Step by step: Hubspot connector from Fivetran to dlthub

A dltHub data engineer describes migrating a HubSpot-to-MotherDuck sales pipeline from Fivetran to dlt, using the existing Fivetran output and its dbt_hubspot package as a reference for schema, transformations, and validation. The process involved reverse-engineering undocumented connector behavior, building a mostly declarative dlt pipeline with custom resources for batch association and fan-out API endpoints, and recreating 13 dbt models as dlt transformations while preserving output compatibility through placeholder fields for unavailable Fivetran metadata. Comparison of the resulting marts exposed hidden differences, including Fivetran’s undocumented renaming of HubSpot’s pipeline property and its separate handling of deal-stage changes, underscoring that row-level validation is more important than code review alone during migrations. The author concludes that building a replacement connector can be well scoped, but verifying parity reveals the main unknowns, and estimates that teams migrating three to five pipelines can complete similar work in about a week.

Use this post, company, and trend context to find content marketing opportunities, perform competitive analysis, or address product feature gaps via the Plushcap MCP server or the Plushcap API.