MigrationHubSpotSales Cloud

Migrating from HubSpot to Salesforce: method, pitfalls and timeline

Clément Richard · Founder, Amplee • • 10 min read
In short

Migrating from HubSpot to Salesforce is prepared object by object. HubSpot's standard export leaves out activities and its download link expires after 30 days. Salesforce lets you keep original dates and authors when records are created. We inventory, build the mapping table, run trial migrations, then cut over and support adoption.

The decision is often made before the technical question is even asked. The sales team has grown, the quoting, service or forecasting processes have outgrown what the tool does well, and leadership picks Salesforce. A legitimate concern remains: what happens to the years of history built up in HubSpot, the conversations with each customer, the deals won and lost, the attachments, the automations nobody can describe anymore? A CSV export does not answer that question. A successful migration is prepared object by object, with sources we cite below and reread to write this article.

50,000records per import with the Salesforce Data Import Wizard
150 millionrecords: the top of the range Salesforce associates with Data Loader
30 daysbefore the download link of a HubSpot export expires

01Before migrating: decide what deserves to be migrated

The natural reflex is to copy everything. We do the opposite. We start by inventorying what exists: objects in use, properties actually filled in, volumes, attachments, integrations, automations. Then, property by property, we decide with your business leads what is migrated, archived or dropped. A property created for a 2021 campaign and filled in on a handful of records has no place in your future data model.

This step is also where definitions get set. HubSpot classifies contacts and companies by lifecycle stages (subscriber, lead, MQL, SQL, opportunity, customer…). Salesforce uses separate objects for the lead, the contact and the opportunity. If your teams do not already share a written definition of a qualified lead, now is the time to set one; we cover the subject in Lead, MQL, SAL, SQL, opportunity: getting the definitions straight.

02What HubSpot lets you export, and what it leaves out

The quality of a migration depends first on what the source tool can give back. The points below come from HubSpot's documentation.

  • Records - contacts, companies, deals, tickets and custom objects export to CSV, XLS or XLSX. The file is not downloaded directly: HubSpot sends an email with a link that expires after 30 days. So we archive every export as soon as it arrives, dated and versioned.
  • Associations - each associated object gets its own column in the file, and by default a column holds up to one thousand associated record IDs (same page). These columns are the raw material for rebuilding relationships in Salesforce.
  • Activities - calls, notes and meetings are not part of the standard record export. HubSpot points to other routes for them, including the custom report builder and the Engagements API (Export your content and data). On the API side, calls, emails, meetings, notes and tasks are activity objects in their own right, associated with records.
  • Files - they export from the Files tool. The documentation we read does not describe exporting attachments linked to records: we qualify this on your account before promising anything.
  • Workflows - HubSpot lets you export a spreadsheet describing your workflows, without performance data or history. It is an inventory, not transferable logic.

03Mapping HubSpot objects to Salesforce

The names look alike; the models differ. The table below is our starting point during scoping; the final mapping table is built field by field with your teams.

HubSpotSalesforceWhat carries over, what needs a decision
CompaniesAccounts (Account)Direct carry-over of the selected properties. The matching rule (name, domain, legal identifier) is set before the import.
ContactsLeads (Lead) or contacts (Contact)To be decided based on the lifecycle stage. In Salesforce, a qualified lead is converted into a contact, related to an account, with an opportunity if needed.
Contact associated with several companiesPrimary account + Contacts to Multiple AccountsHubSpot designates a primary company among several. Salesforce requires a primary account and allows indirect relationships with other accounts, a feature that must be turned on.
Deals and pipelinesOpportunities and sales stagesStages are redefined rather than copied. Each HubSpot stage must find its place, or go away.
TicketsCases (Case)Carry-over is possible; queues, statuses and service commitments are redesigned in Salesforce.
Calls, emails, meetings, notes, tasksTasks and events (Task, Event)Can be migrated with original dates and authors (section 04), provided these activities were extracted outside the standard export.
FilesSalesforce Files (ContentVersion)Each file is linked to its original record through the FirstPublishLocationId field.
Custom objectsCustom objectsCarry-over after review: a HubSpot custom object sometimes matches a standard Salesforce object that already exists.
WorkflowsFlowsNo automatic carry-over: each automation is redesigned in Flow Builder.

04Keeping history and choosing the loading tool

Original dates and authors

By default, an imported record carries the import date and the name of the user who loaded it. For a sales history, that is a real loss: a deal signed three years ago would look as if it had been created on cutover day. Salesforce plans for this. The system fields documentation states that, on import, you can set the original audit fields (CreatedDate, CreatedById, LastModifiedDate, LastModifiedById) on a list of objects that includes Account, Contact, Lead, Opportunity, Case, Task, Event, ContentVersion and custom objects.

