Industry Examples
These are end-to-end blueprints showing how to model User and Event tables for common business types. Each example covers the recommended identifiers, the core tables (with CREATE TABLE schemas), do's and don'ts, and how BPP uses the data for each AI model.
Golden rules (apply to every example)
- User table: 1 row = 1 person, with at least one identifier.
- Event table: 1 row = 1 event, with exactly one timestamp (UTC) and one or more user identifiers.
- Prefer append-only history tables over mutable snapshots.
- Table and field names are illustrative — what matters is grain, identifiers, and timestamps.
:::note Naming convention
Source tables in these examples use plain names (e.g. orders, web_events). BPP creates its own enriched copies suffixed _bpp (e.g. orders_bpp) during reconciliation — don't use the _bpp suffix on your source tables.
:::
Examples
| Business type | Highlights |
|---|---|
| D2C E-commerce | Orders, web behaviour, basket JSON |
| Omnichannel Retailer | Online + in-store, loyalty ID anchor |
| B2B SaaS | CRM, subscriptions (MRR/ARR), product usage |
| B2B Lead Generation | Rich CRM contacts, deal stage history, enrichment |
| Car Dealer | Leads, test drives, aftersales |
| Publisher / Media | Newsletter, paywall, content consumption, UGC |