Columns you need
| Column | Notes | Tasks |
|---|---|---|
| Required | Note body | Task title, Due date |
| Parent identifier | Contact email, company domain, or a Record ID or unique property | Same (optional for tasks, but a task without a parent is hard to find) |
| Activity date | Map it, or HubSpot uses the import time | Map it for historical tasks |
| Owner | Activity assigned to (user email) | Activity assigned to |
| Optional | — | Status, Priority, Type, Task notes |
HubSpot says emails, meetings and notes “must be associated with an object”. Tasks may stand alone, but don’t leave them that way in a migration.
Steps
- Load the parents first. Contacts, companies and deals should already be in HubSpot, ideally with a “legacy ID” property.
- Build one file per activity type. A notes file and a tasks file are easier to fix than one mixed file.
- Start the import: Import > Start an import > File from computer. Choose one file with multiple objects, select Notes (or Tasks) plus the parent object, and map the identifier column to the parent.
- Map the dates and owners. Activity date, Activity assigned to, and Due date for tasks.
- Import 50–100 rows and open a few records to check that the notes sit on the right timeline with the right date and owner.
- Import the rest, within HubSpot’s limits: on paid accounts, 1,048,576 rows and 512 MB per file.
What the native tool does / where it stops
The importer creates notes and tasks reliably. It doesn’t let you update them afterwards: only calls can be updated by import. So a wrong date or owner across 40,000 notes means deleting that import’s notes and running it again. It also won’t carry over the parts that weren’t text: files attached to the old notes need to be uploaded and linked separately.
Where FlitStack AI fits
FlitStack AI moves notes and tasks with their original dates, owners and every association. It checks them in a sample run and can roll back, so mistakes don’t stick. See the HubSpot import guide.