TECHNOLOGY VS BUSINESS NEEDS IN BUSINESS

INTELLIGENCE PROJECTS

Ana-Ramona Bologa, Razvan Bologa

Faculty of Economics Cybernetics, Statistics and Informatics, Academy of Economic Studies, Bucharest, Romania

Adela Bara

Faculty of Economics Cybernetics, Statistics and Informatics, Academy of Economic Studies, Bucharest, Romania

Keywords: Decision Support Systems, Business Intelligence Projects, ERP systems, organization changes, critical

success factors, evaluation criteria.

Abstract: This article presents the development of a real industry business intelligence project, in a company that used

an ERP system. We have focused on the main factors that influenced and affected the project development

and it analyses the system evolution from technical point of view, the solutions found to fulfill the changing

requirements, but also the impact of the company reorganization on the project. This particular experience

can be of use to all those who are involved in building business intelligence solutions to reveal some of

success factors, to prevent or to solve some of the inherent problems related to this type of projects.

1 INTRODUCTION

Many organizations use modern ERP systems for

transaction processing and reporting. Information

from applications within an ERP system is managed

by a relational software database. Managers and

executives need information for strategic and tactical

decision that often requires the combination of data

from ERP and non-ERP application sources. The

ERP’s database complex schema is not adequate for

data analysis and reporting. Business Intelligence

technologies can provide the best solution for

extracting, transforming and presenting the huge

information from within the database.

The specific aspects may come from the way of

implementing the BI project. In our case, the BI

project was implemented by the same developer that

implemented the ERP system, and it was strongly

related to the ERP project implementation, which

continued to be extended along with our BI project.

This paper describes a BI project experience

developed in the context of implementing Oracle

ERP solution into a big company activating in

Romanian oil industry services, with a complex

organizational structure, having over 15000

employees. We’ll analyze the difficulties and the

problems met as the project was developed, and

finally we’ll formulate some lessons learned of use

for future implementations.

2 BI ENVIRONMENT: THE ERP

SYSTEM AND THE BUSINESS

REQUIREMENTS

In (Larissa T. Moss, 2001) we can find out that

60%-70% of business intelligence (BI) applications

fail and these failures are not related only to the

technology but to organizational, cultural, and

infrastructure issues.

The ERP project apparently started on a solid

ground, with Oracle Financial (FI) module

implementation at the central unit. The ERP project

then continued by implementing Oracle FI in all

company branches, sectors and working points.

The implementation generated the need for

processes reengineering and revealed important

flaws of the company structure and management. In

order to complete these tasks, many changes had to

be made. These changes affected company structure,

business information and material flows and the way

they were registered by accountability, management

460

Bologa A., Bologa R. and Bara A. (2008).

TECHNOLOGY VS BUSINESS NEEDS IN BUSINESS INTELLIGENCE PROJECTS.

In Proceedings of the International Conference on e-Business, pages 460-463

DOI: 10.5220/0001910504600463

 SciTePress

and reporting hierarchy. It was also considered

implementing the Logistics and Work over modules.

In this context, the BI project started when the

financial module had already been implemented in

the central quarters. In the beginning, the project

seemed to be a classical one, and we followed the

classical phases described in (Larissa T. Moss,

Shaku Atre, 2003): Planning, Business analysis,

Design, Construction and Deployment.

During project stages we took into account some

specific in modeling BIS (Business Intelligence

Systems), such as:

• BIS are oriented on business opportunities

rather than transactional needs;

• BIS have to implement tactical and

strategically decisions, not only

departmental or operational decisions;

• BIS analysis is focused on business needs.

This stage is the most important of the

process;

• Development process is cyclical, focused on

evaluation and improvement of successive

versions, not only building and major

delivering of a singular final version.

In order to respond to all business needs, the

phases of the project were completed almost entirely

twice: for the initial company structure and for the

modified company structure. During this process,

the BI team faced three important types of problems:

1. Technical Problems – the initially allocated

resources were reduced, as the initial estimations of

the data volume were overcome as soon as the BI

module was implemented in the entire company. It

was estimated that a virtual data warehouse would

be sufficient, but the reporting and analysis needs

were higher, so the performances of the initial

solution had to be somehow improved.

2. Communication Problems – there was a lack

of availability from the company personnel, as they

were already overwhelmed by tasks related to the

ERP implementation and by activities related to the

consultancy companies’ cooperation.

3. Organization changes Problems – the BI

project was from the very beginning under the sign

of change. Beside the changes due to technical

aspects, drastic structure changes in the company

negatively influenced the relevance of the initial

business analysis, and often made some of the

developed reports irrelevant or inapplicable.

