Compare: adv-contradiction-semantic-control-scope

Write the aggregated results to **BigQuery**, your analytics warehouse.

Reasoning:

- **Fit for purpose.** Rollups are analytical artifacts consumed by dashboards, reports, and ad hoc analysis. BigQuery's columnar storage and MPP execution are built for scanning and aggregating large tables; Postgres is optimized for transactional row-level access.
- **Isolation.** A heavy nightly batch writing into your production OLTP database competes for I/O, cache, locks, and vacuum resources with live application traffic. Keeping it in BigQuery protects production latency.
- **Source data locality.** If the raw event/analytics data already lives in BigQuery, doing the aggregation there avoids extracting large volumes out and back in. You can compute the rollup with a scheduled query or `INSERT ... SELECT` into a rollup table/materialized view.
- **Cost/scaling.** Storage and compute scale independently in BigQuery, and rollup tables shrink scan costs for downstream queries.

When it *would* make sense to also write (a slice) back to PostgreSQL:

- The production application itself needs to read these aggregates at low latency to serve user-facing features (e.g. showing a account-level usage counter in the app UI).
- In that case, compute in BigQuery, then sync only the small, needed subset back to Postgres.

Default: BigQuery for the rollup output; mirror to Postgres only for specific operational read paths.

Write the aggregated results to **BigQuery**.

Reasoning:

- **Purpose fit.** Rollups are analytical artifacts — pre-aggregated metrics meant for reporting, dashboards, and ad-hoc analysis. That's exactly what the warehouse is for.
- **Keep OLTP lean.** Your PostgreSQL production database should stay focused on transactional workloads. Adding aggregate tables invites bloat, extra vacuum pressure, contention, and storage growth that competes with live traffic.
- **Query patterns.** BigQuery's columnar, scan-heavy engine is built for aggregation queries over large ranges; Postgres is optimized for selective indexed row lookups.
- **Source of truth.** If the rollup reads from raw data that already lands in BigQuery, keeping the pipeline within one system simplifies lineage, backfills, and reprocessing.

Exception: if a specific rollup exists to serve the *application itself* in real time (e.g., a user-facing "your stats" widget that needs low-latency point reads), write that particular result back to Postgres — or to a cache/serving store. But the default for a nightly analytics rollup is BigQuery.