Glossary
What is data reconciliation?
Data reconciliation compares records in two systems to confirm they match and to find missing, extra or different records.
Data reconciliation is the check that an integration did what it should: the records in the target system match the source system. A reconciliation report compares the two, usually by a shared identifier, and lists records that are missing, extra or different, so problems are found before users or auditors find them.
How does it work?
- Pull the relevant records from the source and the target.
- Match them on a stable key, such as employee ID.
- Compare the mapped fields and classify each record as matched, missing, extra or mismatched.
- Review differences, fix the cause and rerun until the report is clean.
Reconciliation is also the most useful artefact to show a customer at go-live. A clean report, record by record, is stronger evidence than a demo that the integration is complete and correct.
Where it shows up in integrations
- Go-live A dry run with a reconciliation report shows the integration is correct before it writes to production.
- Ongoing monitoring Scheduled reconciliation catches missed webhooks and silent failures.
- Audits Compliance teams need evidence that training and access data is complete.
Common pitfalls
- Comparing record counts only, which hides mismatched fields.
- Reconciling once at launch and never again.
- Matching on names or emails instead of a stable ID, which reports false differences.
- No owner for resolving differences, so reports pile up unread.