How to migrate from Salesforce to HubSpot: 10 steps, in order

You're moving from Salesforce to HubSpot. Here are the ten steps in order, who owns each one, and how long the migration takes from kickoff to go-live.

Download a PDF version

Print it out and tick off each step as you go, or keep it open next to you while you work.

You're all set.
Open the PDF
Oops! Something went wrong while submitting the form.

Table of contents

Before you start

A named HubSpot admin and executive sponsor, buy-in across marketing, sales, and customer success, and HubSpot Enterprise if you want sandbox testing.

How do you move your data, automations, and integrations from Salesforce to HubSpot while your team keeps selling? You work through ten steps in a set order, and you give every step an owner. Here they are, along with the timeline: most go-lives take about 8 to 14 weeks from kickoff.

  1. Take inventory.
  2. Audit and clean the data.
  3. Plan the HubSpot architecture.
  4. Build in sandbox.
  5. Reconnect integrations.
  6. Import sample data and sign off the test.
  7. Deploy the setup to production.
  8. Train and enable the team.
  9. Go live and sign the checklist.
  10. Monitor after go-live.

If you're still deciding whether to move at all, start with Salesforce to HubSpot Migration: Should You Move? and come back here once the decision is made.

What has to be true before technical work begins

Five conditions belong in place before anyone builds anything.

  • Organizational buy-in is confirmed across marketing, sales, and customer success. Low buy-in produces a technically correct CRM that nobody uses.
  • Vendor negotiation is exhausted. If cost is the only reason to move, renegotiate with Salesforce first. A migration takes about 8 to 14 weeks of your team's attention.
  • Licensing is scoped. Seats, hubs, and the expected timeline are confirmed before you buy, and your HubSpot tier decides whether you get a sandbox.
  • A HubSpot admin is named, in-house or contracted.
  • An executive sponsor is named, someone who can remove blockers and communicate the change company-wide.

Your HubSpot tier changes where Steps 4 and 6 happen, because the sandbox (a separate HubSpot test account for building and testing) is included only with Enterprise.

HubSpot EnterpriseHubSpot Starter or Professional
SandboxIncludedNot included
Where you build and test (Steps 4 and 6)In the sandbox, a separate test account, so nothing touches your live dataDirectly in your live (production) portal. RevBlack builds and validates there under a controlled process.

Who is on this project

A migration succeeds when everyone knows what they own. If you're the RevOps lead running this, you're the migration lead.

RoleWhat they own
Executive sponsorThe business case. Removes blockers, communicates the change company-wide, and holds departments accountable for taking part.
Migration leadRuns the project end to end, from discovery through post-go-live monitoring.
Outgoing Salesforce adminKnows why the current system was built the way it was. Essential during inventory and architecture.
Incoming HubSpot adminOwns the platform after go-live. Involved throughout the build, so ownership is real from day one.
Department stakeholdersRepresentatives from marketing, sales, and customer success who validate that pipelines, automations, and reports match how their teams actually work.
IT and integrations ownerCredentials and configuration for every third-party tool connected to the CRM, plus reconnection and testing in HubSpot.

How long this takes

Most go-lives take about 8 to 14 weeks from kickoff, depending on data volume and the number of automations and integrations. The ranges below are estimates, and some steps overlap, so they won't add up line by line. Post-go-live monitoring runs daily for the first few weeks before shifting to weekly, with formal reviews at one month and three months. Set sign-off owners and dates at kickoff so reviews are booked before you need them.

Bars showing about how many weeks each migration stage takes, from about 1 week up to 4 weeks or more, with most go-lives taking about 8 to 14 weeks
StepsDurationWhat happens
Steps 1 and 21 to 3 weeksDocument everything in Salesforce and clean the data. Step 2 runs in parallel with Step 1.
Step 31 to 2 weeksMap every object, pipeline, and automation to its HubSpot equivalent and agree a workaround for every gap.
Steps 4 and 52 to 4 weeksRebuild the architecture in a sandbox and connect third-party tools to it.
Step 61 to 2 weeksImport sample records, test the automations, and sign off the test with your stakeholders.
Step 81 to 2 weeksRole-based training for every team. It overlaps the go-live window so people go live with confidence.
Steps 7 and 9About 1 weekFreeze changes in Salesforce, deploy the validated setup to production, run the full data import and the delta sync, reconnect integrations, sign the go-live checklist, and set Salesforce to read-only.
Step 104 weeks or moreDaily checks on enrollments, data ingestion, and pipeline accuracy, shifting to weekly. Formal reviews at one month and three months.

