How do I migrate my existing customer data into HubSpot CRM without losing deal history in 2027?

The outcome you should expect

A successful migration ends with a HubSpot portal where every open deal carries its true stage, amount, expected close date, and owner, and where each deal is associated to the right company and contacts. That sounds obvious, but the failure mode is subtle: teams import deal rows successfully and then discover the deals are orphaned — no company association, no contact roles, no activity timeline. An orphaned deal is nearly worthless for forecasting because HubSpot's pipeline reports, weighted amounts, and stage-aging analytics all depend on associations and timestamps being intact.

So the realistic outcome has two layers. The first is record parity: the same number of companies, contacts, and open deals you had in your source system, minus duplicates you deliberately merged. The second, harder layer is historical fidelity — closed-won and closed-lost deals from the trailing 12 to 24 months, with their original create dates and close dates, so your win-rate and sales-cycle-length reports in HubSpot match what you saw before. Most teams hit record parity in a week and spend two to four more weeks getting history right.

Expect some loss you cannot prevent. Notes stored as free text inside a legacy CRM's activity body, email threads that lived in a rep's personal inbox, and attachments hosted on a decommissioned server often cannot be migrated cleanly. Decide up front which of those you will archive as a read-only export rather than force into HubSpot. A pragmatic target for most mid-market teams: 100% of open deals with full associations, 95%+ of closed deals from the last two years, and a documented archive for everything older. That is a defensible outcome, and it keeps the project from stalling while you chase the last 5%.

What drives that outcome

The single biggest driver of a clean migration is import order and association mapping. HubSpot associates records by matching a unique identifier you supply in the import file — typically email for contacts and company domain for companies. If you import deals before companies exist, or if your company domain column is inconsistent, HubSpot creates new blank companies or leaves deals unassociated. Import companies first, confirm the domain column is normalized, then contacts, then deals, then activities.

The second driver is preserving original identifiers. Create a custom property on each object — something like "Legacy CRM ID" — and populate it during import. This does two things: it lets you re-run a corrected import without creating duplicates, and it gives you a permanent crosswalk between the old system and HubSpot for auditing. Without it, a failed import that partially succeeded becomes a manual cleanup nightmare.

The third driver is timestamps. HubSpot lets you set create date and close date on import for deals, and activity date for engagements. If you skip this, every migrated deal looks like it was created on migration day, which destroys sales-cycle reporting and makes stage-aging dashboards meaningless. Always map the original dates. The fourth driver is owner mapping: build a user in HubSpot for every active rep and map the source owner ID to the HubSpot owner ID before importing deals, or every deal lands unassigned and your territory reports break.

A fifth driver, often overlooked, is the software you use to transform the data. Spreadsheets work for a few thousand rows; beyond that, a scripting environment or an ETL tool handles deduplication, date normalization, and association building far more reliably. Choose based on volume, not ambition — a 40,000-record migration in a spreadsheet will corrupt silently.

Benchmarks and realistic ranges

Timelines vary widely with volume and cleanliness, so treat these as planning ranges rather than promises. A small team migrating under 5,000 total records with a clean source typically completes a first pass in three to five business days and reaches verified go-live in two weeks. A mid-market migration of 20,000 to 100,000 records usually takes four to eight weeks end to end, with the import itself consuming only a fraction of that — most of the time goes to deduplication, owner mapping, and reconciliation.

Duplicate rates are the number that most often surprises teams. It is common for a source CRM to contain 5% to 15% duplicate contacts and 2% to 8% duplicate companies, accumulated over years of manual entry and list imports. Budget time to merge these before importing, because merging after import in HubSpot is slower and risks breaking associations. On the deal side, expect 1% to 5% of deals to have missing or invalid associations in the source — deals with no company, or contacts that no longer exist. Flag those for manual review rather than letting the import create placeholder records.

For history depth, a practical benchmark is to migrate all open deals regardless of age, plus closed deals from the trailing 18 to 24 months. Older closed deals add import time and storage but rarely change current reporting. If your source system has more than five years of history, export the full dataset to cold storage and migrate only the recent window into HubSpot.

On accuracy, a healthy migration reconciles to within 1% of source counts for each object after deduplication. Anything beyond that gap usually means an import silently dropped rows — often due to required-property violations or malformed dates. Run a count comparison after every import stage, not just at the end, so you catch the failure while it is still small. Finally, plan for a freeze window: stop reps from editing records in the source system for the final 24 to 48 hours before cutover, or you will migrate stale data and lose the edits made during the import.

Risks, edge cases, and failure modes

The most damaging failure mode is importing deals without associations and only discovering it weeks later when pipeline reports look wrong. HubSpot will happily accept a deal row with no company, so nothing errors — the data is just quietly useless. Guard against this by spot-checking 20 to 30 deals immediately after import, opening each one to confirm the company and contacts are attached and the activity timeline shows historical engagements.

A second risk is the partial import that appears complete. If your file has 10,000 rows and 300 fail validation, HubSpot reports the failures, but teams often skim past the summary and assume success. Always download and review the error file, fix the rows, and re-import. Because you stored the legacy ID, the re-import updates existing records instead of duplicating them.

Edge cases worth planning for: deals with multiple companies attached in the source system, which HubSpot handles differently and may need a primary company designated; contacts associated to more than one company, common in agencies and franchises; and currency. If your source CRM stored deals in multiple currencies, HubSpot needs each deal's currency set correctly plus a company currency configured, or every amount converts wrongly and your totals are meaningless.

A subtler failure is owner and team mapping. If a rep left the company, their deals have no valid owner in HubSpot. Decide whether those deals go to a former-rep placeholder user, a manager, or a shared queue — but decide before import, not after, because reassigning thousands of deals later is tedious.

