The big bang approach is defined by going for an enterprise-wide deployment of BI right out of the gate.  In the big bang approach there is a corporate edict that there will be a unified and consistent approach to BI that will be governed and rolled out globally and governed centrally.   For this approach diverse requirements have to be gathered, compiled, vetted, and prioritized.  With diverse business requirements come diverse data requirements that need to be defined and sourced which requires more IT involvement than the silo approach.

In the silo approach specific BI needs are dealt with on a case by case basis.  It could be defined as a ‘one-off’ approach to BI.  As needs come up, the specific need is defined and scoped and a product is selected and implemented to meet that specific requirement.  Very often this is done within the business unit without oversight by IT or with any sense of, or communication with, other BI initiatives that may be going on in the organization.   These types of deployments tend to be functionally focused around a specific process, for instance sales analysis or financial analytics, with little if any cross-pollination of data across functional areas.  This approach can be quick to implement but can also result in a proliferation of data marts without common definitions or consistent data sources.

Think Globally, Act Locally is a hybrid approach that can provide an organization with some of the best of both worlds.   This approach bridges the gaps between the Silo and Big Bang approaches.  Like the silo approach it starts in a division or function and focuses on a specific measurable pain and its scope is limited to resolving that pain but in the context of strategic goals.


For more information about this and other BI topics check out my white papers and webcasts on the IBM Cognos Innovation Center for Performance Management website.