Skip to content
A selective update card passes a metal gate into matching glass contact records while neighboring records remain intact.
Outreach operations7 min read

CRM Import Updates: Keep Good Data Intact

Update existing CRM contacts without losing sales context. Plan matching, field changes, blank values and recovery before importing a refreshed lead list.

A refreshed lead spreadsheet can improve a client’s CRM—or replace the context the sales team actually needs. A newer job title may be useful. A blank owner field, an outdated status or a recycled campaign label may not be.

Before importing updates, decide which records the file may touch and which properties it may change. Then test those decisions against a small set of existing records. This guide uses HubSpot examples and highlights one Salesforce difference; the workflow is a recommendation, not a feature of AllProfiles.

1. Define the update before opening the importer

Write a one-sentence change request: “Update verified job titles and profile URLs for these existing contacts; preserve owners, conversation status and opt-outs.” That is more useful than “import the refreshed list.”

Name an operator and a reviewer. Give the file a version and record when the information was checked. Keep research notes separate from approved replacement values: “possibly moved to another company” should not become a confirmed employer because it occupies a company column.

Split the work into three queues: updates to known records, genuinely new records, and ambiguous matches. This keeps an unresolved identity question from turning into a new contact during routine maintenance. For the earlier identity-resolution step, see our guide to preventing duplicate leads across campaigns.

2. Match the record, not just the person’s name

Export the target records from the correct client CRM and retain their record identifiers. Keep each identifier attached to its row throughout research and review. A familiar name, a changed email address or a current profile URL is not your own proof that two records are the same person.

In HubSpot, an update-only import requires a unique identifier. Its single-object importer supports Record ID, contact Email and Company domain name for their relevant objects; a supplied Record ID takes precedence over other unique identifiers. Map it deliberately. HubSpot’s single-object import guide documents those options.

For this workflow, use the exported Record ID when you are updating known records in that same CRM account. Do not reuse IDs from another client’s account. Keep unresolved rows out of the upload, and separate record identification from any proposed changes to email or employer.

Choose Update existing records only when that is the assignment. If HubSpot cannot match an identifier in an update-only import, the affected row is not imported as a new record. Review that error rather than switching the whole file to create-and-update to make it disappear. See HubSpot’s import error reference.

3. Give every column a change rule

Do not map every column just because it exists in the spreadsheet. Build a short field plan before uploading:

Field in the working fileRecommended treatmentEvidence or review needed
CRM Record IDMatch the existing recordExport from the intended CRM account
Job title or profile URLReplace only approved valuesSource and date checked
Record ownerExclude from this research updateSeparate assignment decision
Conversation statusExcludeSales team controls the live conversation
Opt-out or suppression fieldsExclude from enrichmentSeparate, controlled preference workflow
Research notesKeep outside the import unless intentionally mappedDistinguish observation from confirmed fact

“Exclude” means leave the column out of the upload or explicitly skip it in mapping. A populated column can still contain stale data. A narrower file is easier to review than a complete export with a few cells changed.

HubSpot offers Don’t import column and Don’t overwrite during mapping. These are different decisions: one skips the column; the other protects existing property data. Its multi-object documentation adds a useful nuance: overwrite protection also covers a property that had a value previously but is currently empty. Do not assume it will refill every blank you see. HubSpot’s overwrite guidance explains the distinction.

4. Treat an empty cell as a question

Separate three intentions: leave unchanged, set a value, and deliberately clear a value. An empty spreadsheet cell is not a universal command for the third.

HubSpot’s file importer ignores blank cells; they do not clear an existing property. Its documented alternatives include manual editing or the Edit records workflow action. HubSpot’s file-format guide describes this behavior.

Salesforce’s Data Import Wizard also treats empty cells as no change, but documents #N/A for clearing supported fields, with exceptions for checkboxes and lookup fields. That is specific to that tool. Do not copy the marker into another CRM—or assume its API behaves identically. Verify the chosen operation and field type first. Salesforce’s clearing-field guidance gives the details.

Keep deliberate removals in a separate reviewed change set. This makes the difference between “we did not research a phone number” and “we intend to remove this phone number” visible.

5. Test outcomes with a small, representative batch

Choose records that exercise your decisions, not just the first few rows. Include an existing populated field, a blank field, a protected property and an identifier that you have deliberately held for review. Keep that unresolved row out of the actual upload.

Illustrative example: an agency refreshes three contacts in a client’s HubSpot account. The IDs below are made up. The expected outcomes assume correct mapping, update-only mode and no competing writes.

RecordApproved changeExpected result to inspect
41001Change title from “Manager” to verified “Operations Director”New title saved; owner and conversation status unchanged
41002Leave phone cell empty; update verified profile URLExisting phone preserved; URL updated
41003Keep Department protected with Don’t overwriteExisting department preserved despite a populated incoming department

Apply Don’t overwrite to the entire Department column in this test; it is a column-level mapping choice, not a per-row toggle. Open the records after the test and compare actual values with the plan. Also inspect anything your team depends on downstream: saved views, ownership queues and scheduled follow-ups. Do not infer that an import preserved them merely because the row count looks right.

If a field behaves differently from the plan, stop the remaining batch, document what happened and adjust the mapping or operation. Repeat the small test after the correction.

6. Reconcile results before using the refreshed list

An import finishing is the start of verification. Compare the intended changes with the result summary, download errors and inspect the rejected rows. Keep missing identifiers separate from invalid property values; they require different fixes.

ResultNext action
Intended records updatedSample the saved values and protected fields
Unexpected records createdStop further imports and investigate mode or matching
Rows rejectedCorrect the specific cause before retrying those rows
Values unchanged unexpectedlyReview mapping, overwrite protection and property rules

Keep the original export, approved upload, mapping decisions and result files together. Treat them as one change record. A retry should target the unresolved work, not silently reapply the whole research sheet.

7. Plan recovery before the full import

Do not assume “delete import” means “undo changes.” HubSpot explicitly says deleting records does not reverse updates an import made to existing records. Its import-history guide explains what can be reviewed and managed afterward.

A before-export helps you reconstruct prior values, but restoring it wholesale can overwrite legitimate changes made since the import. If recovery is needed, compare the affected fields again and prepare a reviewed correction. Assign someone to decide which later changes must survive.

Finally, give the refreshed list a clear handoff: what changed, what was preserved, which rows remain unresolved and who owns follow-up. Our lead handoff checklist covers that transfer of responsibility. For spreadsheet encoding and column preparation before this stage, use the CSV import preflight.