Business Intelligence (also known as BI) is big business. In a recent report market analyst Datamonitor predicts business intelligence spend by retail banking in North America, Europe, the Middle East and Asia-Pacific, will increase around 60.7%, from $5.6 billion in 2006 to $9 billion by 2012.

So what is business intelligence, and why would you even need it?

As is typical in the IT industry 'Business Intelligence' is means something slightly different to everyone. So I have come up with my own definition of BI. Here it is 0988293142:

"Business Intelligence is a system which enables organisations to collect, analyse and present business information to support business decisions."

Business Intelligence is a system ...
BI is not a collection of diverse tools you put together and 'hope for the best'. If you want to get good results you need to plan from the start. Don't be dazzled by the impressive offerings of the software giants. Have a clear picture from the start of what you want to get from BI, and how you plan to get there, and your chances of getting real value from Business Intelligence multiply greatly.

One thing that differentiates BI systems from traditional systems like inventory, distribution and finance systems is flexibility - the ability to add measures and outputs as your organisation and its use of BI develop. But this flexibility does not excuse you from the planning stages - in fact it increases the need for planning.

Unlike many traditional computer systems, a single BI system can provide value to all departments within your organisation, but as with any system it is important not to expect delivery on everything at the same time. Build steadily, working down your priority list.

... which enables organisations to collect ...
The first question you might ask is "why do I have to collect information I already have in my database"? The answer is, you probably don't have it already. You might have some of the information you need, and even then it is probably not in the exact form you need it. Much of the other information you need is probably on spreadsheets on various peoples' desks, or doesn't exist at all and has to be collected.

Even if you have all the information you need already (which is unlikely) it is a good idea for the BI system to store it somewhere else. That way the data can be organised and aggregated to make it work quickly and efficiently in a BI system. Often you need to add history to your BI database, which may not be kept in your existing transactional systems.

The most popular way to collect the data is in a specially-designed data warehouse. It takes time and skill to develop a good data warehouse, but in most cases it is vital to an effective BI implementation. A good data warehouse need not be a huge, complex beast - the simpler the design, the lower the cost, and the more chance of success.

A good data warehouse design can be easily extended to allow for unforeseen business reporting requirements.

... analyse ...
Generally speaking, data is best suited to BI reporting when it is

(a) summarised, and
(b) organised in hierarchies

. In large organisations with large amounts of data this number-crunching process can require millions of calculations and is often carried out overnight on a daily basis. Calculations, sometimes between quite diversely held data elements, allow you to create specific 'key performance indicators' (KPIs) such as Profit per Customer and Revenue per Employee.

Data held in multidimensional structures known as cubes contain this hierarchical, summarised information which allows managers to analyse KPIs at any level of the organisation - giving them the ability to see, for example, Revenue per Employee at National, Regional and Area levels, by month or summarised at the year level.