While businesses around the world are rapidly switching onto the power of business intelligence, very few are paying sufficient attention to the utilisation and availability of this information throughout the organisation. This is according to Rian Durandt, head of Informatics at 3fifteen, the Microsoft applications solutions division in the Dimension Data SA Group.
Business information requirement work sessions are held with key stakeholders to define information subjects. Toolsets and templates are used to facilitate these sessions - as they offer a jump-start to many information requirements. Based on the identification of information subject areas, dimension maps are defined that indicate the determinants - when, what, where, who and how many.
The next step is to investigate current and proposed static reports from current operational systems. This will enable the incorporation of such information into the proposed Business Intelligence data architecture. These reports are typically used for static periodic reporting. This information would typically be at a lower level of detail than that of the dimensions that were identified.
At the end of the KPI and Dimension Modelling phase, dimension maps and report definitions are delivered for review. This supplies the input to the next phase - Logical design.
It is important to note that this phase needs to be completed before the physical design of the solution is started - this will ensure that the solution can be delivered in a staggered approach without the need to continuously re-design.
He points out that it's not merely an issue of whether the information is available, but also who is using it, who needs to use it and for what purpose. "At present, the focus of business intelligence is to take data from a variety of sources such as operational databases (e.g. manufacturing data) and functional databases (from HR or sales), and put it into integrated data marts from which information can be easily extracted.”
"However, if businesses took some time to actually examine this information, they could find that by incorporating a single additional attribute, the same data could deliver significantly more intelligence than anticipated.
“For example, sales data is generally used only to track and manage sales. But if that data was made available to the HR department, it could be used to highlight which sales personnel require training in specific areas; the production department could use that information for planning; while the warehouse could use it to better plan the layout of the facility, thereby boosting productivity and efficiency,” he explains.
According to Durandt, this evaluation of available data, what 3fifteen terms an information audit, must take place before the implementation of the business intelligence solution.
"Once the system is in place, it could be too difficult or expensive to effect changes. In addition, unexpected duplications of data such as two departments having information on the same customer, albeit in different formats, could unnecessarily complicate the development of the solution.”
He says the information audit determines the who, when, what, where, why and how of every data source in an organisation. “Business intelligence doesn't depend only on knowing what data is available and where it is located in the organisation, but who needs it, when it is needed, why it is needed, and how it is gathered and delivered to the desktop.”
"In addition, the audit should identify who owns the information, who is responsible for it, who updates it and who ensures it is accurate. All too often, the IT department is regarded as the custodian of the organisation's information when in reality, the IT department is not responsible for its accuracy at all.
"Finally, the information audit must determine the timeousness of the information and who requires it and when. For example, it might not matter too much if the HR department only receives the sales figures for individual sales personnel once a quarter, but the sales director will require that information considerably more frequently in order to make strategic decisions," he says.
"If the information is required on a daily basis, the audit needs to establish when and how that information is updated, daily or even hourly, and then it must determine when and how it is delivered to the desktop of the person who actually needs it.
"Without this accurate information, a business intelligence system - regardless of how sophisticated it is - will not deliver on its full potential," Durandt concludes.