Data Cloud is hyperscale data layer; testing focuses on data flows + identity resolution.
Test areas:
- Ingestion: data flowing in from sources.
- Identity Resolution: unified profiles created correctly.
- Calculated Insights: computed metrics accurate.
- Activation: data flowing out to operational systems.
- Segmentation: criteria match correctly.
Test approach:
- Sample data ingested.
- Manually verify identity resolution outcomes.
- Statistical sampling of large datasets.
- End-to-end: data in source → unified → activated → in CRM.
Identity resolution validation:
- True positives: same person matched.
- False positives: different people merged (BAD).
- True negatives: different people kept separate.
- False negatives: same person not merged (incomplete).
Sampling:
- Random sample of unified profiles.
- Manual review for accuracy.
- Adjust resolution rules based on findings.
Common pitfalls:
- No identity resolution validation — incorrect merges go undetected.
- Test data not representative — production may differ.
- No statistical sampling.
Senior insight: Data Cloud testing is data-centric. Quality of unified profile matters more than feature presence.