How to use this document. The source tables below are the candidate sources identified during discovery. "Columns" means the union of source columns presently mapped to that FACT in our working model. Dvara should validate which table is authoritative. A table being listed does not mean it must finally be ingested into Gold — duplicate, derived, archive, reference-only and obsolete sources can be removed once the System of Record is confirmed. Entries such as leads / leads_cstm / campaign-linked tables, wm_location_data_*, etc. are currently logical/composite discovery mappings, not necessarily a single physical table — these should specifically be challenged during the workshop. The KPI coverage column shows real KPIs from our evidenced model that currently depend on this table, so the business impact of removing a source is visible immediately.
The one rule to keep in mind throughout: Removing a source table does not automatically remove a FACT. If Dvara tells us that three source tables contain the same information, we simply reduce the source lineage to the authoritative table. We merge FACTs only when the business grain itself can safely be merged. That separation — Source Rationalisation vs FACT Rationalisation — will prevent the customer discussion from becoming confusing.