Every migration horror story starts the same way: somebody exported a CSV, imported it, and discovered three weeks later that half the activity history never made it. We map, dry-run and reconcile — and you get a report that accounts for every record.
In short
A safe HubSpot migration has four non-negotiables: a signed-off field mapping document, a dry-run import into a sandbox portal, deduplication and normalisation before load, and a record-level reconciliation report after. Capra Digitals runs all four on every migration. Our largest single move was over 70,000 records with full data integrity.
Signals
Notes, emails, call logs and attachments matter as much as the contact records themselves.
Your legacy CRM has bespoke objects that do not map cleanly to HubSpot's model.
The old contract ends on a fixed day and there is no room to run both systems for a quarter.
Duplicates, dead domains, inconsistent country fields — importing that as-is just moves the problem.
What you get
We pull a full export from the source system and profile it: record counts, field fill rates, duplicate clusters, orphaned associations.
ExploreEvery source field gets a destination, a transformation rule, or an explicit decision to drop it. You sign it before we touch the target portal.
ExploreA full-volume rehearsal into a test portal, so the first time we see an error is not on go-live day.
ExploreThe live load runs in a planned window, then you get a reconciliation report comparing source and destination counts object by object.
ExploreHow it runs
Before promising anything we read the data. Counts, quality, custom objects, attachment volume and how much history is genuinely used.
A working session per object until every field has a home or a documented reason to be left behind.
Full dry run into a sandbox portal. We fix transformation errors there, not in production.
Live load in a planned window, reconciliation report the same day, then a dedupe pass and two weeks of hypercare.
Compare
| DIY CSV import | Capra migration | |
|---|---|---|
| Field mapping | Guesswork at import time | Signed-off mapping document |
| Activity history | Usually lost | Migrated and associated |
| Duplicates | Imported as-is | Detected and merged |
| Verification | Eyeball a few records | Record-level reconciliation report |
| Rollback plan | None | Sandbox rehearsal and cutover window |
Track record
Records in one migration
70k+
Savance ticket history
Duplicates resolved
20k
Merged during the same programme
Data integrity
100%
Verified by reconciliation report
Weeks typical
3–6
Depending on volume and history
Questions
The questions we get asked on almost every call about this work.
Yes. Emails, calls, notes, tasks and attachments migrate with their associations intact where the source system exposes them via API or export. We tell you upfront if a specific object cannot travel.
Most commonly Salesforce, Pipedrive, Zoho, Microsoft Dynamics, Freshsales, Zendesk and spreadsheet-based systems. If it has an API or a full export, we can usually move it.
Typically a single weekend. The dry run tells us exactly how long the live load takes, so we plan the freeze window around a measured number rather than a guess.
We normalise and deduplicate as part of the migration. Deeper remediation — enrichment, bulk correction, ongoing hygiene rules — is covered by CRM data cleanup.
The reconciliation report is the first place we look, and hypercare covers the two weeks after cutover. Because the source system is kept read-only until sign-off, anything missed can be re-loaded.
Next step
Object counts and a note on how much history matters is enough for a realistic timeline and a fixed price.