Migrate from Salesforce to HubSpot CRM without losing what matters
Most Salesforce to HubSpot migrations fail quietly. The record counts match, the deal looks fine, and then a rep reopens an old account and the email thread, the owner, and the close date are all wrong. This guide walks through how to move accounts, contacts, deals, and custom fields cleanly, in the right order, with history intact.
The mapping you need to get right first
The core objects line up, but they are not identical. Salesforce Accounts become HubSpot Companies. Salesforce Contacts and Leads both become HubSpot Contacts, because HubSpot's own Lead object behaves differently (it must attach to an existing contact, so most teams map Salesforce Leads to contacts with a lifecycle stage instead). Salesforce Opportunities become HubSpot Deals. Custom objects either become HubSpot custom objects, which require an Enterprise tier, or get remodeled into existing object types if there is no clean equivalent. Before you move a single record, export the field list for each object and build a mapping spreadsheet: each Salesforce field, its HubSpot property, the field type, and whether the property exists yet or needs to be created. That spreadsheet is the contract for the whole project. Skip it and you will rediscover every mismatched picklist value the hard way, in production.
Clean the data before it moves, not after
Migrating dirty data just gives you a faster way to find the same problems in a new tool. Spend a week or two up front on the parts that actually drive revenue and reporting: contacts, companies, deals, and activities. Deduplicate, standardize formats, and prune records nobody has touched in years. Pay special attention to email, because HubSpot uses it as the primary key for contact deduplication and rejects anything that is not in [email protected] format. A quick regex pass over your export catches most of the bad addresses before HubSpot ever sees them. One sharp edge worth knowing: Salesforce lets two records share an email address, and HubSpot does not, so decide how you will collapse those before import rather than letting the importer pick a winner for you.
Build HubSpot to receive the data, in this order
Set up the destination before you load anything. First create your users and teams, because record ownership depends on the user already existing. Then create every custom property from your mapping sheet, confirming field types up front (HubSpot will not let you later flip a property to or from Score, Calculated, or Date without consequences, and changing a field type can invalidate stored values). Next, build your deal pipelines and stages. This step is non-negotiable sequencing: stages must exist before deals load, or the deals land in the wrong place. Treat this as a redesign, not a copy. If you replicate your Salesforce stages verbatim you have migrated your old problems too. Name stages after buyer behavior, split into multiple pipelines if teams or products differ, and push one sample deal through every stage before you load thousands.
Import in dependency order, and keep associations in one file
Load data in the order things depend on each other: companies, then contacts, then deals, then activities, then custom objects. The reason is mechanical. You cannot associate a contact with a company that does not exist yet. The trap most teams hit is associations across three objects. If you import companies, contacts, and deals from separate files, HubSpot can only associate two objects at a time and you end up with deals that are not tied to both their company and their contacts. The fix is to put the related records on the same row in a single import file with a unique identifier for each object, and let HubSpot create all the associations at once. Use the Company Domain Name to tie contacts to companies, and Record ID once you have them after the first imports. Note that domain-based dedupe is skipped for companies created through the API, so plan accordingly if you go that route.
Preserve owners, history, and a legacy ID
Three things get lost in rushed migrations: owners, history, and traceability. For owners, map Salesforce OwnerId to the HubSpot owner and confirm every record has one, because unassigned companies, contacts, and deals are a mess to untangle later. There is also a hard timing detail at cutover: when you disconnect the Salesforce sync, HubSpot flags records owned by Salesforce-only or deactivated users, and you typically have 24 to 48 hours to reassign them before ownership clears. For history, remember Salesforce stores activities with polymorphic relationships and shared activities, so a single task can touch several people. Preserve every participant, not just the first contact, or your pre-migration pipeline attribution disappears. Finally, add a Legacy Salesforce ID property to contacts, companies, and deals. It costs nothing and it is the difference between a clean post-migration audit and a guessing game.
