Output Targets
Attach transformation outputs to user, event, session, or account properties in the Data Catalog.
Last reviewed September 15, 2026
Every model can attach its output to a specific property in the Data Catalog. Output targets define where the computed values appear. Downstream tools can then use them: reports, segments, destinations, and the Agent.
Target types
| Target | Where it appears | Example |
|---|---|---|
| User property | User Properties in the Data Catalog | predicted_ltv, churn_risk_score |
| Event property | Event Properties in the Data Catalog | attributed_channel, sentiment_label |
| Session property | Session-level attributes | session_value_tier |
| Account property | Account- or organization-level attributes | account_health_score |
Select the target type when you create or edit a model.
Rule-based vs precomputed
Both execution modes write to the same catalog properties:
| Mode | When values are computed | Best for |
|---|---|---|
| Rule-based | At query time (SQL CASE) | Simple logic, always current |
| Precomputed | On a schedule or after source sync | Heavy aggregations, ML scores |
For the user, both appear as properties in the Data Catalog. The difference is the internal execution timing.
Using outputs downstream
When a model output is attached to a catalog property, you can use it in:
- Audiences: build user lists and export them to ad platforms
- Destinations: sync properties to external tools
- Reports: include in cross-channel reporting
- Agent: query and analyze in conversation
- MCP tools: access programmatically from external AI tools
Related
- Data Studio
- Customer 360: built-in calculated property templates
Last updated on