The 10 steps, Salesforce to HubSpot

Each step opens with what has to be true before you move on to the next one. Two steps end in a signed gate: the test sign-off in Step 6 and the go-live checklist in Step 9. Nothing moves forward until each one is signed.

The 10 migration steps in order, with a signed gate after Step 6 (the test sign-off) and Step 9 (the go-live checklist)

Step 1: Take inventory

When this step is done, every object, pipeline, automation, validation rule, formula, piece of code-based logic, integration, user role, and report is documented, and every automation without a direct HubSpot equivalent is flagged.

Document every object you actively use: contacts, accounts, opportunities, leads, custom objects, and junction objects, and how they relate to each other. Document your full pipeline from marketing through customer success. Document all active automations, validation rules, formulas, and code-based logic, every connected third-party tool, your users, roles, and ownership rules, and the reports and dashboards your teams depend on.

Moving the records is the easy part. Tools can transfer data between systems quickly, but they won't carry over your workflows or the process behind them, so that is what the inventory has to capture.

A doctor runs tests before prescribing anything. The inventory is that test, and it's why nothing gets built in Step 4 until this list exists.

Your outgoing Salesforce admin belongs in the room for this. They know why the current system was built the way it was, and structure should never be replicated without understanding its purpose.

The Rat's Nest is the pile of old Salesforce automations the inventory turns up. Each one either gets mapped to a HubSpot equivalent in Step 3 or gets flagged as a gap with an agreed workaround.

Say one automation doesn't make the inventory list. Nobody builds its replacement, so after go-live a deal stage stops updating and nobody knows why. Caught here in inventory, that same automation gets a documented workaround and works from day one.

Step 2: Audit and clean the data

When this step is done, duplicates are resolved and the record list is finalized before anything is exported.

Resolve duplicates, correct errors, and decide which records come over before you export. Clean the data before it moves. Dirty data that lands in HubSpot has to be cleaned inside HubSpot anyway, and carrying records you don't need has a direct cost.

Deduplication tools only match on what's filled in. If the fields a tool matches on are empty, it can't find the duplicates, so fill or remove those records first.

Contacts designated as marketing contacts affect your HubSpot licensing cost, so decide deliberately which contacts you mark as marketing. Super Admins and users with Modify Billing permission can set a maximum number of marketing contacts under Account & Billing before the import runs. Imported contacts are non-marketing by default. The Set these contacts as marketing contacts checkbox in the Details step of the import overrides that default, and it works only for a single-file contacts import, not for a multi-object import. See HubSpot's guide to setting contacts as marketing contacts.

This step runs in parallel with Step 1.

Step 3: Plan the HubSpot architecture

When this step is done, every object, pipeline, and automation is mapped to its HubSpot equivalent, every gap is flagged with an agreed workaround, and your hubs and licenses are confirmed before anyone builds.

Working from the inventory, determine how each object, pipeline, automation, and association (the link between two records, like a contact and the company they work for) will be replicated in HubSpot. For automations without a direct equivalent, identify the closest alternative and write down the workaround you've agreed. Any automation that needs custom code in a workflow requires Data Hub (formerly Operations Hub) Professional or Enterprise, so price that requirement here. If a workaround needs custom code, build it on Node.js 24, because HubSpot no longer lets you create new Node.js 20 actions. See HubSpot's guide to upgrading custom code actions to Node.js 24.

Most of the mapping starts from this short list of equivalents.

In SalesforceIn HubSpot
AccountsCompanies
OpportunitiesDeals
Leads and contactsContacts
Profiles and rolesPermission sets and teams
AutomationsWorkflows, with the closest alternative and an agreed workaround for any gap

Objects are only half of the mapping. Map each Salesforce field you're keeping to a HubSpot property, decide what happens to custom fields and picklist values that don't have an equivalent, and define how record ownership carries over.

Book working sessions with each department in week one. On one migration, we worked from the RevOps team's view of the marketing process for the first two months and didn't bring marketing into the room until after that. The plan needed changes, and fixing it added two to three months to a project we'd scoped at about three.

