Lesson 7 — BI Portfolio, Users, OLAP & Self-Service
One interface does not fit everyone
The BI portfolio must follow user and decision needs. Typical communities include executives, operational managers, front-line staff, analysts, and power users/data scientists. They differ in latency, interaction, repeatability, analytical freedom, support, and control requirements.
Operational reporting and tactical BI
Operational reporting supports current workflow/status and shorter-horizon decisions. It may query transactional systems directly or use ODS/DW data to add integration and context.
Example: “Which orders are stuck right now?” is operational/tactical.
Strategic BI / Business Performance Management
Strategic BI examines historical integrated information for long-term trends, goals, planning, and performance management.
Business Performance Management (BPM) integrates processes/applications for metrics aligned with organizational strategy—budgeting, planning, scorecards, and executive performance are typical clues.
Example: “How have fulfillment delays changed over three years?” is strategic/analytical.
Operational analytics
Operational analytics applies analytical insight to current/near-current decisions. It often needs lower-latency integration than classic historical BI, which is why an ODS or other low-latency path may appear in the architecture.
OLAP
OLAP supports interactive multidimensional analysis: slice, dice, drill, compare, and examine measures across dimensions. It is an analytical technique/capability, not a synonym for the whole warehouse.
Governed self-service BI
Self-service allows authorized business users to create/explore analyses themselves—but the source material does not equate self-service with no control.
Successful governed self-service depends on: - trusted reusable datasets; - common definitions/semantic rules; - Metadata and lineage; - appropriate security/privileges; - training/support; - publication/quality expectations; - governance of shared outputs.
A portal where analysts can build dashboards from certified data using one governed net-revenue definition is self-service. Analysts downloading local CSVs and inventing different formulas is unmanaged analysis.
Managed reporting vs self-service
Managed reporting provides centrally designed/predefined outputs and favors repeatability/standardization. Self-service gives more creation flexibility within guardrails. Choose based on user need and risk—not ideology.
Prototype to learn
A prototype with representative real data and actual user questions can reveal uncertain requirements, bad source assumptions, grain problems, semantic disagreements, DQ defects, and architecture feasibility before a large commitment.
A successful prototype is not production approval; that belongs to the product lifecycle.
Stop and check
One BI tool for executives, analysts, and front-line users despite different needs: missing user segmentation/tool fit.
Users create dashboards on governed data with shared definitions and controls: governed self-service.
Users create dashboards from uncontrolled local copies: unmanaged analysis.
Source anchor: pp. 379–387.