Have you ever heard about Business Intelligence? Chances are that you have not, this is an aspect of the IT/Business world that has started to grow in the last, let’s say, 10 years or so (in other words, this is relatively new).
The idea of Business Intelligence is to drive your core business based on the results of the past.
In other words a business intelligence tool allows you to create reports based on data that is available within your company. These reports can be lots of things, a line chart showing the evolution of your sales, a pie chart shows how many service requests there are of which types, a simple table giving you a small overview of which employees haven’t worked enough hours this week. Everything you want to analyse is possible with a BI-tool.
Typical about the business intelligence setup is that it most of the times runs on either stars or cubes. Sounds childish yet fancy, no?
Let me explain, the idea is rather simple.
Imagine you having a company that has an application that allows you to input data. So for example, you have a system that allows you to enter customer data, sales data, supplier data. This system is built to input data in the most efficient way and will be most likely an ‘OLTP’ system. Online transaction processing. This system has a lot of different, smaller tables which are joined with each other based on different conditions. In theory each table can be joined on every other table in this database.
Business intelligence on the other hand should be able to extract data in a fast and efficient way. So the system mentioned before will not do. In order to make it performant, the data needs to be transformed to either a star or a cube. A star schema is a type of setup which transforms your original tables to facts and dimensions.
A fact is a table which typically contains a few millions records. Because this type of table contains all the information concerning a certain ‘fact’ (which figures..). All calculations are also done on this table. Amount of sales, number of employees, etc…
A dimension on the other hand is a smaller table which contains attributes of a certain subject. Like the name and address of a customer.
These dimensions are all linked to one or more fact table. This means that the fact table linked to the dimension should contain a reference to the dimension. Else it would be impossible to join both table together.
This means that the fact table contains all combinations of information that are available in the OLTP system of a certain fact.
The process of changing the data to a star model is called ETL: Extract Transform and Load.
As I said, business intelligence tools can also used cubes. This is a database object that contains 3 dimensional data. A dimension or a fact is only 2 dimensions, it only has columns and rows. A cube however has walls, which contain more than just rows and columns. They are a pre-calculated set of data about a specific topic.
So a cube can for example contain all sales data for all customers for all products of the past 2 years.
If you want to see data for more than just these 2 years, you’ll have to define a new cube.
This all sounds rather technical, and frankly it also is. But if you are a user of a business intelligence tool, you don’t need to know all of this technical gibberish. If it is set up correctly, you should be able to create reports without ever having to know anything about facts and dimensions.
But I can almost hear you thinking. Why would I use a thing like this? The answers, again, is very simple. If your business intelligence environment is setup correctly and you use the BI-tool correctly this can spare your company a lot of money.
Think of this very simple example. If you create a small report which contains the sales per month per sales representative you can easily track which sales representative is doing a good job and which are not.
You can also track changes in the sales they did. If you see that a certain representative has sold a lot loss this month in comparison with the previous month, you can take actions. This can help your business to grow.
There are quite a lot of business intelligence tools out there. A few big names: Oracle Business Intelligence, Business Objects, Qlickview, Cognos, … .
Most of these products are meant for bigger companies, hence they also cost quite a bit of money (I kid you not, we are talking about a few 10k dollars here). But no worries, there are also smaller products which are more affordable for smaller companies. You even, like almost everything these days, have a few open source BI applications.
I just wanted to share this one more thing with you. The most logical steps in the BI process…
Step 1: Buy BI Tool
Step 2: Analyse
Step 3: Profit
Step 4: Sell company, give me a share, have party in Ibiza.