Why duplicate records appear after a CSV import (and how to prevent them)
A practical pre-import checklist for finding repeated contacts or customers before a CSV creates duplicate records in a CRM or database.
Before you work through this by hand: you can check your own CSV free and see which of these problems your file actually has — duplicate rows, near-duplicates, broken quoting, encoding damage. It runs in your browser tab, nothing is uploaded, and there's no signup or email.
Why an import creates duplicates
Most import tools match on one selected key, such as an email address or customer ID. If that key is missing, has different capitalization, contains invisible spaces, or is not mapped to the destination field, the importer may treat an existing record as new. A repeated export can also contain the same person more than once before the file reaches the import screen.
Pre-import checks
- Keep an untouched backup. Save the original export and work on a copy so the source row count remains auditable.
- Choose the real identity key. Prefer a stable customer ID or a verified email field; do not use a shared company domain, display name, or phone number alone.
- Normalize before matching. Trim leading and trailing spaces, standardize case for comparison, and preserve the original value in a separate column for review.
- Find exact and near-duplicates. Exact duplicates are easy to group, but a changed capitalization, typo, or alternate phone format can hide the same person in a second row.
- Run a small import first. Import a representative sample into a safe test segment, then compare created, updated, skipped, and rejected counts with the source.
Do not merge conflicts automatically
Two rows with the same email can disagree on an address, owner, consent status, or job title. Flag those rows for a human decision instead of silently keeping the first value. Keep a review log and export to a new file so the original remains recoverable.
CSV Repair & Deduper runs locally in your browser: it flags exact and near-duplicate candidates, keeps original values visible, supports review and undo, and exports a separate cleaned CSV without uploading your file. Check your own CSV free before deciding whether the paid review workflow fits.
See also: our CRM import preparation guide, case-normalization guide, and conflicting-data guide.
See CSV Repair & Deduper (offline browser tool) — $19 one-time