The Rise, Fall and Return of Operational BI and Analytic Applications

I find the recent press on operational BI an interesting resurfacing of events.

A few years ago, operational BI was a hot topic. With EII technologies that allowed BI tools to tap directly into source systems, some wondered if it was the demise of data warehousing as we knew it. (It wasn't.) Analytic applications share a similar story with operational BI, as many vendors initially jumped on this band wagon and later retrenched, either exiting this market entirely or rethinking their strategy.The analytic applications market is a crowded one. There are industry- and process-specific solutions. Before it was acquired by Infor last year, Epiphany, for example, focused on CRM analytics. Spotfire (soon to be acquired by TIBCO) has origins in the life sciences industry and offers a number of industry-specific analytics, including clinical trial and research analytics.

BI vendors such as Business Objects, Cognos and Oracle also have analytic applications. SAS derives a sizable portion of its revenues from its analytic applications, dubbed SAS Solutions. Generally speaking, I would describe the solutions from these vendors as more functionally aligned (finance analytics, work force analytics, supply chain analytics) rather than industry specific. Solutions from these vendors share a similar trait: they extract data from source systems, such as SAP, Oracle E-Business Suite and PeopleSoft, and populate data marts with predefined data models. These analytic applications also normally include prebuilt reports.

Here is where the worlds of Operational BI and Analytic Applications overlap; how do you get to the data inside the operational systems? Do these analytic applications understand the thousands of normalized tables in the source systems to handle the ETL process? In some cases, yes. Other times, no.

Vendors such as Noetix and Information Builders take a slightly different approach from analytic applications. Noetix, for example, doesn't extract the data into a data mart; instead, it accesses the source systems directly via NoetixViews. Where Noetix differs from Information Builders (and Business Objects, Oracle and SAS) is that it is pursuing a strategy of letting customers use their BI tool of choice, whether Cognos, Business Objects, Oracle or Excel 2007. Information Builders subsidiary iWay Software offers its application adaptors either for ETL or for real-time reporting. Both have described their solutions as operational BI.

Analytic applications and operational BI share some similarities but have slightly different purposes. As you evaluate the alternatives, first think about what you are most trying to accomplish:

• Are you trying to get to the detailed data in the operational system to complete an operational task (think operational BI), or • Are you trying to improve a business process by analyzing a predetermined set of metrics relevant to that process (think analytic applications)?

In evaluating either operational BI solutions or analytic applications that access transaction systems, also consider these detailed differences:

• Which transaction or ERP systems can be accessed (SAP, J.D. Edwards, Oracle) and how well are configuration-specific deployments handled? • Which specific modules within the ERP are supported? • Is the access real-time or via a data mart? Understand the trade offs of either approach. • Which BI tool is required? • What prebuilt reports and KPIs are provided? • How are upgrades to either the transaction system or the BI tool handled?

If you have recently evaluated any of these solutions, I hope you'll share your thoughts.

Sincerely,

Cindi Howson, founder BIScorecard product reviews.I find the recent press on operational BI an interesting resurfacing of events. A few years ago, operational BI was a hot topic. With EII technologies that allowed BI tools to tap directly into source systems, some wondered if it was the demise of data warehousing as we knew it. (It wasn't.) Analytic applications share a similar story with operational BI as many vendors initially jumped on this band wagon and later retrenched…