The work was performed with gaps, as there was

always some top management decision to make that

affected the BI project. The initial plan was revised

over and over, so it somehow lost its power as

management tool. The entire structure of incomes

and outcomes categories and their corresponding

accounts were changed in the middle of the project.

In the following sections we will focus on the

technical and organizational structure aspects.

3 BIS’S DEVELOPMENT PHASES

The system described in this article includes the use

of the following technologies: data warehouse

realized with aggregate data and virtual model,

OLAP technology, data mining facilities, analytic

SQL functions. The interface is based on Oracle

Business Intelligence Portal that integrates all these

technologies and offers direct access to the system

from anywhere, anytime.

First Phase

The cost factor was decisive when deciding

which of two options to adopt:

• A traditional data warehouse, having aggregate

data, involving an extended data volume, both

elementary and aggregated.

• A virtual data warehouse, which worked

directly with data from the ERP database, much

easier to implement.

For this phase, we used Oracle Discoverer to

build a virtual data warehouse over the database.

Multi-dimension structures like dimensions and fact

tables were transformed and mapped from relational

sources. We started with low data volumes,

generated by the use of the FI module at the central

quarters. Oracle Discoverer Desktop was installed

on the computers of the main decision makers that

were going to use it. We developed reports for

analyzing Cash Flows, Budgets, Costs, Expenses

and Incomes only for the central head quarters. The

number of rows involved in these reports was in

average 100000/month, so we did not need a large

amount of resources.

Second Phase

When the financial module was installed over

the entire company, we faced the problem of having

to go to each of the sectors to install the product and

instruct the managers how to use it and some other

IT persons how to ensure the maintenance and future

development. Inn order to minimize the effort, we

integrated Oracle Discoverer with Oracle

Application Server Portal 10g Release 2, and the

solution became available on-line, allowing

everybody to access information according to their

established privileges. But, this decision had a

negative impact on performances: the working speed

TECHNOLOGY VS BUSINESS NEEDS IN BUSINESS INTELLIGENCE PROJECTS

461

decreased as the required memory space increased

from 32 GB server shared as Discoverer needed, to

64 GB server shared as Oracle Portal needed.

We didn’t manage to get a dedicated server for

Oracle Application Server because of high hardware

resource cost. In this phase, we had to develop

reports for analyzing Cash Flows, Budgets, Costs,

Expenses and Incomes for the central head quarters

and ten sectors with different area of production. So,

the number of rows involved in these reports was

over 1, 5 million/month. Under the circumstances,

with low allocated resources and a large amount of

data per month we needed to find another solution to

increase the speed and reduce the loading time for

reports.

Figure 1: The system’s architecture in the final phase.

Third Phase

When the other ERP modules started to be

implemented, we were facing serious problems. We

decided to construct an enterprise data warehouse

integrating Oracle Data Warehouse Builder and

Oracle Data Mining into Oracle Portal. Unlike

Oracle Discoverer, in Warehouse Builder data are

physically stored into a special repository, build on

the database. The role of Oracle Data Mining was to

supply additional analysis power.

This solution was a success as data were

aggregated in the central data warehouse, the ETL

(Extract, Transform and Load) process was built off-

line, not affecting the reporting analytical time. We

built new types of reports related to production

(costs and allocated resources), work over,

purchasing orders, transactions per account, etc. We

designed one central data warehouse with sub-data

marts for financial, production, purchasing orders

and order entry modules. The final solution is

described Figure 1. The number of rows involved

was now over 3 millions/month, and the

performances were finally satisfying.

4 EVALUATION OF THE FINAL

SOLUTION

4.1 Organization Changes Impact

Major business reorganization is one reason BI

project typically get hosed (Mitchell A. Ocampo,

2007). In this project we faced an important business

restructuring. More, the initial scope estimation

proved to be inexact, and it had to be re-evaluated,

so the project gained new dimensions.

Most of the analysis results had to be re-

evaluated, and database design and repository design

had to be revised and seriously extended, having an

important impact over technical and human resource

requirements. The BI team had to be enlarged and it

was itself a problem, as there were no other BI

specialists available, and other two new persons had

to be employed and integrated.

Many of the reports developed in the first phase

were inapplicable, generating important time loss

and inefficiency; it often happened that after getting

complete specifications, data gathering, and

completing of the report design, during the

validation meeting we were communicated the

whole perspective was changed. This had impact not

only directly upon efficiency, but our team members

got to be confused or demoralized.

But, there was a bright side in all those changes.

As the consultancy company progressed in

elaborating the new business view, they inevitably

brought out the reporting needs. Our team was

invited for discussions, and we had the chance of

proving the great advantages of BI over the existing

reports. A demonstration based on the customized

reports had a relevant impact.

