HubSpot + Datacor Chempax integration

Chemical distributors and blenders buy Datacor because it understands a drum, a tote and a bulk load of the same product, and the safety data sheet that has to travel with each. It prices by customer, container and quantity break, ties every lot to its certificate and its shipment, and prints the placards a truck needs. For years the product was called Chempax, and Datacor renamed it Datacor ERP in 2019.

Datacor also sells a CRM beside it, eChempax until 2019 and Datacor CRM since, which reads pricing, inventory and orders from the ERP directly. Moving marketing and the pipeline to HubSpot gives that up, and the old questions come back. Which accounts are buying less this quarter? What did we charge them for a tote last time? Is the order on the truck? Answering them means moving ERP data into HubSpot on a schedule someone can trust.

The complication is access. Datacor does not publish an API reference, and its support site requires a customer login. What it does describe in public are two copies of the data built for other tools to read: DataLab, a managed data lake with Datacor ERP data pre-loaded, and CRM Premium’s nightly synchronization to your own servers. A HubSpot sync should read one of those rather than the live ERP.

Six objects, all flowing into HubSpot. Each Datacor customer maps to a company and its people to contacts, orders map to deals, invoices sit on a custom object, and saleable pack sizes fill the product library.

The sixth object is the one sellers ask for first. Datacor keeps a customer product profile: what each account buys, with the dollars, volumes and margins behind it. On a custom object keyed by customer number and product code, it shows which accounts are buying less without a phone call to inside sales.

Products need a decision. Datacor rolls pack sizes up to a master product for reporting, but a customer orders a drum or a tote, not a master product. The HubSpot product library carries the saleable pack size, keyed on its product code, with the master product as a property for roll-ups.

What stays out is anything regulated: safety data sheets, certificates of analysis, lot genealogy and hazardous material detail. Datacor generates and distributes those under its own controls, and a CRM copy would drift.

The runtime is a polling watermark sync over the warehouse copy, the reverse ETL guide pattern with Datacor’s pipeline standing in for the warehouse load. Datacor refreshes the copy: CRM Premium daily or nightly, DataLab through managed pipelines under service levels.

Our service connects to the copy with read-only credentials. Before each run it checks when the copy last refreshed, and it stops with an alert if the copy is stale. It then reads rows past a watermark per object and runs batch upserts keyed on Datacor’s numbers, followed by a separate pass that associates contacts, orders, invoices and profiles to their company.

State lives in the service. A replay endpoint re-reads an object for a date range, and a reconciliation endpoint counts rows in the copy against records in HubSpot and reports the difference rather than fixing it quietly. The live ERP never sees a query from us.

What decides the route here is which copy of the data your licence includes. With DataLab or CRM Premium in place, a service reads the copy and the build is predictable, once discovery confirms which tables the copy holds. With neither, discovery asks Datacor two questions: what API access sits behind its open API statement, and whether a scheduled report export can land on an SFTP drop. The answers set the route and the price.

A managed sync on Skyvia is the lighter option when the CRM Premium warehouse sits on your own database server. It suits a fixed mapping of a few objects, and it runs out of room when customer product profiles and pack sizes need logic between the read and the write.

1. Scope. We confirm the Datacor version, which warehouse copy you license and what it holds, and whether Datacor CRM stays beside HubSpot.
2. Architect. We fix the six objects, the Datacor number that keys each, the pack size rule and the owner of every contact field.
3. Build. The service is written against the warehouse copy and a HubSpot sandbox, with read-only credentials and a portal allowlist.
4. Validate. We compare counts and open order values with Datacor’s reports, then review the product profiles with a seller before cutover.
5. Operate. We watch refresh times and run history, replay what a late copy missed, and re-check the mapping when Datacor upgrades.