Scrunch supports two primary patterns for loading data into custom warehouses or data marts: snapshot-based or event-based loading.
1. Snapshot Metrics (Query API)
Pre-aggregated, analytics-ready tables
This approach uses the Query API to load daily or periodic snapshot tables that already contain calculated performance metrics.
Data is aggregated by Scrunch (e.g. by date, prompt, platform)
Metrics such as brand presence, sentiment, position, and competitor presence are pre-calculated
Ideal for building fact tables that power dashboards and recurring reports
Minimizes transformation logic in your warehouse
This model works well if you want:
Fast time-to-value
Consistent metrics aligned with Scrunch reporting
Only need specific views or pivots of the data
Lower ETL complexity
If you don't have an existing data warehouse or enterprise reporting tool, our Looker Studio Connector can build custom reports on top of the Query API in real time — no ETL jobs needed.
2. Fine-Grained Events (Responses API)
Raw response-level events with maximum flexibility
This approach uses the Responses API to ingest every individual AI response as an event in your warehouse.
One row per response (no duplicates)
Full response text, metadata, competitors, sentiment, and citations
Supports incremental loading via timestamps and pagination
Aggregations and metrics are computed downstream in your warehouse
This model works well if you want:
Full control over metric definitions
Custom rollups or advanced analysis
The ability to re-aggregate data as requirements evolve
Choosing the Right Model
Snapshot metrics are simpler and faster to implement
Fine-grained events provide flexibility and long-term analytical power
Reach out to your Customer Success Manager for additional support with your reporting program.