Data Mart
Businesses have to often choose between data marts implemented in some departments in contrast to the bolder alternative of investing in a data warehouse or an enterprise wide real time intelligence infrastructure from the very beginning. A smaller data mart can be a test case and is more likely to be accepted by a minority of technology buffs in an organization while a data warehouse or enterprise intelligence is more efficient. If companies opt for a data warehouse, they will encourage standardization of data definitions and planning for business process management across the enterprise which is not a pressing concern when data marts are implemented in some divisions. Furthermore, data marts require a different kind of technology which is rendered obsolete later when enterprises decide to accept data warehouses or enterprise wide real time intelligence. Also, data marts, duplicate data which has to be cleaned later when companies switch to data warehouses.
One case where a company had to abandon its investments in a data mart is GE Real Estate which has investments in over 8,000 properties around the world. The management of such a portfolio required information that was often available in the individual countries where the operating units were located. GE decided against the ideal solution of a web based data warehouse solution primarily because reconciliation of a variety of data definitions used by its departments seemed much too daunting. After the modest success of its data marts, GE had to entirely abandon its Microsoft server which could not scale for a data warehouse and had to instead change to a UNIX server with an Oracle database.
A more comprehensive survey of 150 technology executives, who have implemented business performance management (BPM) across the enterprise, shows that a step-wise process has been more successful for most companies. Twenty percent of the respondents have implemented BPM and 60% of them implemented them for particular problems before expanding their scope to the entire enterprise.
The decision to buy software for a specific department or the company as a whole is not simply a technical decision. While the more skilled or “power users” are avid consumers who relish new software, the risk of provoking a more political response from other employees is much greater especially if they are either wedded to the tools they have used in the past (typically, Excel) or are not sophisticated users of some arcane business intelligence software. The choice to use more integrated software should be preceded by finding a sponsor with the clout to implement new business intelligence software. Also, companies have to find the means to customize the software according to the roles of each individual in the enterprise.
On the other hand, the benefits from enterprise business intelligence yield the greatest benefit. Companies can more effectively align their strategies with the resources and tactics of their company. They can also monitor performance in real time to ensure that their actions are able to achieve their goals.