Data CloudIntermediatetroubleshootingsegmentsdata-quality
Troubleshoot segment count discrepancies vs source queries
Real World Scenario
A segment defined as "customers with 2+ purchases in 90 days" returns 45k members but a Snowflake query on source data shows 62k.
Expected Answer
• Verify ingestion completeness: missing batches, failed records in quarantine
• Check identity resolution: duplicate individuals counted once in DC vs twice in source
• Confirm segment refresh timestamp vs Snowflake query as-of time
• Review consent and exclusion filters applied in segment but not source SQL
• Validate DMO mapping dropped records failing validation rules
• Compare calculated insight logic to raw SQL aggregation definitions
• Use Data Explorer and query APIs to trace sample individuals in vs out
Follow-Up Questions & Answers
Click to expand — each follow-up includes a direct, interview-ready answer
Direct answer: Verify ingestion completeness: missing batches, failed records in quarantine Also consider: Check identity resolution: duplicate individuals counted once in DC vs twice in source In practice: Confirm segment refresh timestamp vs Snowflake query as-of time Balance speed of delivery with maintainability.
Architect Perspective
Discrepancies are usually identity, consent, or timing—not platform bugs. Architects build reconciliation dashboards proactively.