About Tealium Audience Discovery for Snowflake
Overview of Tealium Audience Discovery for Snowflake, its capabilities, and how it integrates with Tealium for audience activation.
Tealium Audience Discovery for Snowflake is a native Snowflake application that lets you build audiences from Snowflake data and activate them in Tealium without moving data out of Snowflake.
How it works
The app runs entirely within your Snowflake account. It defines audiences from your data and stores each audience as a Snowflake table with a corresponding view for external access. The audience data refreshes on a schedule you configure. Tealium reads from the view using a Snowflake data source to load records for activation.
The overall flow is:
- Create an audience in Tealium Audience Discovery for Snowflake.
- The app stores the audience as a Snowflake table and a corresponding view.
- The app refreshes the table on the schedule you configure.
- In Tealium, create a Snowflake data source that connects to the audience.
- Tealium loads the data for activation through connectors and data workflows.
The app reads data from Snowflake and creates audience tables. It does not modify existing customer data. Audience activation and downstream workflows run in Tealium.
Requirements
- A Snowflake account owner or administrator role to install the app and manage user and data access.
- Access to the Snowflake databases and schemas to use for audience creation.
- A Tealium account with permission to configure data sources.
CloudStream is recommended for easier setup because it automatically creates attributes based on incoming data. If you use AudienceStream, you must manually create attributes and audiences.
For installation instructions, see Install Tealium Audience Discovery for Snowflake.
Use cases
- Data teams identifying high-value or lapsed customers by exploring and querying warehouse data.
- Marketers applying point-and-click filters to Snowflake data without writing SQL.
- Teams using Cortex natural language queries to explore audience ideas and schedule them as datasets for activation.
Next steps
This page was last updated: April 8, 2026