Continue in 2 seconds

Business Intelligence

  • February 01 2002, 1:00am EST

Is your organization, bogged down with expense control, having trouble "breathing?" Does it feel stuffy and somewhat nauseous? Does it cough a bit while not able to perform what, under other circumstances, would be normal activities? There are some effective and not-too-costly business intelligence (BI) initiatives that may cure this general feeling of malaise and get your company back to feeling like its old self again. After a careful diagnosis of the malady, a judiciously placed packaged data mart or BI application might return organizations to optimal health.

In a "data mart in a box" or packaged BI application, the data model (for the most part), dimensions and data interaction are programmed by the vendor, with prebuilt applications reflecting typical queries needed to manage a particular function. Examples are financial consolidation applications, enterprise network and server monitoring applications, workforce analytical applications (e.g., call center optimization, training/skills assessment, worker productivity, etc.) and supplier or procurement management applications. These applications typically provide predefined reports and queries based on deep domain knowledge of one particular functional area. They can save your organization time and money through a buy-versus-build strategy.

"Okay," you say, "but are there conditions or functions for which the prescription of a packaged application isn't a good idea?" Yes. Activities that need data from numerous operational systems, such as customer relationship management applications (e.g., customer segmentation, customer profitability, etc.) that can be significantly different from organization to organization and require customization to be useful are probably not going to realize benefits from a quick-hit packaged application or data mart. These applications need the data structure, integration and technical infrastructure of a full-fledged data warehouse to be successful. Multiple source systems along with a differentiated method of integrating customer information that results in competitive advantage generally preclude customer-type systems from being successfully packaged (although CRM vendors will tell you otherwise!).

Packaged BI apps are best prescribed for functions with a particular (generally non-differentiating) focus, with "stuffy operations" and with the most urgent needs for decision support information. The idea of a packaged BI application is to save implementation time by providing a predefined data model, perhaps with automatic data feeds from an operational application such as a general ledger, supply chain management, enterprise resource management or human resource management system. The packaged app then provides predefined reports and typical queries that people within that organizational function need to perform analysis. With this approach, some savings in data analysis activities can be achieved, although analyzing data for integration cross- functionally will still need to be performed. Savings in data extraction, conditioning and loading can also be realized.

With a predefined data model, the target database structure is essentially already defined, which saves design time. Additional time can be saved in defining the reports and queries that can immediately provide the most business benefit. This is not to say that implementing a packaged BI app or mart is easy, just that it can potentially provide quicker implementation time and therefore realize the benefits of having analytical information at one's fingertips sooner than an application or mart developed completely in house. The vendor, rather than the organization, does a lot of the implementation analysis and design work. The vendor, usually with understanding of the functionally specific underlying operational system, sources decision support data into the mart structure supporting the BI application. The vendor, able to concentrate fully on one function and with deep knowledge of what information is necessary to make that function successful, anticipates the delivery of that information that will provide the most bang for the buck. The vendor predefines performance metrics that executive and operational managers typically like to see.

Keep in mind that with packaged apps and data marts, care needs to be taken to ultimately integrate their data with the enterprise's data warehouse. Stovepipe data marts can fast become "islands of data," just as legacy operational systems once were, unless an effort is undertaken to have the data warehouse be the eventual source for the BI application data. While your organization can begin the healing process with a packaged standalone solution, it will ultimately need to be in sync with the rest of the enterprise's information to yield consistent results.

A packaged BI application or mart may not initially result in information that is integrated with the rest of the organization, but it may be just what the doctor ordered to get a specific function within your organization back on its operational feet. Like an antibiotic, a packaged app can provide quick relief from an illness resulting from lack of particular focused information for decision making. Expect some healing (return on investment) to occur, along with an increased resistance to future infection/inefficiency. Ah-choo? Gesundheit!

Register or login for access to this item and much more

All Information Management content is archived after seven days.

Community members receive:
  • All recent and archived articles
  • Conference offers and updates
  • A full menu of enewsletter options
  • Web seminars, white papers, ebooks

Don't have an account? Register for Free Unlimited Access