Business Intelligence (BI) is just a fancy term for reporting. If I ask a business user what a report is, he/she will be able to bring up an image in their head of what it is and explain it as well. To make it easier to understand BI, just consider it as a reporting mechanism. That begs to the question why we have a fancy term like BI for a simple concept like reports?
Well, consider a simple scenario to explain the subtle difference between the two. The head of a business is looking at a report about the state of his business as of last night. He/She got the report the first thing in the morning based on the closing data of last night. That’s exactly what a report from an ERP system is. Now there might be a situation that the business user wants to look at the data from a previous date. A quick call to the Administrative Assistant and you get the hard copy of the report from a previous date. With a content management system, even that won’t be too difficult with a digitized copy of the report saved perpetually.
An operational system for reasons of efficiency holds only the current state of the business. In a retail store example, the invoice issued will have its own unique number and it will always stay the same. Any changes made to the invoice will just update its data and lose its previous information. However, since the previous state isn’t relevant to any analysis that is the sort of data loss that is perfectly acceptable.
So now the user wants a complete history of a particular customer. That’s easy, they can just pull up each invoice that’s been issued against the customer. So, if you need a breakdown of invoices in a specific region or by customer segment or by value or all invoices in a region by a specific segment of customer that contain your Product X, then you have to get all those various combinations custom-made each time. That is when you realise that the IT department will kill you if you send in the request for all the various reports you want. Too bad, they’ll just have to do it. It’s all good if you’re the President because then you’re requirements get top priority but in the world of middle management you have to share with other middle managers and wait until the IT people get back to you. This is where it gets interesting. Since the reporting tool isn’t something the user would know or even be inclined to know, then they are solely dependant on IT to deliver.
BI is a step towards the Holy Grail of information. An ERP or operational system is basically loaded with data but there’s very little meaningful information you can get from it. It does give you an exact view from an operational perspective, but if you want to do even basic trend analysis or want to view data from a completely fresh and different perspective then the process is so excruciatingly slow that one just gives one.
BI based on a properly modeled reporting database takes you to newer heights of analysis. Since all business concepts have been modeled and the data from the ERP system has been transformed into the reporting model, the BI tools allow the business user to create reports on the fly using a simple drag-and-drop interface. There can be endless possibilities of how you decide to view or analyse your data. The exercise of deploying such a robust platform as to allow such flexibility is time consuming but in my opinion time and money well spent on.
To summarise it all, the operational system or ERP is basically DESIGNED to optimise the performance of your business based on how the business is conducted. It has to make it easier to log in sales, issuing of invoices, etc. What it’s not designed to do is to provide a flexible reporting structure (barring daily reports – they have their own use) and that is exactly where BI comes in. Don’t let this mean that your operational system can’t do what your BI system does, but to make your operational system do that is going to be a huge undertaking, a very painful process and its going to burden your operational system to such an extent (when you pull up reports) that it is going to affect your daily operations.
So ideally the data is extracted from the operational system and stored in a separate location that is dedicated to be your reporting storage. Since you are making a copy of the data, it would make sense to store it in a way that is more tuned to reporting needs than operational needs. So an end-to-end solutions requires data storage (your data mart or data warehouse), a process to bring in data from your operational system to your data storage (ETL) and obviously the presentation of your data/your reporting tool (BI).
BI itself is a broad category and covers various platforms, tools, solutions and concepts. We’ll get to those soon.