Two rows that look alike are not necessarily the same customer
Customers can share names, telephone numbers or addresses. Sometimes the same customer has two records with different IDs. Removing duplicates based only on a name risks combining two distinct buyers and misassigning orders.
A safer three-stage review
- Exact key duplication: identify any repeated stable customer IDs and count their occurrences. Do not auto-delete rows.
- Possible identity overlap: group similar names or contact fields as candidates, not confirmed matches.
- Human approval: retain each original row until the owner approves a merge rule and can trace every changed relationship.
Synthetic example
| Customer ID | Name | |
|---|---|---|
| C-017 | Alex Brown | alex1@example.test |
| C-018 | Alex Brown | alex2@example.test |
The names match, but the records have different IDs and email addresses. The correct first result is requires review, not an automatic merge.
Our before-and-after standard
We count records, identify exact-key duplicates, document proposed changes and list unresolved identities separately. A valid CSV format cannot by itself establish that two customer identities are equivalent.