The hard part of this move is not the export. It is the difference in how the two systems think. Smokeball works in matter records; HubSpot works in contact records. Get that mapping wrong and every report built on it afterwards is wrong too. This page covers the field map, what does not survive, what it costs, and the reconciliation step that decides whether anyone trusts HubSpot in month two.
The two systems, side by side
| Smokeball | HubSpot | |
|---|---|---|
| Primary record | Matter | Contact |
| Grouping | Matter type, which drives the field set | Company associations |
| Activity history | Activity intelligence, notes and automatic time records | Engagements: notes, calls, emails, meetings |
| Documents | Documents on the matter, tightly bound to Word templates | Attachments on records, plus the file manager |
| Getting data out | Vendor-assisted export; local components complicate self-service extracts. | CSV export per object and a well-documented API. |
| Who administers it | None. | None, though marketing usually owns it. |
| Pricing | From about $49 per user per month; four plan tiers. | Free tier through enterprise seats. |
What breaks
- Smokeball works in matter records and HubSpot works in contact records. That is a re-modelling exercise, and it is the first decision, not a mapping detail.
- Grouping changes: Matter type, which drives the field set becomes Company associations. Somebody has to define what a group means in the target before load.
- Activity history moves with partial fidelity. Smokeball stores it as activity intelligence, notes and automatic time records; HubSpot expects engagements: notes, calls, emails, meetings.
- Automations, reports, dashboards and permissions never transfer. Budget them as build work.
- Every connected system has to be repointed and retested: the DMS, billing, e-signature and the shared inbox.
Known hazards on each side
Leaving Smokeball:
- Automatically captured time records are a large dataset with no obvious home in most targets.
- Matter-type-specific field sets mean the same concept appears under several names.
- Document automation templates are Smokeball-specific and get rebuilt, not moved.
Arriving in HubSpot:
- Associations between contacts, companies and deals are what carry structure, and flat exports lose them.
- Marketing email and workflow history does not migrate to a vertical system.
- Property names drift over time because anyone can create them.
The field map
| Concept | In Smokeball | In HubSpot | What to watch |
|---|---|---|---|
| Primary record | Matter | Contact | Different shape. Decide the mapping before any data moves. |
| Grouping | Matter type, which drives the field set | Company associations | The target groups records differently, so grouping is created rather than copied. |
| Activity history | Activity intelligence, notes and automatic time records | Engagements: notes, calls, emails, meetings | Expect partial fidelity. Types, timestamps and authorship survive; structure often does not. |
| Documents | Documents on the matter, tightly bound to Word templates | Attachments on records, plus the file manager | Usually a separate extract and load from the record data. |
| Owner and assignment | Assigned user or advisor fields | Assigned user or team fields | Map departed staff explicitly. Records assigned to nobody disappear from every view. |
| Custom fields | Firm-specific fields added over time | Created to match, or consciously dropped | The audit here is the work. Undocumented fields that reports depend on are the usual surprise. |
Cost and timeline
| Phase | Typical range |
|---|---|
| Data audit and cleanup | $3,000 to $25,000 |
| Migration execution | $5,000 to $50,000 |
| Target configuration | Two to six weeks |
| Ongoing administration | Absorbed by existing operations staff |
The five phases
- Audit Smokeball. Count records per object, find duplicates and orphans, and decide what does not come with you. Get that decision signed off before mapping.
- Map against real records. Use the table above as the starting point, then walk fifty live records field by field.
- Dry run into a HubSpot sandbox. Load the full set, not a sample. Sample loads hide the failures that only appear at volume.
- Cut over on a quiet window. Freeze writes in Smokeball, load, repoint every integration, unfreeze. Keep Smokeball read-only for at least a quarter.
- Reconcile. Compare HubSpot to Smokeball record by record and field by field, and produce a variance report.
The full version of this plan, with the failure evidence behind it, is in the CRM migration playbook.
The compliance clause
State retention rules and the firm's duties over the client file survive the migration. A matter that lands in the target without its history is a records gap, not a cosmetic issue. Keep the field map, the variance report and the mapping sign-offs with the records, because the decisions you made in phase two are part of the record too.
Where Caddi fits
Caddi does the mechanical half and then keeps doing it. Show it the task once and it reads Smokeball, applies your field map, loads HubSpot, and then compares both sides record by record and hands back a variance report you can act on. Because it runs over APIs across your stack, the same agent covers the parts a migration tool ignores: documents filed to the right contact with the right metadata, and the post-cutover hygiene that otherwise becomes someone's Monday.
Caddi automation for these systems: Smokeball and HubSpot.
Related migrations
- Clio Grow to HubSpot
- Clio Grow to Smokeball
- Clio Manage to HubSpot
- Clio Manage to Smokeball
- Filevine to HubSpot
- Filevine to Smokeball
Migration without the manual pass
Move and reconcile Smokeball into HubSpot
Caddi reads Smokeball, applies your field map, loads HubSpot, then compares both sides record by record and reports the variances.
Frequently asked questions
How do you migrate from Smokeball to HubSpot?
Audit the Smokeball data first, then map fields against fifty real records rather than against the schema, dry run the full set into a HubSpot sandbox, cut over on a quiet window with writes frozen, and reconcile record by record afterwards. Vendor-assisted export; local components complicate self-service extracts. HubSpot then needs the records loaded in dependency order, with documents as a separate pass.
What does not transfer from Smokeball to HubSpot?
Smokeball works in matter records and HubSpot works in contact records. That is a re-modelling exercise, and it is the first decision, not a mapping detail. Grouping changes: Matter type, which drives the field set becomes Company associations. Somebody has to define what a group means in the target before load. Activity history moves with partial fidelity. Smokeball stores it as activity intelligence, notes and automatic time records; HubSpot expects engagements: notes, calls, emails, meetings.
How long does a Smokeball to HubSpot migration take?
Six to twelve weeks is realistic for a clean mid-size move, and three to six months where HubSpot is being implemented at the same time. Firms with years of history in Smokeball should budget several weeks of cleanup before any data moves.
Who administers HubSpot afterwards?
None, though marketing usually owns it. That is one of the reasons this move is lighter than a platform migration.
What are the compliance implications for law firms?
State retention rules and duties over the client file survive the migration. A gap in matter history is a finding, not a cosmetic issue, so keep the field map, the variance report and the sign-offs with the records.