Panorama Business Intelligence

Standardizing BI – Application vs. Platform 

If you ask the “industry” about the optimal way of standardizing BI, the consensus definitely leans toward the platform.  Choose a single platform to standardize on and use it to access data dispersed across the organization.  Then use one set of reporting tools to gain insight about business trends and opportunities.

It sounds great.  But it’s also an ideal that is difficult to achieve.  The industry is moving in that direction, but the progress is deliberate.  Currently, there are two main platform standardization approaches being advocated in the marketplace.  But they do not always help organizations achieve the best reporting, analysis and performance management from data across the organization. In addition, standardizing on one platform requires an organization to sacrifice functionality from existing investments, allot significant budget and demand effort from already overburdened resources.

The first approach focuses on standardizing on a specialized BI platform like Cognos Powercube or Business Objects Universe.  These solutions are designed to be agnostic so that they work with commonly used Line of Business (LOB) systems that already have their own optimized BI platforms. Examples include offerings from some key players including SAP, Oracle and Microsoft.      

Most organizations choose optimized platforms because of the automation.  Everything, from sourcing LOB data, building models, cubes, metadata and aggregation, is automated.  Having these capabilities eliminates the need for manual processes, which are expensive and time consuming.  But the automation and optimization is tailored to the applications (for example, SAP’s ERP suite) and the supporting BI platform.
 
So to make Cognos’ or Business Objects’ agnosticism work, organizations must extract data or metadata from existing BI platforms and systems and build the models and metadata on the third party BI platform.  Doing so means giving up the optimization inherent in a vendor’s BI platform because it is virtually impossible to recreate it on a third party platform. 

For example, SAP can build the data warehouse, cubes and Queries automatically in a way that optimizes analysis and reporting on data from the SAP business suite.  But if a company were to standardize on Business Objects Universe by recreating models and metadata from SAP, it would lose the optimized capabilities inherent in SAP.  None of the LOB BI platforms has been optimized or designed to enable the porting of its optimization into third party platforms.

The second approach has organizations choosing one platform where most of its data resides, already optimized.  For example, many organizations working with SAP would choose to standardize on the Netweaver BI platform.  They would take data from other LOB systems and bring it to the SAP platform.  And that’s when issues arise. 
First, SAP doesn’t support other types of data in the same it supports its own, which limits BI capabilities.  Second, Seibel CRM data, for example, has already been optimized for its specialized BI platform in the same manner that SAP has optimized its BI platform for its own applications.  Once an organization moves it over to SAP, it is giving up the optimization inherent in the Seibel platform, curtailing the potential of reporting on the CRM data. 

So how can an organization realize the benefits of standardization without undertaking the complexity of recreating the models and metadata or sacrificing existing data optimization?  It simply requires a change of thinking about how to standardize. Instead of approaching standardization from the perspective of the platform, why not standardize on the application? 

Isn’t the goal to take data from across the organization, amalgamate it, analyze it and present it as reports, scorecards or dashboards?  That can all be done on the application level.  Standardizing on the application enables organizations to access and use data from multiple sources without having to deal with the hassle and expense of bringing it together on one platform.

How is it done?  The application translates queries using multi-dimensional expressions (MDX), a BI standard.  The leading platforms and systems including SAP BW, Microsoft SQL and Oracle support MDX.  As a result, the application can originate queries, retrieve results from the different platforms and conduct the required calculations to enable the data from different platforms to work together for reporting and analysis. 

If you’re interested in the potential of BI application standardization, where do you start? If your thinking and focus has been solely on the platform, it can be challenging to find the right type of provider amidst all of the available offerings.  Determining the BI capabilities you want should be the first thing you do.  Look at the systems you have and the data you want to analyze.  Is it SAP-based?  Is it Microsoft-based?

The ideal solution should be able to connect seamlessly to any platform without the need to rebuild models and metadata and provide the capabilities to bring the data together at the report, cross tab or dashboard levels. 

It is important to note that standardizing on the application is a step in the evolution toward one platform.  In an ideal world, every organization would be able to bring all of its data together in one place, access it, analyze it and report on it.  But the existing technology available in the marketplace has yet to mature to that stage.  So standardizing on the application is evolution, not revolution.  It’s a way to take advantage of the platforms and optimization that LOB vendors offer with less risk, less work and virtually no manipulation or rebuilding of data.

Naysayers refute the application approach by claiming that no one application supports all platforms.  And that’s true.  There isn’t one application that can connect natively to all optimized platforms.  But with 60% of the market using either Microsoft or SAP, there are proven applications that can connect natively to both of those platforms, enabling 60% of organizations to realize the potential of BI application standardization.

In addition, focusing on the most widely used platforms facilitates a depth of analysis that would be almost impossible if the application supported six or seven different platforms.  BI application vendors designed and developed their solutions to address the pervasive platforms so that the largest majority of companies could benefit.  As time goes on and platform technology matures even further, applications will grow more sophisticated in tandem to support a broader range of platforms.

When it comes to BI standardization, the approach needs to be evolution, not revolution.  Many organizations are missing out on opportunities because they are spending time and money needlessly extracting data and rebuilding process.  The notion of standardizing on one platform has become so ingrained that they feel they don’t have another choice.  But they do.  The right application can bridge the gap between platforms without the need for all the work in the back end, opening up cross-platform communication and data analysis opportunities.