The two workstreams that are not part of a normal import
Two things almost never come along for the ride, and assuming they do is a classic mistake. Attachments and files use a separate data model in both systems. CSV imports cannot carry binary data, and HubSpot has no native bulk import for attachments, so file migration is its own workstream built on the API, often adding a week to the timeline. Import the CRM records first, get the HubSpot record IDs, then attach files only for records that still matter to active business. The second workstream is automation. Salesforce Process Builder flows, validation rules, assignment rules, and Apex do not transfer. The native HubSpot to Salesforce connector is a sync tool for keeping two live systems aligned, not a migration engine. Document each automation by what it does for the business, then rebuild it in HubSpot Workflows by intent. And pause active Salesforce automation before you export, or it will fire mid-migration and corrupt your data.
Validate against business questions, not record counts
Matching record counts prove nothing. A migration can reconcile perfectly on numbers and still be broken everywhere that matters. Run a full count reconciliation as a baseline, then go further: pull your key reports in both systems and compare the numbers, test every rebuilt workflow trigger, and have a few reps reopen old accounts and deals to confirm the timeline, owner, close date, and associations look right. Treat any unexplained discrepancy as a blocker to go-live rather than a cleanup task for next quarter. Keep both systems running in parallel during validation. HubSpot reports that focused teams usually complete the move within about three months running a phased rollout, and a phased approach (test batch, validate, historical load, active data, cutover) is what keeps the work honest.
How Pineido fits
We run CRM migrations as a fixed-scope 30-day sprint at a flat $8,500, which suits this work well because a good migration is a finite, well-defined project, not an open-ended retainer. The first week is the mapping spreadsheet, the property and pipeline build, and a test batch. The middle weeks are the phased loads in dependency order with reconciliation after each one. The last stretch is rebuilding the automation that earns its place, handling attachments as their own workstream, and the business-question QA that catches the problems counts cannot. If your move is mostly clean core objects, you may not need help at all, and this guide is meant to get you there. If you have heavy custom objects, shared activities, attachment history, or a wall of Apex, that is exactly where a sprint pays for itself.
Get operations insights weekly
Join service-business operators who get one actionable insight on ops, automation, and growth every week.
No spam. Unsubscribe anytime.
You’re subscribed! Look for our next issue in your inbox.
That did not go through. Please try again, or email [email protected].
Ready to ship?
Tell us where you are stuck. We will review it and send a focused recommendation within one business day.
Frequently asked questions
How do Salesforce objects map to HubSpot objects?
Accounts become Companies, Opportunities become Deals, and both Leads and Contacts become HubSpot Contacts. HubSpot's own Lead object works differently because it must attach to an existing contact, so most migrations map Salesforce Leads to contacts with a lifecycle stage rather than to the HubSpot Lead object. Custom objects either become HubSpot custom objects, which need an Enterprise tier, or get remodeled into existing object types.
Can I just use the native HubSpot to Salesforce connector to migrate?
No. The native connector is a sync tool designed to keep two live systems aligned, not a one-time migration engine. It handles core objects but struggles with custom objects, full association preservation, and historical activity at scale. For anything beyond a few hundred thousand clean records, plan on CSV imports plus API-based work for custom objects, associations, and history.
Why do my deal associations break during import?
HubSpot can only associate two objects at a time when you import from separate files, so deals end up tied to a company or to contacts but not both. The fix is a single import file where companies, contacts, and deals sit on the same row, each with a unique identifier. HubSpot then creates all the associations at once. Import in dependency order: companies, then contacts, then deals.
Do attachments and files migrate automatically?
No. Files use a separate data model in both systems, CSV imports cannot carry binary data, and HubSpot has no native bulk import for attachments. File migration is its own workstream built on the API, and it often adds about a week to the timeline. Import the CRM records first, get the HubSpot record IDs, then attach files only for records still tied to active business.
What happens to my Salesforce automation and Apex?
It does not transfer. Process Builder flows, validation rules, assignment rules, and Apex were built around Salesforce's object model and will not run in HubSpot. Document each automation by the business outcome it produces, then rebuild it in HubSpot Workflows by intent. Pause active Salesforce automation before you export, or it can fire mid-migration and create duplicate or corrupted records.
How long does a Salesforce to HubSpot migration take?
HubSpot reports that focused teams usually complete the move within about three months using a phased rollout with both systems running in parallel. A clean core-object migration moves faster. At Pineido we scope CRM migrations as a fixed 30-day sprint at a flat $8,500, which fits because a good migration is a finite, well-defined project rather than an open-ended engagement.