SAP ECC and HubSpot integration

What moves between them

Customer and vendor master records flow bidirectionally. SAP ECC customers and vendors are polled on a scheduled interval and synced to HubSpot as companies and contacts, with SAP fields mapped to HubSpot custom properties (vendor number, credit limit, payment terms, plant assignment). Changes to companies and contacts in HubSpot (phone, address, contact name) trigger an update back to the corresponding SAP ECC customer or vendor record via BAPI_CUSTOMER_CHANGE or BAPI_VENDOR_CHANGE. HubSpot deals and line items are mapped to purchase orders in SAP ECC only when explicitly configured per customer.

How ml-connector handles it

ml-connector stores the SAP ECC on-premises agent endpoint (the RFC gateway hostname and port) and the HubSpot API token encrypted, then periodically polls SAP ECC BAPI endpoints via RFC_READ_TABLE or OData to fetch the current vendor and customer rosters. Each record is checked for a unique identifier (vendor number or customer number) and matched to an existing HubSpot company or contact. If a match is found, the HubSpot custom property values are updated; if not, a new company or contact is created. When HubSpot records are changed through the UI or API, ml-connector watches for those changes via HubSpot's webhook system, retrieves the updated fields from HubSpot, maps them back to SAP ECC field names, and calls the appropriate BAPI change function with the ERP transaction commitment. SAP ECC BAPIs require an explicit commit call to finalize changes, and ml-connector includes that in every write request. HubSpot rate limits are managed through request queuing and exponential backoff on 429 responses. On-premises RFC connectivity is validated on each poll cycle, and if the agent is unreachable, an alert is logged and the next scheduled poll is attempted after a brief delay.

A real-world example

A mid-market B2B software company runs SAP ECC on-premises for procurement and vendor management, and uses HubSpot for sales. The sales team sources prospects from industry directories and builds deals in HubSpot, but frequently discovers that deals reference a vendor or customer already in SAP ECC under a different contact name or address variant. When a prospect converts to a customer, the sales team must manually enter the new account into SAP ECC so procurement can issue purchase orders to the same entity. With SAP ECC and HubSpot connected, existing vendors and customers in the ERP appear automatically in HubSpot as company records, and when a new deal closes in HubSpot, the prospect is pushed into SAP ECC as a customer in the same transaction. The procurement team sees new customers immediately, and the sales team no longer re-enters customer data across the two systems.

What you can do

• •Sync vendor and customer master records from SAP ECC into HubSpot companies and contacts on a polling schedule.
• •Map SAP ECC customer numbers, credit limits, and payment terms to HubSpot custom properties so sales has current ERP data.
• •Push new customers and updates from HubSpot deals back to SAP ECC customer records via BAPI_CUSTOMER_CREATE and BAPI_CUSTOMER_CHANGE.
• •Authenticate to SAP ECC via HTTP Basic Auth for OData and RFC gateway credentials for BAPI calls, and to HubSpot via Private App Bearer token.
• •Track and retry failed records with a full audit log, so data mismatches are immediately visible.

Questions

Which records move between SAP ECC and HubSpot, and in which direction?

Vendor and customer master records flow from SAP ECC into HubSpot as companies and contacts. Changes to company and contact details in HubSpot (phone, address, contact person) are pushed back to SAP ECC as customer or vendor record updates. HubSpot deals are read by ml-connector but typically synced to SAP ECC purchase orders only when explicitly configured per customer, since every deal does not map to a procurement need.

Does the integration work with SAP ECC on-premises, or does it require the cloud version?

ml-connector works with SAP ECC on-premises only. It cannot reach SAP ECC directly from the cloud, so it requires an on-premises SAP .NET Connector or Java Connector agent running on the customer's network to originate RFC and BAPI calls. The agent endpoint (gateway host and port) is stored encrypted in ml-connector's configuration and used for every RFC request.

How does ml-connector handle the differences in how SAP ECC and HubSpot represent vendor and customer data?

ml-connector maps SAP ECC vendor and customer fields to HubSpot companies and contacts using a custom field schema. SAP-specific identifiers like vendor number, plant assignment, and dunning status are stored in HubSpot custom properties. When a record is updated in either system, ml-connector translates the field names and types (for example, SAP date format YYYYMMDD to ISO 8601) so both systems remain in agreement.