Question 1

Which HubSpot objects can a Databricks model write to?

Accepted Answer

Companies, contacts, deals, tickets, products and line items, activities such as calls, meetings, emails, notes and tasks, plus invoices, services, custom objects, static lists and custom behavioural events. That covers most of the CRM, so a propensity score can land on a company record, product milestones can be logged as events, and a static list can track a model you maintain in the lakehouse.

Question 2

How does Polytomic read from Databricks?

Accepted Answer

Through the SQL endpoint you configure, using the Databricks SQL driver. The documented setup points at a SQL Warehouse. Reads are ordinary SQL queries rather than direct access to Delta files or the underlying object storage, and Polytomic does not provision a cluster or warehouse of its own. That also means the staging storage a Databricks destination needs is not involved here, since a source only connection requires no bucket at all.

Question 3

Will the sync wake up a stopped SQL warehouse?

Accepted Answer

Yes, and it is worth planning for. Polytomic reuses the warehouse you configured rather than creating one per sync, so if that warehouse has auto stopped between runs, the next query wakes it. That adds startup latency and compute cost to the run. The connection retry allows roughly ten minutes for compute to start, so a sync will usually wait rather than fail. If schedules are spaced wider than your auto stop interval, expect a warm up on every run.

Question 4

Can Polytomic create HubSpot properties?

Accepted Answer

No. Existing custom properties can be written to, but property creation is not exposed, so create the property in HubSpot first and then map to it. It is one extra step and worth keeping in the process, because it means a mistyped mapping fails cleanly rather than quietly littering a CRM other people depend on with stray fields.

Question 5

What changes when a score lands on the record?

Accepted Answer

Which accounts get worked. A propensity or health score sitting in a table is an artefact nobody outside the data team sees, and asking reps to open another tool to check it does not survive a busy week. Reverse ETL writes it onto the company record, so the model output lands where the decision happens, using the same definition your reporting uses rather than a parallel one rebuilt in the CRM.

Sync Databricks to HubSpot

Sync Databricks to HubSpot with Polytomic

Most of the CRM is writable

No bucket for reads

Incremental on a tracking column

Or sync HubSpot to Databricks

Trusted by businesses worldwide

Step-by-step integration guide

Connect Databricks in Polytomic

Connect HubSpot in Polytomic

Create a sync between Databricks and HubSpot

Map fields and enable the sync

Frequently asked questions

The sync and ETL platform that does it all

Powerful data sync

Advanced analytics

Enterprise security

Easy setup

Ready to connect Databricks with HubSpot?