A migration can complete with zero technical errors and still be wrong. Contacts can attach to the wrong organizations, owners can be missing, transformed fields can contain unexpected values, and historical records can be incomplete. Validation needs to test business consistency as well as execution success.
Validate before production
Use staging to check required fields, identities, value formats, duplicate rules, relationship references, and unresolved mappings. Pre-production validation prevents known bad data from becoming a production cleanup problem.
Validate execution completeness
Track eligible records, successful writes, failed records, retries, skipped items, and explicit dispositions. The execution ledger should make it possible to account for every source item in scope.
Key takeaway: The useful test is whether the process preserves business meaning, ownership, evidence, and the ability to verify what happened. A faster handoff is not enough if those controls disappear.
Reconcile relationships and ownership
Compare whether contacts still belong to the right organizations, opportunities retain owners, activities attach to the expected records, and historical links remain intact. These checks catch errors that row counts alone cannot detect.
Require business verification
Automated reconciliation can identify discrepancies, but final acceptance should include a permitted human reviewing the result and outstanding exceptions. Completion is an evidence decision, not only a worker state.
Practical checklist
- Pre-production validation
- Execution accounting
- Count reconciliation
- Relationship reconciliation
- Ownership checks
- Transformation checks
- Exception review
- Human sign-off
See the connected product context
This guide targets a narrow operating problem. The related Infrakinetic capability page shows how that problem connects to the wider product architecture and adjacent workflows.
Explore the related capability