Migrating From Another CRM How to move off HubSpot, Salesforce, Zoho, Pipedrive and the rest without creating a mess you spend months cleaning.

Operations, Data, Apps and Import. Sixteen connectors, covering the tools most workspaces arrive from.

Category	Connectors
CRM	HubSpot, Salesforce, Zoho, Pipedrive, Freshsales
Work management	Monday, Notion, Jira, Linear, Asana, Trello
Finance and commerce	Stripe, Shopify, QuickBooks
Documents and people	PandaDoc, Zoho People

Before You Import Anything

Migration goes wrong in preparation, not execution. Four things first.

1. Create your custom fields. Anything your old system holds that this one does not ship with needs a field to land in. A column with no target is a column you import again later. See Custom Fields.
2. Set up your pipelines and stages. Deals arrive needing a stage. Without your pipeline defined, they all land somewhere wrong and you re-stage them by hand.
3. Create your hubs, if you trade from more than one entity. Records attach to a hub, and retrofitting that is painful. See Hubs in the Settings section.
4. Decide what not to bring. This is the step everyone skips. A CRM you have run for five years contains dead leads, duplicate companies and contacts who left. Migration is the one clean chance to leave them behind.

Order Of Import

Parents before children, always:

1. Companies - so contacts have somewhere to attach.
2. Contacts - so deals have people.
3. Deals - now they can link to both.
4. Everything else - invoices, tasks, notes.

Importing deals first produces deals with no company and no contact, which is the single most common migration mess. The links are not repaired retroactively.

Running The Import

1. Start the connector and authorise, or supply an API key from that tool.
2. Choose what to bring, where the connector supports selection.
3. Map fields. Your custom fields appear as targets here.
4. Review before confirming.
5. Import.
6. Check the Imports tab in Data Management for what actually happened.

Field Mapping

Take the time here. Some notes from experience:

• Owner fields need people to exist first. Import your team before your records, or ownership lands empty.
• Dates carry format ambiguity. Confirm which is day and which is month on a sample before importing thousands.
• Picklists rarely match. Your old system's stage names are not your new ones. Map deliberately rather than accepting a default that collapses distinct values.
• Do not map notes into a text field if the old system had many per record. They belong as separate entries.

After The Import

Do not build anything on the data yet.

1. Reconcile counts. Compare against your old system.
2. Spot-check twenty records by hand. Counts can be right while content is wrong. A column landing one position off passes every automated check and corrupts everything.
3. Work the queues in Data Management - no owner, no company, duplicates.
4. Check your pipeline. Are the deals in sensible stages?
5. Only then reactivate automations and build reporting.

Sync Versus One-Off

Some connectors sync continuously; others import once. Check which yours does rather than assuming.

If you are migrating, a one-off import is usually what you want - you are leaving the old system. Continuous sync is for tools you will keep running alongside, and it brings its own question: which system is authoritative when both change the same record.

Running A Pilot

For anything over a few thousand records, import a slice first.

1. Import one hundred companies and their contacts.
2. Check the mapping properly.
3. Fix what is wrong.
4. Import the rest.

The hundred records take an hour to verify. Getting the mapping wrong across forty thousand takes a month to unpick.

Common Questions

My import partly failed. Fix the rejected rows and import only those. Re-running the whole file duplicates what already succeeded.

Can I undo an import? Not as one action, which is why the pilot and the review step matter.

Should I import historical closed deals? Yes if you want historical reporting; it is the only way to get a year-on-year comparison. Import them closed so they do not appear in your live pipeline.

How long does a migration take? Preparation is most of it. The import itself is quick; the field mapping and the clean-up are the work.