Infoverity successfully completed a business intelligence project for a client in the healthcare products and pharmaceutical distribution industry. The client faced performance issues after a recent implementation of SAP Business Objects.
Background
The client had deployed an SAP Business Objects solution in support of a large SAP transformation. The solution was plagued with slow performance, complaints of information availability and inaccuracy, and redundancy of reports. The client recognized the need for an effort to evaluate all aspects of environment design, usage patterns, data lineage and governance to improve performance, user comprehension and collaboration. Infoverity was asked to evaluate report design and consumption, the consistency of key measures and dimensions across universes, and build assets to improve user comprehension of the environment.
An initial assessment verified critical pain points:
- There was not broad knowledge of what reports were being run, by whom, or for what purpose;
- There were no catalogs identifying what information was present on a report;
- Report creation was left to individual best efforts. This impacted report quality, but a side effect was substantial redundancy without consistency;
- There was no formal mechanism to communicate design changes to impacted users due in large part to the lack of visibility to report content, stakeholders, and approved consumers.
Approach
The optimization program consisted of three major work streams:
- A team to prioritize trouble tickets and deploy fixes;
- A team to focus on the methods and procedures to qualify new requirements and train users;
- A team focusing on clarifying existing capabilities, establishing consistent definitions and sourcing, and unifying consumption.
Infoverity planned and led the third work stream. Our approach analyzed reporting activity, prioritizing reports and queries of greatest importance to the business users and having the greatest performance impact on the infrastructure.
Five iterations of analysis and reengineering were conducted. Each iteration lasted two weeks and:
- Was scoped around prioritized universes and the most frequently executed reports from each universe.
- Identified stakeholders by report and universe, as well as by functional information domain. A process was designed to engage owners with key findings from the project, and they became advocates for the implementation of recommended improvements.
- Investigated and clarified the lineage of data used on key reports.
- Identified optimizations and recommendations that were presented to executive steering team members such as:
- Reports that were candidates for consolidation;
- Non-standardized measures and hierarchies;
- Reductions of multiple-universe queries.
Results
Issue Ownership and Resolution
The project Increased information ownership, identified accountable points of contact, and engaged stakeholders of issues, new reporting requests, and governance decisions. Methods for measuring issues’ status through closure were implemented and continuously monitored.
Increased Comprehension, Fewer Tickets, Faster Analysis
A singular reference was deployed which was effective for minimizing time spent investigating report content, sourcing, and meaning. Called the Reporting Enablement Cross-Reference, it displayed report contents, data lineage, and data definitions for the top 400 most frequently requested reports to minimize duplication and maximize comprehension and trouble-shooting effectiveness.
Data lineage was traced from over 2,000 reporting objects back to the operational data store for the benefit of report designers and viewers. Our analysis evaluated over 500 design documents, which identified flaws in the design and implementation that impaired ongoing report analysis. Methods and tools to establish and maintain complete data lineage were explored and recommendations provided that contributed to funding requests for software tools.
Benefit Metrics
- Recommended the consolidation of 25 universes and identified the reduction in resource consumption, analysis hours, and server performance.
- Reduced the average response time of reports by 50%.
- Reduced the number and duration of analysis to resolve end-user support requests by 50%.
- Implemented 279 reporting measure definitions to production. These were viewable by report writers in Business Objects and streamlined report creation and comprehension.
