Calculation of the value of truth that the strategy is being fulfilled (E)

As a premise, must be ensured that the VGIm value was more true than false for all strategic goals.

A strategy is being fulfilled if and only if all the important strategic goals are being met.

A strategic goal is important if there are critical success factors that justify its approach.

An important strategic goal is being met if and only if all the indicators defined in the BSC for its measurement are being met.

Based on the principles stated above and using compensatory fuzzy logic to compensate the indicator defined in the BSC, given as:



E: Value of truth that the strategy is being fulfilled.

VGm: Value of truth that the element "j" is a key success factor and in turn advises the strategic goal "m".

VGIm: Value of truth that the strategic goal "m" is being measured by the indicators defined in the BSC.

VIk: Value of truth of the criterion of measurement of indicator "k".

To calculate we propose to use the sigmoidal membership function.


S =: Value of truth of the criterion of measurement of indicator "k".

X =: Calculated value of the indicator "k" according to the company.

Gamma (γ): Value acceptable. It would be equal to the value at which the indicator is considered acceptable.

Beta (β): Value almost unacceptable: It would be equal to the pre-image of a symmetric sigmoidal function for the optimal value defined for the indicator, or it would be the same β = (Value at which the indicator is acceptable - Value from which the indicator is optimal).

Alfa (α): Sigmoidal function parameter.


The proposed framework, as well as all the tools that conform it, help to improve the availability of the necessary information for the decision making process, based on the integration of the business and technological domains. Beside as complement to the tools which compose the framework, we included an indicator useful for the measurement of the strategy performance based on the indicators behavior defined in a BSC. It gives a solution to the problem found in the literature according to the absence of management control indicators, capable of integrating in a single value the different results of the BSC and to formulate strategic plans based on the received feedback.


In order to facilitate the application of the specific procedures defined into the framework for the final users is necessary to design and to integrate a graphical interface software package. Beside with the enterprise architecture and the proposed fulfillment rules in the framework, it is possible to define an ontological model to support the improvement of the mapping process between the enterprise architecture and BI tools. Finally we want to design an indicator based on compensatory fuzzy logic to measure the degree of alignment between business and technological domains.