A professional walking up the stairs of a bright office atrium

A CRM migration checklist.

Twenty-five steps, in order.

A CRM migration checklist runs in six stages: set the reason and owners, design the new platform with field and association maps, clean and deduplicate the data, test with trial loads and reconciliation, cut over with a freeze and final load, then track adoption and data quality. Add AI agents only after that.

Who should careCTO, COOMetric it movesData quality, adoption rate

At a glance

The checklist on one page.

  1. Stage 01 · 5 stepsBefore you moveReason, owners, inventory, scope, data protection
  2. Stage 02 · 5 stepsDesignDefinitions, field and association maps, access, rebuild list
  3. Stage 03 · 3 stepsPrepare dataDeduplicate, standardise, keep source IDs
  4. Stage 04 · 4 stepsTestTrial loads, reconcile, user and integration tests
  5. Stage 05 · 4 stepsCut overFreeze, final load, training, support
  6. Stage 06 · 4 stepsAfter go-liveAdoption, data quality, archive, then AI

A working model of the checklist below: the red square is the final step, adding AI agents once the data is clean.

Use this as a working list. Each step has an owner and a sign-off before the next begins.

Before you move anything

  1. Write the reason. The business outcome the move must deliver, and the date it matters by.
  2. Name the owners. One executive sponsor, one project lead, and a data owner for each market.
  3. Inventory the old system. Objects, fields, record counts, workflows, integrations, reports and users.
  4. Decide what not to move. Stale records, unused fields and retired processes stay behind or go to an archive.
  5. Check data protection. Legal basis and consent for each market, and where the new platform hosts data.

Design the new platform

  1. Agree definitions. Lifecycle stages, deal stages and required fields, signed off by marketing, sales and service.
  2. Map every field. Old field, new property, type, allowed values and transformation rule.
  3. Map associations. How contacts, companies, deals, tickets and custom records link, so history keeps its context.
  4. Map users and access. Teams, permissions and record owners by market.
  5. Rebuild, do not copy. List each workflow and report, then rebuild only those still needed, in the new platform's way.

Prepare the data

  1. Deduplicate. One record per company and contact, across markets and old systems.
  2. Standardise. Countries, phone formats, currencies, industries and picklist values.
  3. Keep source IDs. Store the old system's record ID on every migrated record, so you can trace and fix.

Test before cutover

  1. Trial migration. Load a sample, then a full copy, into a test environment.
  2. Reconcile. Record counts, totals such as open pipeline value, and spot checks of history on key accounts.
  3. User testing. Each team runs its real daily tasks and signs off.
  4. Test integrations. ERP, website forms, messaging channels and billing, end to end.

Cut over

  1. Freeze the old system. Read-only from an agreed time, so nothing is lost in between.
  2. Final load and delta. Move the last changes, reconcile again, then switch integrations.
  3. Train by role. Short sessions on the tasks each person does, in their language.
  4. Support the first weeks. A named contact and a daily issue list.

After go-live

  1. Watch adoption. Logins, records updated and deals moved, by team and market.
  2. Check data quality. Duplicates, missing fields and stale owners, weekly at first.
  3. Keep the archive. Retain the old export for as long as policy requires, then retire the old system.
  4. Then add AI. Agents only after the data is clean and the process is agreed.

Moving to HubSpot from a specific platform? See Salesforce, Zoho, Dynamics or Pipedrive.

Plan the move before the data moves.

A strategy call scopes your inventory, data map and cutover plan.