HUBSPOT ENGINEERING
When HubSpot configuration becomes engineering
There is a point in most HubSpot builds where the work stops being configuration and starts being software engineering. Recognising that line early changes how a project is scoped.
Configuration is choosing between options the platform already offers: pipelines, properties, forms, standard workflows, templates. It is fast, reversible and well documented.
Engineering begins when the requirement no longer maps onto an existing option — when you are designing a data model, deriving state, or making one system agree with another.
Signals you've crossed the line
- Custom objects are needed to describe something the standard objects can't.
- A value must be derived from several sources rather than entered.
- Behaviour depends on events arriving from outside the portal.
- Correctness matters enough that you need to handle failure explicitly.
Why the distinction matters
Engineering work needs modelling, review and a rollback story. Treating it as configuration is how portals accumulate half-finished logic that nobody wants to touch.
A practical split
- 01Model the objects and the state before building anything.
- 02Keep derived values in one place and reference them everywhere else.
- 03Move anything non-trivial out of workflow branches and into code you can test.
- 04Document the intent next to the implementation.