Step 4: Build in sandbox

When this step is done, the full architecture is validated in a sandbox and every gap and workaround is documented.

A sandbox is a separate HubSpot account for building and testing, so nothing you do touches your live data. On Enterprise portals, recreate the full architecture there, including custom objects, pipelines, properties, workflows, sequences, segmentation, subscription types, email templates, forms, landing pages, dashboards, and reports. When it's validated, Step 7 deploys it to your live (production) portal. On Starter and Professional portals there is no sandbox, so this step happens in your live portal instead, as the table in the prerequisites shows.

HubSpot doesn't deploy everything from a sandbox. Here is what carries over and what doesn't, based on HubSpot's list of assets eligible to be deployed to production:

What you builtWhat happens at deploy
New supported assetsDeploy
Edits to assets copied from productionDo not deploy
WorkflowsDeploy set to Off. A workflow that depends on an asset type that can't deploy won't deploy either.
Custom property access, sensitive data settings, segment settings, and workflow organizational settingsRedo by hand in production
New forms and new automated marketing emailsDeploy, but their settings don't copy
Regular marketing emailsNot deployable, so rebuild them in production
IntegrationsIntegrations in production won't connect automatically to the sandbox
Contact, company, deal, ticket, and custom object IDsDiffer between sandbox and production

Write down every gap and every agreed workaround as you hit them. Also keep a second list of anything that won't deploy, such as regular marketing emails and the settings above, because Step 7 rebuilds those by hand.

Step 5: Reconnect integrations

When this step is done, every third-party tool is connected to the sandbox or a test configuration, field mappings are verified, and nothing is writing bad data back into the new system.

Connect each tool to the sandbox or a test configuration and verify it: data flows in the right direction, field mappings are correct, and nothing writes bad data back into HubSpot. Integrations in production won't connect automatically to the sandbox, as noted in Step 4, so plan this as its own workstream. Your IT and integrations owner holds the credentials and runs the tests.

If you keep HubSpot and Salesforce connected while you test, our HubSpot and Salesforce integration guide explains how the sync works.

Step 6: Import sample data and sign off the test

When this step is done, representative records are imported, the automations are tested, and the test sign-off is documented and signed by your stakeholders before you move to Step 7.

Export a representative set of records from Salesforce and import them into the sandbox. Then test against that sample: records behave as expected through the pipeline, enrollments (records entering a workflow) trigger correctly, associations are intact, ownership is preserved, and validation rules and formulas behave as designed. Test permissions and pipelines the same way.

Email automations are the exception. Emails to contacts who aren't sandbox users won't send, so email automation behavior is verified in production at Step 9 instead. Three other checks move to Step 9 because they need production data to mean anything: full record counts, the report comparison against Salesforce, and the production integration smoke tests.

This sign-off is the gate. Don't begin Step 7 until it's documented and your stakeholders have signed it.

Step 7: Deploy the setup to production

When this step is done, the validated sandbox setup is live in your production portal and your integrations point at production.

Deploy the validated sandbox setup to production. This moves your configuration. Your records come over in Step 9. Rebuild by hand anything the deployment doesn't support, working from the list of assets that won't deploy that you kept in Step 4. Then switch your integrations to the production environment and reconnect them. Workflows arrive switched Off. When you switch one on, HubSpot asks whether to enroll records that already meet its trigger or only records that meet it from then on, so decide that for each workflow first. See HubSpot's guide to creating workflows.

Step 8: Train and enable the team

When this step is done, every team has had role-based training and your incoming HubSpot admin is ready to own the platform.

The HubSpot interface, its objects, and the daily workflows are different from Salesforce, so plan training by role and build each session around the processes that team actually runs. Training starts before go-live and overlaps the cutover week, so nobody's first day working in HubSpot is also their first look at it. Your incoming HubSpot admin is enabled here to own the platform going forward.

Step 9: Go live and sign the checklist

When this step is done, the full data import is complete, the go-live checklist is signed, and Salesforce is set to read-only.

This is the step where your team starts working in HubSpot for real. Schedule a change freeze in Salesforce in advance, with no new automations and no structural changes. The full data import runs into your production portal. After it finishes, a delta sync, a second and smaller transfer of anything created or changed in Salesforce since the test import, catches the remainder. Salesforce stays intact and untouched until validation passes. Agree your go/no-go criteria (the specific checks that decide whether you go ahead) before the window opens. If validation fails, Salesforce is still intact, so you are never left without a working CRM.

