Need for Business Intelligence
INFORMATION AT YOUR FINGER TIPS May 2007
Wisdom InfoTech
18650 W. Corporate Drive Suite 120
Brookfield WI 53045
Phone: 262-792-0200 Fax: 262-792-0202
E-mail: [email protected]
ABRAHAM PABBATHI
Principal Consultant – BI Practice
TABLE OF CONTENTS
BUSINESS INTELLIGENCE ...3
TOP 10 SIGNS YOUR COMPANY MIGHT REQUIRE A BI SYSTEM ...3
WHAT IS A BI SYSTEM? ...4
COMPONENTS OF A TYPICAL BI SYSTEM...5
APPROACH TO BUILDING A BI SYSTEM ...7
Business Intelligence
Over the years with increasing globalization the business landscape has grown larger, more diverse, more competitive and more complicated. To survive in this treacherous landscape, businesses need to be smart, flexible and responsive to changes in the market. Till now businesses have only concentrated on capturing data and using it to run their day to day operations. But of late businesses have been leveraging the data that they have collected over the years to study market trends to help them adjust to the fast changing business environment.
This analysis of stored data to discover patterns and trends is known as Business Intelligence and its delivering rich dividends to businesses in terms of better insight into market trends and the factors affecting the performance of their businesses which in turn gives them a competitive advantage.
Top 10 Signs your company might require a BI System
Inability to access data quickly and easily:
Typically if you want to know the latest sales numbers you need to call your sales manager who in turn needs to dig through a bunch of spreadsheets on her laptop to come up with the numbers. You don’t have easy to use reporting tools.
Inability to perform in-depth analysis:
Although you have a report which gives you the total sales at each of your territories you don’t have an easy way to figure out which products are doing good in each of those territories. You don’t have the tools to analyze your data.
Data scattered over multiple systems:
Your Sales data is stored in an MS-SQL Server database while your production data is stored in an oracle database and you don’t have a way to combine the data from these two databases into one report.
Historical data not available:
You remove all data older than 2 years from your databases. Hence you cannot run a trend report to see how your sales did over the past 5 years.
Lack of data privacy and security:
You don’t have a way to apply restrictions on the data seen by different employees.
You don’t have a way to securely share your data with your customers and partners.
Inability to comply with Government Regulations:
You plan to go public some day but you don’t have the infrastructure to support all the government reporting requirements that you need in order to go public.
Inability to compare performance against goals:
You have set strategic goals for your company but you are unable to track your performance on a daily basis to see if your company is on track to achieving the desired performance goals.
Inability to use Data to facilitate decisions making process:
Your business decisions are based on your gut feeling rather than facts. You are unable to derive information out of the data that you have in your databases to help you in your decision making process.
You are losing market share:
You are slowly losing market share to your competitors. You are unable to manage your resources to the optimum level to achieve maximum efficiency in the way you do your business.
What is a BI System?
Business intelligence (BI) is a Business Management term which refers to the services, applications and technologies that enable organizations to gather, manage and analyze data, transforming it into usable information which helps them to make better decisions.
The example below shows how analyzing the sales numbers against targets helps you to concentrate your energies on the most pressing problems at hand.
Figure 1: Analyzing Sales Number Example
At the core BI provides query, reporting, and analysis capabilities. It provides Dashboards and other visualization techniques which help users to quickly get an understanding of the position of the business. This is complemented by data quality and data integration services which help in consolidating data from multiple sources.
The Figure 2: Dashboard Example below shows how a typical dashboard for hotels business looks like.
Figure 2: Dashboard Example
The Advanced analysis capabilities provided by BI can help users to view data across multiple dimensions (for example: product, geography, customer, time etc.) and slice-and-dice the data to look at various combinations, such as the sales of certain product in all geographic locations in certain year and quickly switch to seeing the sales of all products in a particular geographical location during the same year. Advanced analysis functionality also enables organizations to define hierarchies so, for example, sales could first be viewed for each year and then the user could drill down to view the sales in each month in certain year and consequently drill down to a particular week of a month. All these operations can be done with the click of a mouse button.