There was also a change in the attitude of end-

users, previously reluctant to changing their old

tools (Microsoft Excel and other Office applications)

and to learning how to use a new one. They were

now eager and motivated as the old reports were no

longer valid. And the truth is that they come to love

it in very short time, as the new reports were very

flexible and easy to use and they interfaced directly

with the ERP application and other data sources.

ORACLE DB

ORACLE E-BUSINESS

SUITE

DATAWARE HOUSE

ORACLE DATA

WAREHOUSE AND

ORACLE

DISCOVERER

DATA MINING

ORACLE PORTAL

BI REPORTS

ORACLE APPLICATION

SERVER

ORACLE PORTAL

ICE-B 2008 - International Conference on e-Business

462

The end-user involvement is a very important

factor in developing functional BI solutions. As

project requirements and scope are set by users, they

know exactly the data they need and they will test

and use the result of the project. There are voices

claiming there was a shift of authority and

ownership in BI projects, from IT to business side

(Larissa T. Moss, Shaku Atre, 2003).

4.2 Criteria for Evaluating Business

Intelligence systems

Deploying BIS involves many risks: system design,

data quality, and technology obsolescence. Large

budgets and strategic information are at steak in

deploying BIS systems, a good reason to establish

rigorous criteria for evaluating this type of system.

These criteria are discussed below.

1. Decisions based on Business Process

BIS should not be viewed only as a data repository

or a large set of data. Instead, system’s

implementation should be concern with

conceptualizing new data models, processes, and

indicators that form the content of the BIS.

2. Performance

It reports to the response time the system provides

to its users, depending on the complexity of the

database and the queries being requested. Most

responses should range from few seconds to a

maximum of 30 seconds for routine queries.

3. Flexibility and Scalability

Flexibility determines whether a BIS solution can

adapt to changing business conditions after the

system’s delivery. BIS should be able to

accommodate changes in any type of business

processes and functions, as well as in new laws and

regulations, requiring the capture of different types

of data. BIS should be expandable to accommodate

data growth and changes in organizational structure,

without a slowdown in performance.

4. Integration

Integration involves two types of issues: data

integration and system integration. BIS will be

particularly effective if it can overcome the

challenge of information fragmentation, allowing

managers to measure features of business processes

that involve information from inside and outside of

the organization. System integration involves two

aspects: the ability to extent the BIS software with

new capabilities and modules and the system’s

ability to coexist with other enterprise solutions.

5. Friendly user Interface

BIS should be designed to provide managers a fast,

easy, and understandable way to navigate into data

and identify trends and patterns. BIS should allow

the user interface to accommodate different degrees

of technical knowledge.

We notice these criteria are related not only to

technical aspects, but also to business aspects, as

managers offer the most important feedback to the

quality of business intelligence systems.

5 CONCLUSIONS

The most important conclusion of this real project

experience is that, more than technical reasons,

organizational and infrastructure dysfunction

endanger the success of a project. When there are no

supporting business decisions, the BI project is

paralyzed. So, the main factor in conducting such a

project is business rather than IT. Following this

direction, after our project implementation was

ready, the maintenance and further development of

BI reports were entrusted to some end-users,

specially trained (non IT).

Only combining BI technologies on the top of an

Enterprise Resource Planning system will never

provide a viable response to the changing

management needs. A competitive advantage can be

obtained only by adding business-oriented

personalization so that the information provided can

fulfill the particular needs of end-users and empower

dynamic analysis and decisions.

REFERENCES

Kaniclides A., Kimble, C. 1995. “A Development

Framework for Executive Information Systems”,

Proceedings of GRONICS '95, Groningen, the

Netherlands, Ed T LOURENS;

Larissa T. Moss, Shaku Atre, 2003. Business Intelligence

Roadmap. The Complete Project Lifecycle for

Decision Support Applications, Addison-Wesley

Professional

Larissa T. Moss, Organizational Barriers to Business

Intelligence (Part 2), http://www.businessintelligence

.com /ex/asp/code.135/pagenum.1/xe/article.htm

Larissa T. Moss, 2001. Organizational and Cultural

Barriers to Business Intelligence, Cutter Consortium,

Executive Reports, http://www.cutter.com

Mitchell A. Ocampo, 2007. Overcoming the Barriers to

Business Intelligence Success, EMC Corporation

Publications, http://www.emc.com/solutions/microsoft

/pdf/H2740_overcom_barriers_bus_intell_success_ep_ldv.

pdf

TECHNOLOGY VS BUSINESS NEEDS IN BUSINESS INTELLIGENCE PROJECTS

463