Skip to content

The problem

Contacts, companies and deals only go so far.

The default model fits a simple sales funnel. Your business has properties and the entities that own them, distributors and the stores they supply, portfolio companies and the customers they share. Squeezed into three objects, the data ends up in notes fields and spreadsheets.

  • Fields that mean different things on different records
  • Relationships kept in people's heads
  • Platform limits found halfway through the build

The outcome

A model you can read.

Every object, association and owner mapped before the build and documented after it. Records show what each role needs, including data from outside HubSpot. Platform limits are known up front, not discovered in production.

  • Objects and associations that mirror the real relationships
  • Cards and pages that bring outside data onto the record
  • Permissions designed by role, schema kept as code

What we build.

All of it on HubSpot, inside the limits of your edition, with code where configuration runs out.

  • 01

    Custom objects and associations

    Properties, orders, agreements, locations: modeled as their own objects with labeled associations, so a contact can be the principal of one entity and the guarantor on another. Custom objects need HubSpot Enterprise; on Professional we design with companies, deals and association labels, and tell you where that will strain.

    • custom objects
    • association labels
    • schemas API
  • 02

    UI extension cards

    React cards on deal, company and contact records, built with HubSpot's UI extensions SDK. They run in HubSpot's sandbox and fetch outside data through hubspot.fetch: property details, valuations, sale history, demographics, maps, KPIs.

    • @hubspot/ui-extensions
    • React
    • TypeScript
  • 03

    App pages

    Full pages inside HubSpot for jobs that don't belong on one record: an intake page that resolves an address against a data provider, checks for duplicates, and creates the company and deal in one step.

    • developer projects
    • app pages
    • hubspot.fetch
  • 04

    Serverless Workers behind HubSpot

    Cloudflare Workers that hold the secrets, cache the lookups and do the work HubSpot cannot: decay timers, queues, external sends. Every call from HubSpot is signature-checked. HubSpot's own serverless functions need Enterprise; the Workers in our field notes run behind a Professional portal.

    • Cloudflare Workers
    • Durable Objects
    • signature v3
  • 05

    Custom workflow actions

    Workflow steps backed by your code: an app-defined action that holds records in a queue and releases them on a schedule, or sends through your own mail server. On Data Hub Professional, custom code actions run inside the workflow itself.

    • workflow actions
    • callbacks
    • custom code
  • 06

    Permissions, teams and schema as code

    Teams and permission sets designed around who does what. Apps, cards and workflow actions kept as files in a HubSpot developer project in git and uploaded from the CLI; every workflow backed up before we edit it.

    • developer projects
    • HubSpot CLI
    • git
FIG. 01DATA MODEL11:40 CTSample HubSpot data model: standard company, contact and deal objects plus custom location and contract objects and synced orders, joined by labeled associations.

How it works.

  1. Map the business

    Who and what you track, how they relate, and which decisions depend on each relationship.

  2. Check the limits

    Edition, object, property and API limits on your subscription, measured before the design depends on them.

  3. Build in a test account

    Schema, cards and automation built in a developer test account, deployed from code, then promoted to your portal.

  4. Migrate and hand over

    Existing records moved into the new model with their history, and documentation for every object and workflow.

Property data cards in HubSpot

Property data on every record

Fifteen custom cards that put property, valuation, sale history, area and map data on HubSpot deal and company records, with the property data behind a signed, cached Worker.

Read the field note
PROPERTY CARDS10:15 CT

FAQ

Questions buyers ask.

Do we need HubSpot Enterprise for custom objects?

Yes. Custom objects need an Enterprise subscription. On Professional we model with companies, deals and association labels, and we tell you where that design will strain before you build on it.

Can you work inside our existing portal?

Yes. Most work starts in a live portal. We read the current schema and automation first, back up anything we change, and test on a handful of records before touching the rest.

Do cards slow the record down?

Cards load on their own, after the record. The Workers behind them cache lookups so a repeat view does not wait on, or pay for, a second call to the data provider.

How do you handle platform limits?

We check them before designing: calculated-property caps, the search API's result ceiling, which features your edition includes. Where HubSpot cannot do a job, a Worker outside it does, and writes the result back.

Who owns the code?

You do. It lives in your repositories and your accounts, with a README for every component.

Figures show illustrative sample data. Names, companies and values in them are invented.

Show us what doesn't fit.

The relationship you track in a spreadsheet because HubSpot has nowhere to put it. That is where we start.