Multi-party insights¶

About the template¶

This template demonstrates a three-party data clean room use case built on Snowflake Collaboration Data Clean Rooms. It brings together three distinct datasets representing different parties in an advertising and measurement workflow:

• Publisher (Exposures): Ad exposure data including hashed email, IP address, impression date, ad group, and campaign ID, representing audience impressions on an ad platform.

• Advertiser (Purchases): Purchase transaction data including hashed email, purchase date, platform, location, and purchase amount, representing customer buying behavior across mobile, desktop, and CTV.

• Identity Partner (Customer Spine): Identity resolution data including hashed email, state, and filing type, serving as a third-party spine that enriches the collaboration with geolocation and demographic attributes.

The template joins all three datasets using standardized join keys (hashed email SHA-256) to surface aggregated insights on purchase behavior broken out by geolocation. Specifically, the analysis computes the total number of purchases and average spend amount per state and ad group across the three parties.

This three-party join is a key capability of Snowflake Collaboration Data Clean Rooms, going beyond the one-to-one provider-consumer clean room models. The collaboration specification controls which parties can run which templates and access which data offerings, enabling flexible permissioning. For example, the identity provider can contribute data without having access to run any analysis templates.

Two SQL templates are included:

• Audience Overlap: A simple overlap query that returns distinct hashed emails found across the joined datasets, useful for audience activation.

• State Spend Analysis: An insights query that joins all three tables to compute total purchases and average spend per state and ad group, delivering actionable campaign performance insights.

Collaboration roles¶

Key use cases¶

• Identity resolution: Securely match customer records across publisher, advertiser, and third-party identity providers using hashed join keys, enabling a unified view of audiences without exposing raw PII.

• Audience overlap analysis: Identify shared audiences between a publisher, advertiser, and identity data provider to evaluate match rates, refine targeting strategies, and activate matched segments for campaigns.

• Purchase attribution by geography: Attribute purchase behavior to specific ad groups and geographic regions by joining advertiser transaction data with publisher exposure data and a third-party identity spine.

• Campaign performance optimization: Aggregate total purchases and average spend by state and ad group to understand which campaigns and regions are driving the most value, enabling data-driven budget allocation.

Get the worksheets and template¶

You can run this example in two ways:

• Single account: Run the entire example in one Snowflake account, where one account plays all three roles.

• Three accounts: Run the example across three separate Snowflake accounts in the same organization and the same cloud hosting environment, with each account playing a different role (publisher, advertiser, identity partner).

See instructions to upload a SQL worksheet into your Snowflake account.