```json { "table_name": "b2b_sales_funnel_pipeline", "table_type_tags": [ {"tag": "dimension", "confidence": 0.90}, {"tag": "path", "confidence": 0.15} ], "primary_type": "dimension", "insufficient_evidence": false, "entity": { "columns": ["opportunity_id"], "unique_per_row": true }, "time_axis": { "columns": ["engage_date", "close_date"], "kind": "event_time" }, "reasoning": "opportunity_id has Uniqueness 1.000 — each row is unique information about one sales opportunity. The timestamps engage_date and close_date are descriptive attributes of the opportunity (when engagement started, when it closed), not an axis ordering state transitions. The deal_stage column shows current status (Won/Lost/Engaging/Prospecting) but lacks transition timestamps — we see the outcome, not the lifecycle. Sibling tables (accounts, products, sales_teams) suggest this is the catalog of opportunities those tables reference. Dimension is the clear primary reading; path gets weak consideration only because close_value could theoretically accumulate, but there's no evidence of replay semantics or state derivability." } ```