In my previous article, Creating an Agile Data Warehousing Environment, I discussed several ways to create an agile data warehousing environment, an environment that enables users to quickly perform analyses in support of their decision-making processes. These included cloud-computing and virtualization techniques for minimizing upfront hardware investments and facilitating quick deployments, enhanced search and delivery capabilities for leveraging existing parameter driven reports and analyses templates, and organizational techniques such as relaxed development standards for one-time requests.
In this article I will focus on the desirability of having operational systems include strong analytic capabilities so that some (albeit certainly not all) analyses can be performed and generated directly from them. Although this will never replace the need for a data warehouse, it can be a useful adjunct for helping to satisfy many end-user analytic needs. Our goal should be to provide our organizations with the necessary agility to facilitate their decision-making processes while you can still take advantage of an opportunity or you can cost-effectively resolve a problem.
(Commentary continues below)
Related Resources
White Papers:
The Hybrid Approach to Data Warehousing for Maximum Flexibility
Infrastructure for Dynamic Warehousing: IBM Industry Models
Webinars:
Mixing Advanced Analytics and Transactional Workloads: The New Generation of Data Warehouses
When Search and BI Collide
(Commentary continues)
One example of the inclusion of business intelligence functionality is Oracle's Daily Business Intelligence suite, first released in 2003. Offered as an add-on to several operational systems, it included hundreds of key performance indicators, dashboards, and drill-down reports for a wide variety of business functions and business user roles. Rival SAP augmented its enterprise software applications with composite applications (SAP xApps), several of which were focused on analysis.
Currently many vendors are using the term "embedded analytics" to describe their included or add-on BI functionality. This, of course, is not a new concept as almost all design efforts for operational systems include asking users what reports and analyses they would like generated. Although while most of these were in support of tactical operational requirements, they also included requests for analyses needed to satisfy more strategic goals.
I believe that operations and analytics go hand in hand. When organizations evaluate an operational software application, they should consider its analytic capabilities -- available either as part of the package or as an optional add-on -- as well. Any business intelligence functionality that accompanies an operational software package shouldn't be expected to replace the need for a data warehouse, especially if there is a requirement to integrate data from multiple heterogeneous systems. They should, however, be considered a part of an overall data warehousing environment and deployed when appropriate. That said, organizations need to be somewhat cautious and recognize that any BI functionality that goes directly against an operational system may adversely affect the performance of that system.
The bottom line is that all organizations should strive to create an agile data warehousing environment that can quickly respond to new user requirements. They may sometimes find that these analysis requirements can be met through their existing operational systems. Even if not a perfect fit, they might be good enough to satisfy some top-level requirements and provide enough time to develop a more complete solution. Organizations should leverage the analyses capabilities of their operational systems and consider how it could be a valuable component of their overall data warehousing architecture.
Michael A. Schiff is a principal consultant for MAS Strategies. He can be reached at [email protected]
Related Resources
White Papers:
Webinars:
Mixing Advanced Analytics and Transactional Workloads:
The New Generation of Data Warehouses
July 28, 2010
Speaker: James Kobielus
When Search and BI Collide
August 25, 2010
Speaker: Wayne Eckerson