Finally, compliance and retention. If your existing customer data includes personal data subject to GDPR or similar regimes, migrating it into HubSpot is a new processing activity you should document. Confirm your lawful basis, honor deletion requests that were pending in the old system, and do not migrate records you were legally required to purge. Skipping this step creates risk that far outweighs any reporting benefit. Also verify that any field containing sensitive categories — health, financial, or biometric data — is either excluded or handled under the appropriate safeguards before it lands in HubSpot.

A practical rollout plan

Run the migration in six phases, each with a clear exit criterion.

Phase one, audit and map. Export every object from your source CRM and inventory the fields. Build a mapping document that pairs each source field to a HubSpot property, noting which need custom properties created first. Identify your unique identifiers — email for contacts, domain for companies, legacy ID for deals. Exit criterion: a signed-off field map covering every property you intend to migrate.

Phase two, clean and deduplicate. Normalize company domains, standardize phone formats and country codes, and merge duplicates in the staging data before they ever reach HubSpot. Resolve missing associations. Exit criterion: staging files with duplicate rates under 1% and every deal linked to a company.

Phase three, import companies and contacts. Create the legacy ID custom properties first. Import companies, verify counts and domain matching, then import contacts with company association. Exit criterion: company and contact counts reconcile to within 1% of source, and a sample of 20 contacts shows correct company links.

Phase four, import deals and activities. Map stage names to your HubSpot pipeline stages carefully — a mismatch here silently corrupts every pipeline report. Import deals with amount, close date, create date, stage, and owner, then import notes, calls, and emails with original timestamps. Exit criterion: open deal count matches source exactly, and 20 spot-checked deals show correct stage, amount, owner, association, and activity history.

Phase five, reconcile and train. Compare win rates, pipeline totals, and average deal size between source and HubSpot. Investigate any gap over 2%. Train reps on the new pipeline stages and on how to read the migrated history. Exit criterion: reports match within tolerance and reps can find their deals.

Phase six, go live and archive. Freeze the source system, run a final delta import for anything changed during the project, then retire write access to the old CRM and store a full export in cold storage. Keep the legacy ID property forever — it is your audit trail.

Related questions

How long does a HubSpot data migration usually take?

For under 5,000 records, two to three weeks. For 20,000 to 100,000 records with history, plan four to eight weeks. The import itself is fast; deduplication, owner mapping, and reconciliation consume most of the timeline.

Can I migrate deal history without the original CRM's API access?

Yes. A full CSV export of deals, contacts, companies, and activities is enough. You lose nothing essential as long as you capture stage, amount, close date, create date, owner, and associations in the export.

What happens to deals if I import them before companies?

They become orphaned — HubSpot accepts them but they carry no company association, which breaks pipeline and forecasting reports. Always import companies first, then contacts, then deals.

Do I need a custom property for the old record ID?

Strongly recommended. A "Legacy CRM ID" property lets you re-run corrected imports without creating duplicates and gives you a permanent crosswalk for auditing after go-live.

Should I migrate closed deals from five years ago?

Usually no. Migrate open deals plus closed deals from the trailing 18 to 24 months, and archive the rest in cold storage. Older history rarely changes current reporting and slows the project.

FAQ

Will migrating my existing customer data into HubSpot lose deal history? Not if you import associations and original timestamps. Deal history survives when each deal keeps its stage, amount, close date, owner, and linked company and contacts, and when activities carry their original dates. Skipping associations or dates is what loses history.

How do I avoid creating duplicate records during the import? Deduplicate in your staging file before importing, and store a legacy ID custom property on every object. On re-import, HubSpot matches on that ID or on email and domain, updating existing records instead of creating new ones.

Can HubSpot import activities like calls, emails, and notes with historical dates? Yes. HubSpot's import tool and API support engagement imports with original timestamps, so migrated activities appear on the correct date in the contact and deal timelines rather than all clustering on migration day.

What is the correct import order for a HubSpot CRM migration? Companies first, then contacts with company associations, then deals with company and contact associations, then activities. This order ensures every association target already exists when the dependent record arrives.

How do I handle deals owned by reps who have left the company? Create a placeholder user or assign those deals to a manager or shared queue before import. Decide the rule in advance, because reassigning thousands of unowned deals after go-live is slow and error-prone.

Does migrating data into HubSpot create GDPR or compliance obligations? Yes. Moving existing customer data into a new system is a new processing activity. Confirm your lawful basis, honor pending deletion requests, exclude sensitive fields you are not cleared to process, and document the migration.

Sources

• HubSpot Knowledge Base — import and migration documentation: https://knowledge.hubspot.com
• HubSpot Developer Documentation — CRM API and associations: https://developers.hubspot.com/docs/api/overview
• HubSpot Academy — CRM data management courses: https://academy.hubspot.com
• Salesforce Help — data export and migration guidance: https://help.salesforce.com
• Microsoft Learn — Power Query for data transformation: https://learn.microsoft.com/power-query
• GDPR official text and guidance: https://gdpr-info.eu
• Pipedrive Knowledge Base — exporting CRM data: https://support.pipedrive.com
• Zoho CRM Help — data export and backup: https://help.zoho.com

Related on PULSE

• [How do I deduplicate contacts and companies in HubSpot without breaking associations?](/knowledge/sw0102)
• [What is the right way to map legacy CRM pipeline stages to HubSpot deal stages?](/knowledge/sw0114)
• [How do I set up HubSpot custom properties for a clean data migration?](/knowledge/sw0121)
• [Which CRM migration mistakes most often break forecasting reports?](/knowledge/sw0133)
• [How do I archive a retired CRM so historical data stays auditable?](/knowledge/sw0140)