Source system primary keys and ETL Lineage: You must document the exact column names from the source system to ensure the rough model is a 1:1 mirror of the operational database structure.

Name, business definition, and aggregation rule (additivity): You must define what the indicator represents (the business logic) and, crucially, how it behaves when rolled up across dimensions (e.g., Sum, Average, Count, or Last Value for semi-additive measures).

Visual formatting, color thresholds, and icons: You must define how the indicator will appear in the final dashboard (e.g., "Red if < 50%") so that the database engine can pre-render these visual states for faster retrieval.

Data type, precision, and storage bytes: You must define the physical database characteristics to ensure the rough model fits within the allocated storage capacity of the server.