Transforming Data into Delivery: PostNL’s BDC Journey
Overview
| Experience | In Person |
|---|---|
| Track | Data Sharing & Collaboration |
| Industry | Transportation |
| Technologies | Databricks SQL, Delta Sharing, Unity Catalog |
| Skill Level | Intermediate |
| DOWNLOAD SESSION SLIDES | |
PostNL is transforming how it uses data to drive operational efficiency, innovation, and better customer experiences across its logistics network. In this session, PostNL shares its journey with SAP Business Data Cloud, highlighting how a unified data platform enables faster insights, stronger collaboration between teams, and scalable AI initiatives. Learn how PostNL is building a future-ready data foundation that supports decision-making at the speed of delivery and unlocks new business value.
Session Speakers
Roel Schouten
/Domain Architect
PostNL
Full Summary
Inside PostNL's dual‑platform data journey with SAP BDC and the Databricks Platform
PostNL, a 225‑year‑old Dutch postal and parcel leader, is rebuilding its data operating model around federated governance and domain ownership. Rather than force a single technology, the company paired SAP Business Data Cloud (BDC) and the Databricks Platform to match the distinct needs of its business domains, while keeping a single governance plane. The session offers a grounded view of what it takes to make two major platforms interoperate in pursuit of a data mesh ambition, and where a modern Databricks Platform fits.
FAQ
Domains differed in SAP footprint, authorization complexity, and migration constraints. Finance and HR were tightly coupled to S/4HANA semantics and BW heritage, while other domains had no SAP dependency. Matching platforms to domain needs proved faster and lower risk than forcing uniformity.
For BDC to Databricks, teams replicate HANA models to BDC's object store as local tables, declare a data product, and share via BDC Connect, then manually mount the share in the Databricks Platform. For the reverse, dbt invokes the BDC Connect SDK to register products and, with naming conventions, auto‑mounts them into the correct BDC space.
There is no clean workaround. Schema changes require new products, which nudged teams to publish wider datasets to avoid frequent rework, a tension with narrowly scoped data products.
No formal benchmarks were run, but querying large tables, such as order‑line detail, performed better than expected in day‑to‑day use. A compute premium applies in the Databricks Platform when consuming BDC Connect data.
Start with governance: a clear domain map, strict responsibility boundaries, and productized interfaces. Confirm the platform team can support two stacks and keep pace with frequent product updates, then select technology accordingly.