How to prepare a CSV for a CRM import without creating duplicate contacts
A practical pre-import workflow for checking headers, identity fields, formatting, blank records, and duplicate contacts before a CSV reaches your CRM.
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.
1. Keep the export and make a working copy
Save the original export unchanged, then work from a dated copy. If an import changes records unexpectedly, the untouched file is your reference for what came from the source system.
2. Match the destination schema before cleaning rows
Write down which column will identify a contact or account, then compare the CSV headers with the destination fields. Remove unused columns only after you have confirmed they are not needed, and rename headers consistently rather than carrying both Email and email_address into the same file.
3. Normalize the fields used to recognize the same person
Trim leading and trailing spaces, use one case convention for email addresses, and preserve identifiers such as account numbers as text. Do not normalize every column blindly: notes, names, and free-form addresses may need their original spelling for display or audit purposes.
4. Find exact and near-duplicate contacts before import
Check for exact duplicates on the fields that define identity, such as an email or customer ID. Then review near-duplicates separately: [email protected] and [email protected] may or may not represent the same person, while a shared email domain is not enough evidence to merge two records. See our duplicate-removal guide for the distinction.
5. Check the rows that are not useful contacts
Remove blank rows and obvious header rows repeated inside a merged export. Flag records missing the identity field for review instead of guessing which person they belong to. Also inspect uneven row lengths and quoted commas before importing, because a shifted column can put a phone number or note into the wrong field.
6. Validate a small import first
Export a small, representative sample to a new file and use it for a dry run or limited import if the destination supports one. Compare the resulting field values and contact count with the source sample before importing the full copy. This catches mapping and formatting mistakes while they are still easy to undo.
See also: our CSV cleanup checklist, schema-validation guide, and backup/versioning guide.
See CSV Repair & Deduper (offline browser tool) — $19 one-time