All about Business Intelligence

This article aims to clarify some of the topics related to business intelligence. Starting with the definition of business intelligence.

Business Intelligence
There is no one single agreed definition for business intelligence. Below are the ones that I believe are the widely used definition.

Gartner - BI is an umbrella term that includes the applications, infrastructure and tools, and best practices that enable access to and analysis of information to improve and optimize decisions and performance.

Wikipedia - BI can be described as "a set of techniques and tools for the acquisition and transformation of raw data into meaningful and useful information for business analysis purposes".

Forrester - A set of methodologies, processes, architectures, and technologies that leverage the output of information management processes for analysis, reporting, performance management, and information delivery.

My Version - BI is the process* of deriving information from data*  efficiently to enable informed decision-making in order to improve Business.
  • Process includes technologies, methodologies and most importantly the People.
  • Data refers to all types and from all the sources whether internal to Business or not.  It doesn't matter how small or big the data is, where it is, how it is captured and which type it is.
BI can be considered as the black box to which data is passed as input and the output is useful information based on which decisions can be made to improve Business. 

DATA + BI = Improved Business.
Businesses are not necessarily  commercial businesses, it could also be non commercial activity like police department, not-for-profit organizations, charity hospitals, etc.

Why Businesses need Business Intelligence?

In short, to measure and improve business. As someone rightly said, If you can't measure it, you can't improve it.
  • To provide right information to the right person at the right time.
  • To understand customers, employees, partners, suppliers, distributors and competitors. Basically to understand all parties involved in the business.
  • To ensure all parties involved in the business get the same information.
  • To make fact based decisions.
  • To be compliant with legal and regulatory requirements.
  • To minimize expenses.
  • To stop revenue leakages.
  • To find growth markets/products/services
  • To increase revenues
BI to business is equal or more important than a dashboard for a car. The more intelligent the car dashboard is, the better results you get. Can a car be driven without a dashboard? Yes. Can you drive better when you have a dashboard? Yes.  This is exactly what BI does to business. You can obviously run a business without BI. However with BI you can run it better. And in this competitive world where you need to run better than your competitors you definitely need BI. Like the dashboard of the card indicates what speed the car is going, What is the optimum speed, in how much time you will reach your destination, which route to take, which road to avoid, how many kilometres you have driven in this trip, what is the total number of kilometres driven till date, etc. BI indicates how the business is performing, how will it perform in the future and what changes should be done to achieve the objectives of the Business.

Should a BI Solution always have a Data Warehouse?
Short answer is no. A BI solution does not always have a data warehouse.

A strategic and tactical BI solution most often  has a data warehouse in the backend of the BI solution. This is because data from the transactional systems needs to be historized and versioned.  Operational BI solution most often do not have a data warehouse. Examples of operational BI solutions without data warehouse are BMC Remedy Analytics for BSM and JIRA dashboards. In these examples the BI reports are directly based on the transactional databases or mirror databases and data is not historized like in the case of a data warehouse. A BI solution with a data warehouse can answer 100% of the questions whereas a BI solution without a data warehouse will fall short in answering historic and versioned data based questions and hence will also fall short in providing predictive and prescriptive BI.

What exactly is the difference between Data Warehouse and Data mart?
Unfortunately there is no simple answer. Because the answer depends on two main points; 1) Building approach 2) Data stored

Building approach:
If bottom up approach (Kimball methodology) is  used  for building the data warehouse then data marts are the building blocks of the data warehouse. If top down approach (Inmon) is used for building the data warehouse then data marts are the subsets of the data warehouse. 

Data Stored:
For explanation purpose lets consider that both data warehouse and data marts are types of data stores that are at the backend of the BI solution. Data warehouse stores or plans to store data from multiple subject areas whereas data mart stores data of only one subject area. Example - A data warehouse may contain data from various subject areas like customers, employees, distributors, sales, finance, purchase, marketing, HR etc., in the data store whereas a data mart contains data only from one subject area for example a data store based on job application system that is used by HR team.

Whether a data warehouse gets built or data mart gets built largely depends on who owns the BI solution? Is it a enterprise level initiative or a department level initiative? In general if it is a enterprise level initiative then a data warehouse gets built and on the other hand if it is a department level initiative then a data mart gets built. Exceptions are definitely there.  A department level initiative can also result in a data warehouse if the data store stores multiple subject areas.

For further reading on BI topics click on the BI page.


Popular posts from this blog

ETL developer vs Data engineer

KABI - The new Agile Methodology for BI Projects - Implement BI projects quicker happily

BI Architect course and BI Tool question