How to migrate CRM data without losing history
Field mapping, deduplication rules, and the dry-run process that catches problems before they reach production.
Most CRM migrations lose something. Usually it is activity history, sometimes it is the relationship between a contact and their company, and occasionally it is a duplicate mess that takes a year to untangle. All of it is preventable with a dry run.
Map fields before you export
Open both systems side by side and write the mapping down: source field, destination field, transformation, and what happens when the value is empty. The fields with no obvious destination are the interesting ones — they usually encode a business rule nobody documented.
Decide your deduplication rule in advance
Email is the usual key, but it fails for households, shared office addresses and B2B contacts who changed jobs. A composite rule — email, or phone plus surname — catches more without merging people who are genuinely different. Whatever you choose, apply it consistently and keep a log of every merge so it can be reversed.
Migrate in this order
Users first, so ownership can be assigned. Then companies, then contacts linked to them, then deals, then activities and notes. Attempting it in any other order produces orphan records that have to be relinked by hand.
Run it twice
The first run goes into a sandbox. Compare record counts, spot-check fifty records against the source, and have someone from sales look at their own accounts — they will notice missing history that no count reconciliation catches. Fix, then run the real one.
Keep the old system read-only
For thirty days, not thirty hours. Something will be missing, and having the source available turns a crisis into a five-minute lookup.
Want this running in your business?
Skyra ships with the templates, sequences and automations described here. Try it free for 14 days.