Begin with the CRM's requirements and the purpose of the import. Prepare and review the file separately, then use the CRM's own import process. Data cleanup helps with preparation; it does not guarantee that an importer will accept every record.
Decide whether you are creating or updating records
A new-record import and a bulk update have different consequences. For new records, you may want to exclude people already represented in a CRM export. For updates, existing record identifiers may be essential. Do not filter out matched records just because matching is possible.
Write down the object you intend to import: contacts, companies, leads or another supported record type. Check the destination's current import settings directly before making changes to the source file.
- Confirm create versus update behavior.
- Identify required fields and accepted values.
- Preserve the record identifiers required for updates.
Check columns and record structure
Inspect the header row and representative records. Map the source columns to the destination fields on paper or in your import checklist. A column named Company in one export may correspond to Organization in another, but the label alone does not establish the intended relationship.
Watch for identifiers with leading zeros and date strings whose order is ambiguous. Agree the required representation instead of letting a generic formatting step decide. Completely blank rows are different from records with one required field missing.
Clean values and review duplicates
Trim surrounding whitespace and standardize selected values to the conventions your destination expects. Review proposed replacements before applying them, especially when changing a status or category label.
Then decide what makes a record unique. An exact repeated row can be redundant; two rows sharing a contact email may contain different information. Review duplicate groups and the proposed survivor before removal. A record with more non-empty fields is not automatically more accurate.
Compare with a relevant CRM export
For a new-record import, compare the incoming file with a recent export covering the CRM records you need to check. Map matching fields and use explicit normalization rules. A restricted or stale export can make existing records appear new because they are absent from the reference.
Review new records, matches and records only in the reference. Filtering the incoming file does not change the live CRM. For a bulk update, use the comparison to understand overlap while retaining the records and identifiers required by your update plan.
Verify the handoff before starting the import
Inspect changed values and removals, then compare the final count with the starting count. Export a separate clean CSV and check it against the destination requirements. Keep the original source and any supporting outputs you need.
Use the CRM's available preview or validation process before committing an import where that process is provided. Review its reported errors and results afterward. RowDesk prepares files; it does not write back to the CRM or replace those destination-side checks.
- Confirm columns, required values and identifiers.
- Check create/update settings in the CRM.
- Verify the final CSV and record count.
- Keep a copy of the source and reviewed output.