Build your marketing data warehouse with Claude

In this series, our CEO walks through the full process of building a digital marketing data warehouse in BigQuery: from an empty warehouse to millions of rows from 60 ad accounts unified in one place, with every ad’s image stored in your own cloud.

The whole thing happens in conversation with Claude, connected to Detrics via MCP: exploring the data sources, designing the schema, creating the tables and transfers, loading 3 years of history and leaving the data to refresh itself forever.

What you’ll have when you finish

• One BigQuery table per platform (Meta Ads and TikTok Ads) with all your ad-level data, day by day.
• The full history loaded before the platforms delete it.
• Your ad images stored in your own bucket, with no expiring URLs.
• A unified view that merges Meta and TikTok into a single table, ready for dashboards, alerts or agents.

What you need

• A Detrics account with your data sources connected.
• Claude (web or desktop) with the Detrics MCP connector.
• Your own Google Cloud project where the warehouse will live.

Video links

• Connect Claude to Detrics: app.detrics.io (custom connector: mcp.detrics.io)
• Guide to connect BigQuery to Claude: support.detrics.io/bigquery/connect-to-claude
• What to build on top of this data: the dashboard series with Claude Code

You can watch the full video straight through, or go chapter by chapter: each one has its summary, the transcript and the exact prompts used in the video, ready to copy.