These advanced analysis capabilities make it easy to compare the results of one time period with that of another, so this year’s sales can be compared with sales of a previous year across geography, customer, and salesperson. BI also provides capabilities which allow predictive analysis to discover hidden patterns and enable what-if analysis.
Components of a typical BI System
The Figure 3: Components of Business Intelligence System below shows the important components that go into building a robust BI system.
Figure 3: Components of Business Intelligence System
On the left we show your existing database systems which might be in RDBMS, Mainframes or Spreadsheets. The arrows show the flow of data out of the existing database systems into the components of the BI system. Each of the components are sold separately and need to be installed and tied together to create a BI system. Below each component we have mentioned the popular vendors who build that particular BI component. Below is the description of each of the BI components.
Data Integrator: This is a specialized software tool (also called ETL tool) which can extract data from varied systems like RDBMS, Mainframe and spreadsheets and load the consolidated data into a “data warehouse”. ETL tools also take care of “cleaning” the data if the data in the source systems is bad. By “cleaning” we mean removing duplicated data, correcting name/address data and the like.
Data Warehouse: The database where all the important data from an analysis stand point is stored. Selected data from the existing databases systems (OLTP systems) is collected and loaded into this database. This database could be loaded with fresh data weekly, daily or real- time depending on the user’s requirements.
OLAP Engine: OLAP stands for “On-Line Analytical Processing”. Certain specialized software tools called OLAP tools provide the capability to run reports which have slice and dice, drill down and drill through capabilities. Typically these tools would build another database on top of the data warehouse called the multi-dimensional cube (a highly de-normalized database) to allow for quick query response. These reports are used by Business managers to analyze the
Reporting Tool: Reports are built on top of the data warehouse using special Reporting tools.
These tools can provide pre-built reports (canned reports) which could be refreshed as per certain schedule or give the users the capability to build their own reports (ad-hoc reports).
Typically these kinds of reports are required by middle level managers to run the operations.
Digital Dashboard: This is again provided by yet another type of software tool which allows for building reports which are easy to comprehend and allow for visualizing the data in the data warehouse in different formats, for example Charts, Gauges, and Alerts. Typically these reports would report all the key performance indicators of the business on one page so that the executive managers can understand the health of the business with one glance of the report.
Approach to building a BI System
Building a complete business intelligence system spanning over all the departments/business processes of the enterprise can turn out to be an expensive proposition. But it is not necessary that you implement the BI System for the entire enterprise in one go. A good approach would be to implement the system for one department/business process at a time. You add more departments/business process to your BI system as your company grows and expands. This way the cost to build the BI system is more manageable and the lessons learnt during one implementation can be leveraged during the next implementation. Each cycle of implementation will have the below mentioned phases.
Figure 4: Approach to build BI System
Initiate: This is the most important phase of implementing a BI system as it sets the course for the rest of the process of building the BI system. In this phase we would set the course for the BI System by defining a road map. This will be a guide in deciding which business processes would be incorporated into the BI system and in what sequence. Typically we would want to divide all the business processes in the enterprise into 3-4 groups and attempt at building the BI system for one group at a time. This way we would be able to cover all the business processes in 3-4 cycles. Once we have narrowed down the scope for the first cycle to a group of business processes we would identify the users who are involved with the business process who would be the first to start using the BI system. Next we need to survey the market and decide on the
Plan: In this phase we execute the typical steps we would do in any IT project. First we make a project plan followed by Requirements analysis. Once the requirements analysis is done we would proceed with designing of the system.
Build: In this phase we actually build the system and unit test all the components built.
Test: In this phase we conduct various tests like System Test, Load Test and User Acceptance Test on the BI system. Once the system passes all the above mentioned tests, we plan the deployment of the system into production environment.
Deploy: In this phase we install the BI System in production environment followed by user training so the users know how to use the system. Finally we also setup a support team for the application and train them in handling issues which may arise in the system during the course of its usage.
The above mentioned steps are followed for each cycle of implementation of the BI system.