Data Migration & Quality
Move operational data into SupaChrome with dependable mapping, deduplication and validation.
Open interactive lesson and quizWhat you will learn
- Prepare import-ready source files
- Map standard and custom fields
- Resolve duplicates and relationships without losing data
Three teams maintain separate CSV files with different headers, duplicate contacts and company names that do not match exactly. You need one safe import plan.
Lesson 1
Profile before you move
Count rows, identify keys, inspect blanks and document source meanings before importing. A migration succeeds when business meaning survives, not merely when every row loads.
Start with a sample and a source inventory: record count, unique identifiers, mandatory values, date formats and relationship keys. Inspect several rows from the beginning, middle and end of the file. A column named Date could mean creation date, purchase date or renewal date; its meaning must be decided before mapping.
Document which source is authoritative when values conflict. Preserve a copy of the original file and create a separate cleaned version. Check for repeated headers, hidden spaces and values such as N/A in numeric columns. Resolve those issues before uploading a large batch.
Lesson 2
Map fields intentionally
Match each CSV header to a destination field, including custom fields and relationship lookups. Leave irrelevant columns unmapped rather than storing noise.
Select the destination object first, then map each source header to a field with the same meaning and compatible type. A Membership Start column belongs in a date field, while Membership Level should use the configured picklist choices. Multi-select values must correspond to the allowed choices; they are not a way to create several object records in one row.
In Data Handler → Import, download the destination template, add your business fields, then upload it. SupaChrome generates record IDs; do not add an ID column. The mapping table shows each source header and sample value beside the destination field. Map every required field once, validate the file and review the mapped three-row preview before importing. For relationships, use an existing record ID or supported matching email, name or SKU for that lookup. Review company creation and linking when importing contacts; ignored columns should be deliberate. In Data Handler → Export, choose an object and fields, preview five rows and download the BU-scoped CSV. Export requires Data export plus View and Export access for that object; import remains a separate permission and workflow.
Lesson 3
Choose a duplicate policy
Use a stable identifier such as normalized email, domain, SKU or record ID. Skip protects existing records; update is useful only when the source is trusted and authoritative.
Duplicate handling must follow a clear matching key. For products, a SKU is often more stable than a display name. For people, email may work when the business does not share addresses. Before choosing Update, identify which existing values the source is allowed to overwrite. Older source files can otherwise replace newer CRM information.
Test Skip and Update on a small file containing one new record and one known existing record. Compare the results with your intended policy. SupaChrome creates its own record ID. For CSV, existing Contacts match by email, Companies by domain or name, Products by SKU, Deals and custom records by name, and Cases by case number if supplied. Tasks import as new records. An external integration can use its own supported external identifier through its authenticated API; do not put that identifier into the system record ID column.
Lesson 4
Validate and reconcile
Dry-run first, fix rejected rows, then reconcile source counts against created, updated and skipped totals. Review a sample of linked records after loading.
A pre-import validation catches structural errors before commit, but you still need to review the final results. Record how many source rows were created, updated, skipped and rejected. Reconcile these categories to the number of rows submitted, then inspect a sample of the saved records and their relationships.
Download the import error report for the exact row numbers and correct only the affected rows when practical. Validation also checks required custom fields, picklist choices and common field formats. Do not repeatedly import the whole file without understanding the duplicate policy. If the totals differ from your expectation, check rejected values, missing permissions, relationship ambiguity and filters in the destination view before assuming data was lost.
Practice · Run a governed contact import
- Download the object-specific template
- Map source headers and enable company creation
- Validate, correct rejected rows and review the completed import job