Designing for Business Intelligence Success

Just as patterns and similarities exist in solutions across a variety of organizations, some consistent patterns surfaced in MultiValue organizations that are not getting all they can from the business intelligence efforts.  These three key factors must be addressed carefully in any reporting project.

MultiValue systems and staff are the key to any business intelligence effort

A hallmark of MultiValue is the high level of subject matter expertise among the IT staff in MultiValue shops.  A byproduct of the incredible productivity and longevity of our environment, it is not unusual to find programmers who have been personally involved with implementing every significant business initiative in institutional memory – in fact, these programmers arethe de facto institutional memory.  Any effective business intelligence initiative must involve and leverage this knowledge.

Mainstream reporting solutions can fail to capture the complete picture

Pressured by demands for enhanced reporting, some companies have rushed to off-the-shelf solutions that do not leverage the expertise in their MultiValue systems and staff.  The promise of quick solutions that focus on "tables" and "columns" can miss the rich content in I-descriptors, routines, and less structured data.  Additionally, one-time, scheduled, or real-time migration of large or volatile datasets for data warehousing or integration can introduce complications that even the best-engineered operational system cannot cope with without specialized tools and techniques.

Effective analytical reporting requires specialized techniques  

To avoid problems with off-the-shelf solutions, other organizations have tried to build a custom analytical reporting layer into their MultiValue applications.  Problems with managing several types of change – problems addressed in formal data warehousing approaches – create the biggest challenges to one-off solutions.  Data warehousing strategies and techniques can be applied to MultiValue reporting applications, but they should be applied at the design stage – and that's where we can help.