Read the data never the customer.
A micro-firewalled schema abstraction layer sits in front of every publisher. You segment against semantic headers and clusters; raw records stay inside the publisher's own environment.
Transaction ledgers, CDRs or booking history connect to the publisher's own instance of the abstraction layer. Nothing leaves their cloud.
PII is removed, identifiers hashed, and records partitioned into macro cohorts against aggregate thresholds.
Only non-PII headers surface in your builder — and only while that one publisher is selected.
Queries return sizes and cohort definitions. No raw row ever crosses the firewall or blends with another publisher.
| Publisher | Source system | Refresh | Records | Consented | Data residency |
|---|---|---|---|---|---|
| Bank Partner | Core transaction ledger | Daily 02:00 | 4.2M | 82% | Publisher VPC (Mumbai) |
| Telco Partner | CDR + recharge ledger | Hourly | 8.4M | 68% | Publisher VPC (Pune) |
| Airline Partner | Booking + FFP history | Daily 04:00 | 1.8M | 91% | Publisher VPC (Delhi) |
| Retail Partner | POS + loyalty events | Every 15 min | 2.6M | 74% | Publisher VPC (Bengaluru) |
| Header | Origin | Grain | PII |
|---|---|---|---|
| monthly_spend_band | Derived | Cohort | Stripped |
| mcc_affinity_top3 | Derived | Cohort | Stripped |
| tier_status | Actual | Cohort | Stripped |
| city_cluster | Actual | Cohort | Generalised |
| recency_bucket | Derived | Cohort | Stripped |
| consent_flag | Actual | Row (internal) | Hashed id |
Data can be ingested from systems outside the LR ecosystem. Wherever it resides, the abstraction layer is what advertisers query — so first-party advertiser data can also power self-led segmentation, campaigns and loyalty without ever mixing with publisher records.