Then you work the checklist with your stakeholders, against concrete criteria.

  1. Record counts by object match between source and destination, accounting for records you intentionally excluded during cleanup.
  2. Associations are intact on a random sample: contacts linked to the right companies, deals to the right contacts, activities to the right records.
  3. Every migrated automation has been triggered at least once in testing and behaved as designed.
  4. Every flagged gap has a documented, agreed workaround.
  5. Email automation behavior is verified.
  6. Pipelines and stages match your real sales process, confirmed by the sales stakeholders who use them daily.
  7. Key reports produce numbers consistent with Salesforce for the same period, including MQLs, SQLs, deals created, and revenue by stage.
  8. Record ownership is preserved.
  9. User permissions match the agreed access design.
  10. Subscription and consent statuses transferred accurately.
  11. Email sending is authenticated on your domain.
  12. Each reconnected integration has passed a smoke test in production, a quick end-to-end check that data flows and nothing errors.
  13. Stakeholders from marketing, sales, and customer success have each signed off on their area.

Email sending is its own piece of work. When you move it to HubSpot, plan for domain authentication and a gradual sending ramp-up so deliverability and sender reputation are protected.

Once the checklist is signed, Salesforce goes read-only for an agreed archive period. Keep read access for that whole period. Don't terminate the instance immediately.

Step 10: Monitor after go-live

When this step is done, the monitoring cadence is running and a formal review is on the calendar.

In the first weeks, run daily checks on automation enrollments to catch records falling through the cracks, monitor data ingestion quality, and watch pipeline accuracy. Compare key metrics against your Salesforce baseline, including MQLs, SQLs, deals created, and contacts added, so an unexpected drop shows up as a number. As the team settles in, monitoring shifts from daily to weekly.

Book formal reviews at the one-month and three-month marks. Each one asks whether automations need to be added or adjusted, whether the data governance rules are working, and whether reporting is accurate and genuinely used.

Plan follow-up enablement sessions once your team has been on the platform a few weeks. Real usage surfaces questions that pre-launch training can't anticipate.

One signal is worth watching closely: teams working around the system. A workaround is a symptom of a gap that still needs addressing.

Frequently asked questions

How long does a Salesforce to HubSpot migration take?

Most go-lives take about 8 to 14 weeks from kickoff, depending on data volume and the number of automations and integrations. Monitoring continues after go-live, daily for the first few weeks and then weekly, with formal reviews at one month and three months.

Can we migrate on HubSpot Starter or Professional?

Starter and Professional portals don't include a sandbox, which Steps 4, 6, and 7 rely on. RevBlack builds and validates directly in the production portal under a controlled process.

What happens to our email history?

Logged emails, calls, meetings, notes, and attachments can all be migrated, but volume drives the effort and sometimes the cost. We usually bring over full activity history for open deals and recent records and keep a lighter archive for older, inactive ones. We agree before the migration starts what comes over and what stays in an export archive.

What doesn't HubSpot deploy from a sandbox?

Regular marketing emails can't be deployed, and edits to assets copied from production don't deploy. Settings such as custom property access, sensitive data settings, segment settings, and workflow organizational settings have to be redone by hand in production. Workflows deploy set to Off. Step 4 has the full list.

Who needs to be involved?

You need a migration lead, an incoming HubSpot admin, your outgoing Salesforce admin, a stakeholder from each of marketing, sales, and customer success, and whoever owns your integrations and their credentials. If you're the RevOps lead running the project, you're the migration lead.

Do we have to freeze changes in Salesforce?

Yes, for the go-live window. Schedule the freeze in advance, with no new automations and no structural changes, so the delta sync can catch anything created or changed since the test import. Salesforce stays intact until validation passes.

RevBlack has completed more than 300 CRM audits, and the inventory is where this migration is won or lost. It's a step that's hard to do well from inside the system you're leaving. If you want the inventory, the architecture map, and the go-live checklist run by someone who has mapped Salesforce automations onto HubSpot equivalents before, you start in HubSpot with a signed list of what moved and who approved it.

BOOK A MIGRATION CALL
Pedro Wolff
Written by
Pedro Wolff
RevOps Consultant

LinkedIn
Guides

Don't miss these

Get started with revblack today

Ready to see these results for your business?

Fill out form