Before investing in any business intelligence (BI) solution, a prudent question to ask the BI vendor is whether its proposed BI solution offers business information or true end-to-end enterprise-wide BI. The clients should investigate thoroughly all the various activities that a proper, enterprise-wide BI system should enable.
Too many vendors provide BI only up to a point - and organisations end up with only one benefit: a historical data view. Real BI goes well beyond the mere provision of a rear-view mirror look at the organisation. It adds powerful analytics to quality historical data, enabling accurate, reliable intelligence and forecasts that trigger appropriate action.
Any BI implementation must obviously include powerful access to disparate data; the true integration of this data, some sort of extraction, transformation and loading (ETL) activity - which includes quality assurance and intelligent storage. It should not stop there. At this stage, all that has been achieved is a historical data store.
In addition, the organisation needs a statistical or analytical layer. This starts to add real business intelligent value to the data. An intelligence distribution layer closes the BI loop.
Access to disparate data is absolutely key to any BI system. It must be able to take data in any format, from any platform, be it a legacy mainframe system, an enterprise resource planning (ERP) system or the many disparate accounting systems within a multinational group of companies. This data then needs to be integrated, so that an organisation obtains an enterprise-wide view. For example, the fact that some human resources (HR) data is stored in a HR system, while another section of it is contained in an ERP system, should not affect the holistic view of the HR data. All this disparate information is needed to answer crucial HR business questions intelligently.
Real BI goes well beyond the mere provision of a rear-view mirror look at the organisation.
Carel Badenhorst, business intelligence product manager for SAS SA
ETL activity follows on from data integration. This must contain a fully integrated quality component ensuring only reliable and quality data is stored in the warehouse.
The organisation will now have good quality, clean data to work with. However, it is of little use if this information is not accessible to users in a timely and usable manner. Intelligent storage is vital. Companies should make sure the BI infrastructure in which they invest is optimised to store data intelligently, and to access required data rapidly.
Relational data base management systems, ERP or online transactional processing, for example, while all able to gather data, have not been designed for the optimal exploitation of data. BI should be supported by systems, data models and intelligence architecture able to work with huge volumes of data very quickly, including online analytical processing.
So far, so good. But if organisations stop with BI at this point, all they have achieved is good quality, BI-enabled, but nonetheless historical data.
Businesses will only be able to look at what happened in the past. For example, they could ask: "How many did we sell yesterday?" They would not able to obtain a forward-looking view. It is only when they begin to add analytical and statistical processes to the historical data that they can begin to ask forward-looking questions, and start reaping intelligence from their historical information.
Intelligent projections and forecasts will allow organisations to begin asking questions like: "What happens if rainfall increases by 20mm per annum in the major maize areas in Mpumalanga?" "What will be the effect on the maize crop in 2010?" "What do I need to do to ensure I have a contingency plan in place?"
We now have good quality, reliable intelligent data, but the BI system is still incomplete. This information must now be made actionable - and for this to happen, it must be distributed to those who need to act on the intelligence.
These people - including business analysts, IT users and decision-makers - must receive this intelligence in the format with which they are most comfortable. We need to make sure the right people get the right intelligence at the right time so that they are empowered to take action. Some may want it via an Internet browser, others in a Word or Excel document, and yet others in a Java interface. A true BI system should be able to do this, so that BI adds real value within the organisation - by leading to action.
Share