The goal of Business Intelligence (BI) is none other but to provide organizations with real information about their internals and the surrounding environment. However the goal seems simple, its materialization can be achieved through various solutions.
Two major circumstances define the best approach for the given situation. One is the volume and quality of data assets that should be available at the company. The other is the aspects and cases of using those assets. For the later we can highlight the following:
Information that is mainly to measure the company from an accounting point of view, based on accounting and financial data, from which the majority is governed by regulations;
Information gathered from customer data. Concerned items are customer value, customer behavioral patterns and analysis of customer risk;
Evaluation of transactional indemnity;
Data describing the market environment, effects of competitors;
Calculations providing assurance for compliance with special, e.g. Basel II, directives;
Needless to emphasize that the pertinence of decisions influencing the company’s future are primarily based on the quality of present and past data, which describe how the company functions. The cornerstone that affects the quality of the data is that only one occurrence of the data exists for one particular event (not several, slightly different ones) and that it is accurate. It seems to be trivial but its implementation needs extensive consideration. As well it is an important, although not always possible, principle that the company’s rudimentary (a.k.a. not derived) data are gathered in one instance, one coherent to the other, consistent and maintained. This enables individuals utilizing the data asset to access the information properly. Following this principle the solution is the data warehouse.
There is one, even more elementary, based on common sense, requirement: apart from the possible IT solution, we have to be aware how, at which point of the processes, in which system, with what content and quality these business data are arisen and how long they are available there. This is the company’s data asset dictionary, or business data catalog. Its practicality and quality define the value and usability of the IT solution that utilizes it.
Based on the selected IT solution’s architectural design there are several approaches which reside in distinct ranges of short-term, long-term return and data quality dimensions.
A central data warehouse, the specialized data markets fed from it and the specialized logic (engine) create high quality data, its return is long-term;
Specialized data markets with built-in specialized logic (engine) without a data warehouse bear a risk in data quality, its return is mid-term;
Fulfilling custom calculation demands below the data market level, where data quality is risky, have a short-term return, but looking at the company as a whole the cost of the mini solutions will overcome the cost of data warehouse and data market model solutions in the long-term.
A broad number of tools and applications can be selected, as for data asset management, to exploit data asset business or profession wise. These are, without limitation, SAP BusinessObjects, Oracle BI and Cognos. One of these may be selected based on the company’s preferences, demands, or IT environment.
Utilization of the data asset can be supported with scheduled reports that are generated based on already known demands, or in a different manner the analysis can be done by an individual in an intuitive, spontaneous way. Multidimensional data organization, although relevant for the first case, is critical in the latter. Having available the rudimentary data for derivative operations (calculations or aggregative) is crucial as well.