CRM IMPORT
Bring your customer list across with a preview first.
Map spreadsheet columns to CRM fields, review the proposed changes and keep an import batch that can be checked and, where appropriate, undone.
| Full name | Email address | Organization |
|---|---|---|
| Alex Winter | [email protected] | Studio Nord |
| Morgan Reed | [email protected] | Northline |
Review before importing
Migration is a data decision, not just a file upload
An old customer list may combine years of naming conventions, blank columns and repeated people. Uploading it directly into a live CRM can turn a messy spreadsheet into a messy shared database. Tsumira’s import flow separates reading the file, mapping its columns, preparing a plan and applying the approved batch. The preview gives the person responsible a chance to understand what will be created, updated or skipped.
Read and map before changing customer records
The importer detects supported text encodings, including UTF-8 and Japanese Shift_JIS exports, and maps headers to the declared target fields. Review that mapping, especially where a source column could mean several things. A telephone number must not land in a note simply because the headers differ. The contact target accepts a name, reading, phone, email, address and note; company and deal imports have their own targets and field meanings.
- Encoding-aware file reading
- Column-to-field mapping
- Separate contact, company and deal targets
Column mapping
Review before importing
Understand updates, warnings and skipped rows
The contact plan normalizes values using the same CRM rules as ordinary records. Existing contacts can match by email, normalized phone, or an exact name-and-reading combination. The plan also accounts for repeated contacts within the uploaded list. Missing names are skipped. If a Japanese reading is absent, the written name is used as a flagged fallback rather than a guessed pronunciation. Review these outcomes before accepting the batch.
- Proposed creates and updates
- Matching within the existing data and the file
- Visible missing-reading warnings
Treat undo as a controlled correction
An applied batch retains the information needed for its undo path. That is useful when the mapping or source file was wrong, but it should not be treated as permission to skip review. Undo requires the relevant update and delete permissions. If a record has changed after the import, the current implementation preserves the concurrent change rather than blindly replacing someone else’s work with an earlier snapshot. Read the undo outcome as carefully as the original preview.
- Batch-based correction
- Permission checks for undo
- Preservation of subsequent concurrent changes
A concrete workflow
Prepare a representative file
Check a small set of real column patterns and remove information that should not be imported. Keep an unchanged source copy.
Review encoding and columns
Confirm that names are readable and each column maps to the intended CRM field.
Inspect the proposed batch
Review new records, updates, skipped rows and warnings, including repeated contacts and missing readings.
Apply and verify
Apply the reviewed plan, inspect several resulting records and use the batch correction path if a genuine import error is found.
An illustrative working example
Illustrative workflow: replacing a shared customer spreadsheet
An agency wants to move a long-used customer list into CRM. The list contains both email addresses and telephone numbers, plus several repeated contacts. The operations lead uploads a representative export, checks the column mapping and reviews which rows update existing people. A missing name is corrected in the source. Only after checking the preview does the lead apply the batch and ask the customer team to review a sample of records.
Questions about this topic
Does import immediately write every row?
The flow provides a mapping and planning stage before applying the import. Review the intended creates, updates, warnings and skips before accepting it.
How are duplicate contact rows recognized?
Contact matching uses email, normalized phone, then the exact written name and reading together. This is a defined matching rule, not a guarantee of resolving every real-world identity.
What happens without a Japanese name reading?
The contact import uses the written name as a fallback and records a warning. It does not invent or promise an accurate phonetic reading.
Can undo overwrite a colleague’s later edit?
The current version checks the post-import values before restoring or removing records. Subsequent concurrent changes are preserved and reported rather than blindly overwritten.
Explore connections around this workflow.
Ideas for intended connections. Check the actual availability, permissions and supported data for each app before using a scenario.
Microsoft Excel
Workflow concept: Prepare a clean import workbook with explicit field and customer matching decisions.
HubSpot
Workflow concept: Review exported contact and company rows before a controlled migration.
Google Sheets
Workflow concept: Resolve missing names and ambiguous identifiers before uploading the table.