Moving data from a legacy CRM into a new platform is one of the riskiest parts of any implementation. Most of the risk can be removed with careful preparation.
Before you move anything
- Profile every source. Understand volumes, field usage, formats and obvious quality issues.
- Agree what not to migrate. Old, incomplete or duplicate records may not deserve a place in the new system.
- Define the target model first. Mapping is easier when the destination is settled.
Mapping and transformation
- Document every source-to-target field mapping, including transformations and default values.
- Choose external ID fields so records can be upserted and traced back to the source.
- Decide how to handle relationships, owners and record types.
- Plan how files and attachments will move, and where they will be stored.
Trial runs
- Run at least one full trial migration in a sandbox.
- Reconcile record counts per object and spot-check key fields.
- Ask business users to validate real records they know well.
- Time the run so you can plan the production window.
Cutover
- Freeze changes in the source system, or plan a delta load.
- Disable automation that should not fire during the load, and re-enable it afterwards.
- Reconcile again in production before users start work.
- Keep the source system read-only for a defined period as a safety net.
After go-live
Expect questions. Keep a short log of data issues raised by users and resolve them quickly — early confidence in the data drives adoption.
This article is a draft for review.