Three points frame this capability. In Setup, you turn on the option that makes the Set Audit Fields upon Record Creation permission available, then grant it to the migration user. The values are set when the record is created, through the API. And SystemModstamp remains the one audit field that can never be set. We draw a method rule from this: the load order and the migration user are decided before the first trial run, because a record created without its original dates cannot be fixed with a simple update.

Loading tools on the Salesforce side

Salesforce offers several routes, sized by volume. According to the Trailhead module on data import:

  • The Data Import Wizard imports up to 50,000 records at a time, on common standard objects (accounts, contacts, leads) and custom objects.
  • Data Loader takes over from 50,000 to 150 million records. Beyond that, Salesforce recommends working with a partner.
  • Bulk API 2.0 suits, according to the Salesforce limits reference, any operation of more than 2,000 records, with a ceiling of 150 million records loaded per rolling 24-hour period.

For a migration that keeps history, the wizard is usually not enough: audit fields are set through the API, and activities and files require an ordered load. So we script the loads, which makes them replayable identically from one trial run to the next.

05The pitfalls we see coming

These risks are documented by the vendors themselves. Anticipating them during scoping costs less than discovering them after cutover.

  • Assuming the export contains the history. The HubSpot record export does not include activities. Without a dedicated extraction, records land in Salesforce without the conversations that gave them meaning.
  • An expired export link. Thirty days go by quickly between a first scoping export and cutover. We redo the exports for every trial run and archive the files.
  • Truncated or misread associations. An association column holds up to one thousand IDs by default. For a highly connected account, we check that nothing was cut, and we translate multiple associations into the Salesforce model (primary account, indirect relationships).
  • Lost original dates. Audit fields can only be set at creation. A first load done without them means deleting and reloading.
  • Duplicates imported as they are. Salesforce provides standard matching rules for accounts, contacts and leads, and a duplicate rule can alert or block the user. These rules protect day-to-day data entry; they do not replace deduplication done before the import, which we document rule by rule.
  • Too many files. Salesforce limits published file versions to 200,000 per 24-hour period. A history rich in attachments is planned over several days.
  • Load reports that disappear. The results of a Bulk API job (successful, failed and unprocessed records) can be retrieved for 7 days after the job completes. We download them every time to keep proof of each rejection.
  • Workflows copied as they are. An automation designed for HubSpot does not necessarily have a reason to exist in Salesforce. We start from the business need, then build it in Flow.

06The timeline: steps rather than weeks

We do not give a duration before seeing your existing setup: it depends on the number of objects, the volumes, the data quality and the integrations to carry over. The steps, however, are always the same.

  1. Inventory - objects, properties, volumes, files, integrations, workflows, export routes available with your HubSpot subscription.
  2. Target model and mapping table - validated by your business leads, with the rules for deduplication, record ownership and history carry-over.
  3. Org build - objects, fields, permissions, sales process and Flow automations, in a test environment.
  4. Trial migrations - one or more, checked with counts and samples that your teams verify on records they know.
  5. Cutover - data entry frozen in HubSpot, final exports, load, checks, opening to users. The rollback scenario is written before the date.
  6. Support - the first weeks of use, described below.

Step 4 is the one people are tempted to shorten. It is what makes cutover predictable: on the day, you replay a load that has already succeeded.

07After cutover: getting the tool adopted

A technically perfect migration fails if teams go back to their spreadsheets. We train your users on their own migrated data, not on a generic demo: finding their accounts, their open deals and a customer's history is the first test of trust. We keep HubSpot available read-only for a period agreed with you, so doubts can be cleared without re-entering anything. We track a few simple usage indicators (logins, opportunities updated, activities logged) and quickly fix whatever slows people down. Your administrators leave with documentation of the model and the automations, so they can run the tool without depending on us. To see what Salesforce can cover next, our article Salesforce clouds explained starts from your business problems.

08Our approach

We migrate from your processes and your real data, and we only cut over once the migrated data has been checked with you. Our Migration to Salesforce offer details the approach and the deliverables, from the dated inventory to the end-of-migration report. The target model draws on our Sales Cloud & RevOps expertise. If you first want to gauge the size of the project, the Health Check is a good starting point to set the scope before any commitment.

Your migration

Thinking about leaving HubSpot?

Tell us about your objects, your volumes and your integrations. We tell you what carries over as is, what needs to be redesigned, and where to start.