A HubSpot to Salesforce migration is a serious project. Done well, you ship a more powerful CRM with cleaner data than you started with. Done poorly, you spend six months in a half-broken state where both systems disagree on the truth, deals slip, and pipeline visibility evaporates.

This post is the 6-phase playbook we use with every client. It assumes you've already decided Salesforce is the right destination (we covered that decision in our HubSpot vs Salesforce comparison). Here's how to actually move.

Why HubSpot → Salesforce Is the Harder Direction

Both vendors will tell you migration is easy. It is not.

Salesforce has a more rigorous data model than HubSpot. Where HubSpot lets you store almost anything as a contact property, Salesforce wants you to think in objects, relationships, and fields with defined types. The mismatch shows up in every column of your export. Phone numbers come over fine. Custom multi-value fields, calculated properties, and HubSpot-specific concepts (lifecycle stages, lists) don't have one-to-one Salesforce equivalents.

The automation rebuild is the other major lift. Workflows you've spent years tuning in HubSpot have no direct migration path. Every routing flow, every nurture, every internal notification has to be redesigned and rebuilt in Salesforce Flow or Apex.

Plan for a 12-16 week project for a mid-size B2B company. Smaller orgs (one product, ~50 users) can compress to 8-10 weeks. Larger orgs (multiple business units, regulated industries) routinely take 6+ months.

Phase 1: Audit and Decision Architecture (Weeks 1-2)

Before you touch any data, you need to make decisions about how the target Salesforce org will be structured. The audit phase is about getting clear on what you have today, and what you want.

What to inventory:

HubSpot objects in use (Contacts, Companies, Deals, Tickets, Custom Objects).
All custom properties on each object, with usage counts (how many records actually have a value).
All active workflows, with their triggers and actions.
All active lists (smart and static).
All active forms.
All active sequences.
All integrations.
All reports and dashboards in heavy use.

Decisions to make:

Object model. Will you use standard Salesforce objects (Lead, Account, Contact, Opportunity) or also custom objects? Most B2B teams need at least 1-3 custom objects.
Lifecycle stage equivalent. HubSpot's lifecycle stage doesn't have an exact Salesforce match. Most teams build a custom picklist field on Contact + Account that mimics the stages.
Lead vs Contact. Salesforce makes a distinction HubSpot doesn't. Decide your lead-to-contact conversion rules early.
Property cleanup. Use this as the forcing function to delete unused properties. Average B2B HubSpot instance has 200+ properties on Contact. The destination Salesforce org probably needs 50-80.
Custom objects. What HubSpot custom objects (or repurposed standard objects) need to become Salesforce custom objects?

Deliverable for Phase 1: a written architecture decision document. Get sign-off from RevOps, Sales leadership, and Marketing before proceeding.

Phase 2: Salesforce Org Build (Weeks 3-5)

Build the destination Salesforce org from scratch in a sandbox. This is not data migration — this is configuration.

Create custom objects, custom fields, and record types.
Configure profiles, permission sets, and role hierarchy.
Build page layouts, dynamic forms (Lightning), and search layouts.
Configure opportunity sales process (stages, probabilities).
Set up products, pricebooks, and quote templates if relevant.
Build validation rules.
Configure email-to-case, web-to-lead, web-to-case if needed.

Critical principle: don't build automation yet. Get the data model and UX right first. Automation comes after data lands.

Deliverable: a fully configured Salesforce sandbox org that mirrors your target production state, minus data and automation.

Phase 3: Data Cleansing in HubSpot (Weeks 4-6, parallel)

This phase runs parallel with the Salesforce build. While the destination is getting built, clean the source.

The cleansing operations:

Deduplicate contacts. HubSpot's native dedupe + a tool like Insycle. Aim for under 5% duplicate rate.
Standardize key fields. Phone numbers, country names, state abbreviations, industry values. Pick a normalization scheme and apply it.
Backfill required fields. If Salesforce will require Industry on every Account, make sure HubSpot has it before export. Enrichment via Apollo, Clearbit, or ZoomInfo is faster than manual.
Archive ancient records. Contacts that haven't engaged in 3+ years probably don't belong in the migrated dataset. Export them to a cold storage spreadsheet and exclude from migration.
Resolve ownership. Records owned by former employees need new owners assigned. This is a forcing function for an ownership audit you probably should have done years ago.

Deliverable: a clean HubSpot instance ready to export, with a documented exclusion list (what's being archived vs migrated).

