Companies can use their existing data and skills to launch an analytics system to support business functions, but improving the business value of analytics systems is a continuous process, says business analytics consultancy BI-Blue solutions manager Anthony Rodriguez.

Business analytics should be applied to deal with a specific business problem or objective, and its use can be broadened and its business value increased as more data are related to the underlying models.

“This process of relating new sources of unstructured data to the specific model, in effect making it structured data, is the continuous process required to maintain and improve the business value and relevance of analytical systems,” he explains.

Business analytics systems are often considered to be highly complex and high-volume data processing systems. However, they are mainly used more surgically to support various decision- making processes based on empirical data and modelling.

“Business analytics is not a new function. Many disciplines have the necessary statistical knowledge to design the relations between data that underpin business analytics systems. What is new is the speed at which these models can be run, from the old retrospective business analyses to the near-real-time analysis of today.”

Companies can use their technical skills to design the logical relational map that will be the blueprint for the design of a business analytics system.

Business analytics should be designed to enable the execution of the company’s strategy, but companies must also remember to include new sources of information, should they be required.

“While data quality is necessary to produce good-quality analyses, companies must not become bogged down by focusing on this issue. Data quality is a process.”

The quality of the data required by analytics is important, which means that storing unnecessary data is counterproductive, adding costs and data management risks, he notes.

“These systems must address business objectives. Does the analytics system fit the business strategy and does it help to fill a gap in the strategy?”

Business units can potentially use data from other divisions as a quick way of gaining access to data and can then look at the consistency and the relationships these data have to the unit’s analytics models.

Further, once unstructured data have been linked to the model, the data must also be subjected to the same-quality rigours of the structured data used by the system. It can then potentially be used in other analytics models dealing with other business problems or processes.

“To properly structure data before using it is a time-consuming and costly challenge. A more effective way is to focus on the data source to ensure that data are processed and correctly structured when captured, which should only be done to meet a business strategy objective.”

If the implementation of a business analytics system takes longer than three months, then the objectives are not precise enough.

“Business analytics is not expensive to implement – the only requirements are the ability to access information at the lowest level and, if possible, in real-time without constraints for data relationship modelling, as well as for users to recognise the strategic and business value of these systems and then support them accordingly,” concludes Rodriguez.