ga4-export-fixer 0.4.1-dev.2 → 0.4.1-dev.4

package/columns/columnDescriptions.json

@@ -157,7 +157,7 @@
         }
     },
     "session_first_traffic_source": {
-        "description": "The data extracted from the collected_traffic_source struct of the session's first event. Similar to the session_traffic_source_last_click but does not include last non-direct attribution. Meaning that sessions that started without a source of traffic are always reported as direct in this field. Can be useful for building custom attribution models.",
+        "description": "The data extracted from the collected_traffic_source struct of the session's first event. Can be used for building custom attribution models.",
         "columns": {
             "manual_campaign_id": "Manual campaign ID (utm_id)",
             "manual_campaign_name": "Manual campaign name (utm_campaign)",
@@ -174,7 +174,7 @@
         }
     },
     "session_traffic_source_last_click": {
-        "description": "Session-level last-click attributed traffic source across Google Ads and manual channels. Aggregated as the first non-null value in the session (the field is session-scoped in the GA4 export and identical across events in the same session)",
+        "description": "Primary session attribution data. Session-level last non-direct click attribution.",
         "columns": {
             "manual_campaign": {
                 "description": "Last-click manual campaign attribution",
@@ -202,14 +202,14 @@
                 }
             },
             "cross_channel_campaign": {
-                "description": "Last-click cross-channel campaign attribution",
+                "description": "Last-click cross-channel campaign attribution. The primary session attribution data within the session_traffic_source_last_click struct",
                 "columns": {
                     "campaign_name": "Cross-channel campaign name",
                     "campaign_id": "Cross-channel campaign ID",
                     "source_platform": "Cross-channel source platform",
                     "source": "Cross-channel source",
                     "medium": "Cross-channel medium",
-                    "primary_channel_group": "The highest-level traffic categorization (e.g., Organic Search, Paid Social). Use this as the default dimension for all general 'traffic source' questions",
+                    "primary_channel_group": "The default dimension for ranking traffic popularity and acquisition performance. Prefer this over source/medium for general summaries. Use this as the default dimension for all general 'traffic source' questions",
                     "default_channel_group": "Default channel group for the session's last non-direct click"
                 }
             },

package/columns/columnTypicalUse.json

@@ -15,8 +15,8 @@
     "ecommerce": "Transaction and revenue data. Filter to purchase or refund events (WHERE event_name = 'purchase') for ecommerce reporting",
     "items": "Product-level data array. Unnest with CROSS JOIN UNNEST(items) for item-level analysis. For product revenue and units sold, filter to event_name = 'purchase' and use item_revenue and quantity. Use item_category to filter or group by product category",
     "collected_traffic_source": "Event-level UTM parameters and click identifiers. For session-level attribution, prefer session_first_traffic_source instead",
-    "session_first_traffic_source": "First-touch traffic source for the session. Use for session-level acquisition and campaign reporting",
-    "session_traffic_source_last_click": "Google-attributed session traffic source. For high-level 'traffic source' reporting, ALWAYS use .cross_channel_campaign.primary_channel_group as the primary dimension. Use .cross_channel_campaign.source as a secondary drill-down when granular detail is required",
+    "session_first_traffic_source": "Typically NOT used as the default field for general traffic source reporting. Use when building custom attribution models that require the session's first touchpoint, including sessions that started without a source (reported as direct)",
+    "session_traffic_source_last_click": "The default field to use for session-level traffic source analysis. For high-level 'traffic source' reporting, ALWAYS use .cross_channel_campaign.primary_channel_group as the primary dimension. Assume that the user wants the data on the channel grouping level. Even if they ask for \"traffic sources\". Use .cross_channel_campaign.source as a secondary drill-down when granular detail is required",
     "data_is_final": "Data stability flag. Use WHERE data_is_final = true when you need only stable data that will not change in future refreshes",
     "export_type": "Source export type (daily, intraday, fresh). Use for debugging data freshness or filtering by export source",
     "row_inserted_timestamp": "Pipeline metadata. Use to identify when data was last refreshed or to debug incremental update issues"

package/columns/tableAgentInstructions.json

@@ -99,7 +99,7 @@
         },
         {
             "field": "session_traffic_source_last_click",
-            "note": "Primary field for traffic source analysis. When a user asks a general or vague question about 'traffic sources' or 'where traffic comes from', DEFAULT to .cross_channel_campaign.primary_channel_group. Only drill down to .source if the user explicitly mentions 'source' or specific site names.",
+            "note": "The primary source of truth for all session-level traffic analysis. For high-level reporting (e.g., 'most popular sources'), always default to the primary_channel_group field within this struct. Only drill down to .source and .medium if the user explicitly asks for the \"source / medium\" level of detail.",
             "dependsOn": ["session_traffic_source_last_click"]
         },
         {

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "ga4-export-fixer",
-  "version": "0.4.1-dev.2",
+  "version": "0.4.1-dev.4",
   "description": "",
   "main": "index.js",
   "files": [