GA4 Data Modeling for Marketing Attribution in BigQuery
I show why GA4-to-BigQuery users should move beyond ad hoc queries to purpose-built data modeling driven by a clear use case, using marketing attribution as the main example. I argue attribution is primarily a strategy problem, not solvable by a single “best” model or measurement tweak, and outline the core attribution ingredients: identity, touchpoints, and conversions. The session details anonymous (GA4 pseudo_id) vs identified users (user_id/email), how to raise identification rates, and how to stitch identities via journey mapping and a mapping table. It then covers touchpoint design (often session_start) and presents an activity-schema-based BigQuery architecture with lean staging tables, accounts/anon tables, identity resolution, touchpoints, attributed touchpoints, and an analytics layer combined with ad cost data.
00:00 Why Data Modeling Matters
01:25 From Ad Hoc Queries to Models
04:26 Pick a Use Case First
06:03 Attribution Layers Explained
07:24 Attribution Is Strategy
10:35 Attribution Data Ingredients
15:18 Identity Basics in GA4
15:58 Anonymous vs Known Users
19:57 User ID and Email Strategy
23:51 How You Identify Users Today
24:30 Identity Resolution Basics
26:41 Journey Mapping Workshop
28:03 Hashing and Passing IDs
29:53 Mapping Table Modeling
32:56 Filtering Touchpoints
34:52 Activity Schema Modeling
36:51 Boost Identification Rate
38:24 Touchpoint Design Choices
44:46 Attribution Architecture Blueprint
48:15 Wrap Up and Next Steps
