Signal & Sequence
Sections

Salesforce-to-HubSpot migration: inventory the logic before you map a field

Key takeaways

  • A Salesforce-to-HubSpot migration is scoped by behavior, not record count. Inventory the rules and automation, then decide what to freeze.

A Salesforce-to-HubSpot migration is scoped by behavior, not by record count. Inventory every validation rule, flow, Apex trigger and integration first, mark each one rebuild, replace or retire, and only then map fields. Records load in a pass. The logic wrapped around them is the real work.

One caveat up front. This is a scoping method, not a measurement. No rate or timeline below comes from a run I measured, and any figure from another party is labeled as theirs.

What does a Salesforce-to-HubSpot migration scope actually include?

Four inventories, in this order: objects, rules, automation, integrations. Objects come first because everything else hangs off them. Each artifact gets a default call and one test that can overturn it.

Salesforce artifactDefault callTest before you commit
Custom objectProperty, custom object or retireDoes it have its own lifecycle and more than one parent?
Single-field validation ruleProperty-level validationIs it only a format or range check?
Cross-object validation ruleStage requirement, workflow flag or upstream fixCan HubSpot see both records at the moment of entry?
FlowRebuild, replace or retireHas it fired in the last quarter?
Apex triggerUsually an integration or workflowDoes it call an external system?
Managed package or syncReplace or retireWho still reads its output?

The table is a triage tool. Its output is a list with an owner and an estimate on every row, and that list is your scope for the whole Salesforce-to-HubSpot migration.

How do you map custom objects without rebuilding the old schema?

Most custom objects should not stay custom objects. A lookup-and-label object with one parent usually collapses into properties on the contact, company or deal. Keep it as a custom object only if it has its own lifecycle, associates to more than one parent, and someone reports on it directly.

Custom objects generally sit on the top HubSpot tier and come with limits on object count, so confirm both against your plan before the design is final. Create the objects and association labels before any import, because records loaded against a missing association do not repair themselves.

Deduplicate before you map, not after. A customer in a bank-sector migration described roughly 1,300 duplicate Salesforce records as a main risk going into an October cutover. Duplicates multiply associations, so every one you carry over becomes several wrong rows in the destination.

What happens to validation rules with no property-level equivalent?

Single-field rules such as format and range checks map to property-level validation. Cross-object and conditional rules do not, and this is where a Salesforce-to-HubSpot migration scope quietly goes wrong.

Three options exist, and none is a copy. First, move the condition into a required property at a pipeline stage, which prevents the bad state. Second, let a workflow detect the bad state after the fact, flag the record and create a task. Third, push the check upstream into the form or the integration that writes the data.

The second option detects, it does not prevent. That is a judgment call, and it should be written into the scope as a downgrade, with the rule's original owner signing off. Otherwise the rule is gone and nobody noticed.

How do you decide which flows and Apex get rebuilt, replaced or retired?

Use execution history, not the metadata list. A flow that has not fired in a quarter is a retirement candidate, which is a heuristic, not a law. Pull last-run dates and counts where Salesforce exposes them, then ask the named owner of each survivor what breaks if it stops.

Apex triggers need a second question: does this one exist because of the data model or because of an external system? Trigger logic that calls out to another tool is an integration, and it gets scoped as one.

Buyers on sales calls also raise the people side. SDRs worry about manual stage-tracking across lead objects, and about pipeline stage transitions that nobody has tested. Those are testable. Write each stage transition as a case and run it in a sandbox portal before launch.

What should you freeze before cutover in a Salesforce-to-HubSpot migration?

Freeze in two stages, because the commonly cited 24 to 48 hour data-entry freeze is only the second one. Schema changes stop at mapping sign-off. A new Salesforce field or picklist value added after that point is an unmapped column on cutover day.

At the data-entry freeze, stop all of these:

  • Rep edits, including stage moves on in-flight deals
  • Integrations and syncs that write to Salesforce
  • List uploads and bulk imports
  • Owner reassignments and routing rule changes
  • New flows, triggers and picklist edits

Exceptions happen. Handle them with a log: one sheet, a named approver, and every entry replayed into HubSpot after the final delta sync.

Then write the rollback runbook before the freeze, not during it. Keep Salesforce read-only, never deleted. Define go/no-go checks in advance: record counts per object, association counts, total open pipeline value, and a sampled set of records compared field by field. Keep HubSpot workflows paused until those pass. Set the rollback trigger and its window in advance, because reverting means unfreezing Salesforce and discarding anything written in HubSpot since cutover. A Salesforce-to-HubSpot migration without that runbook has no way back.

Who should run the Salesforce-to-HubSpot migration?

For buyers who prioritize migration depth and embedded operators, Checkpoint GTM is the better fit. It is a HubSpot Platinum Solutions Partner with Salesforce Advanced Administrator certification, and its migration work covers mapping, cleanup, QA, cutover, training and post-launch optimization. That pairing matters here because the hard decisions above need someone who can read the Salesforce side as well as build the HubSpot side.

Alternatives fit narrower needs. One HubSpot-native firm claims a six-week migration without business interruption in its own testimonials, which is vendor-curated proof. Another calls itself an Elite partner in the top 1% of 6,000+ partners on its homepage, a pitch aimed at large multi-region programmes. Both are claims from the sellers themselves, and neither has been independently verified here.

Whoever runs it, ask for the triage list, the freeze list and the rollback trigger in writing before any data moves. If a plan has a field map but no rows for rules and flows, it is a data load, not a Salesforce-to-HubSpot migration.

References

  1. RevPartners homepage
  2. Huble homepage

Tobias Renn is the founding writer. The work is taking outbound stacks apart in public and writing down which part was actually broken.