Phase 4: Test Migration in Sandbox (Weeks 6-8)

Migrate a subset (typically 10-20% of records, but a representative sample) to the Salesforce sandbox. Don't migrate everything yet.

The test migration validates:

Field mapping works correctly (data lands where you expect).
Calculated fields compute as expected.
Ownership assignments transfer correctly.
Associated records (Contacts on Accounts, Deals on Contacts) maintain relationships.
Records that fail to import are flagged with reasons.

Tools that handle the actual transfer: Salesforce's Data Loader is the workhorse. For more sophisticated mapping, tools like Trujay or HubSpot's own migration toolkit handle some of the lift. We typically build a custom n8n workflow for complex mappings so we can apply transformations on the way through.

Run UAT (User Acceptance Testing) with a sample of users from each team — sales, marketing, ops. Have them find specific records, run specific reports, and confirm the data looks right. Fix issues. Iterate.

Deliverable: a Salesforce sandbox with a clean test data set and signed-off UAT results.

Phase 5: Automation Rebuild (Weeks 7-11, parallel)

Rebuild all critical HubSpot automation in Salesforce. This is the largest lift of the migration.

The automation inventory from Phase 1 becomes the build backlog. Triage it:

Must-rebuild. Lead routing, lifecycle stage management, deal automation, internal notifications, customer onboarding triggers. These are non-negotiable for go-live.
Should-rebuild. Nurture sequences, scoring updates, reporting refreshes. Get these in within 30 days post-launch.
Won't-rebuild. The automations that nobody actually relies on. Use migration as the forcing function to kill them.

Build the must-rebuild automation in Salesforce Flow, with Apex for the cases where Flow can't do the job (covered in our Flow vs Apex post).

For email nurtures, decide where they'll live. Salesforce Marketing Cloud, Pardot/MCAE, or a third-party tool like Customer.io. Each has different setup costs.

Deliverable: a Salesforce sandbox running the must-rebuild automation, tested against the migrated data.

Phase 6: Cutover and Hypercare (Weeks 12-16)

The cutover weekend is the most stressful 72 hours of the project. Done right, it's anticlimactic.

Pre-cutover (Week 12, Thursday-Friday):

Freeze HubSpot changes (announce it 1 week in advance).
Final delta export from HubSpot.
Final import into Salesforce production.
Validate record counts match.
Switch DNS / form actions / integration endpoints to point to Salesforce.

Cutover (Week 12, Saturday-Sunday):

Activate Salesforce automation in production.
Run sanity-check reports.
Verify a few critical workflows fire correctly (test lead routing with a fake lead, test Closed Won automation with a sandbox deal cloned to prod).

Hypercare (Weeks 13-16):

Daily standup with the project team for the first week.
Twice-daily Slack standup with users for the first week.
Open ticket queue for issues, triaged hourly during business hours.
Resolve issues in priority order. Don't promise to rebuild "Should-rebuild" automation until critical issues are stable.

Deliverable: a stable Salesforce production org with users actively working in it and a closing ticket queue.

The Five Mistakes That Derail Migrations

1. Trying to migrate without a clean source. Migrating bad data into Salesforce produces an even worse Salesforce org. Do the cleansing first.

2. Skipping UAT. "We tested it ourselves" is not UAT. Real users finding real bugs in real workflows is UAT.

3. Underestimating automation rebuild. This is 40-50% of the project effort and most teams underestimate it by 2x.

4. No training plan. Reps who don't know how to use the new system stop using the new system. Build a training plan, schedule sessions, and have a designated champion on each team for the first 60 days.

5. Dual-system limbo. Running both HubSpot and Salesforce as "the source of truth" for any meaningful period (more than 2 weeks) creates data drift that's expensive to reconcile. Pick a cutover date and commit.

Is a HubSpot to Salesforce Migration Right for Your Team?

It's the right move when you've outgrown HubSpot's structural model — specifically: complex sales motion, multi-product, multi-segment, custom objects in heavy use, regulatory requirements that need Salesforce's audit story, or integrations that have first-class Salesforce support and weak HubSpot support.

It's the wrong move when you're frustrated with your HubSpot implementation but the platform isn't actually the constraint. We've seen teams migrate to Salesforce and discover their problems were process problems, not tool problems. Salesforce just made them louder.

At Ops Automators, we run CRM migrations for B2B companies — both directions, including this one. If you're contemplating the move, the audit phase alone is worth a conversation.

Ready to automate? Book a free discovery call and we'll scope your migration honestly.

HubSpot to Salesforce Migration: The 6-Phase Playbook We Use With Every Client