Pimcore PIM catalog for 100k+ SKUs
Within the project team, I worked on the backend/PIM scope: catalog model, bulk loading, Excel tools, barcode-based media matching, product variants, and data-exchange flows with external systems.
- Role
- Backend developer in the project PIM team
- Context
- Beauty retail and e-commerce growth
- Scale
- 100k+ product positions at launch
- Timeline
- 3 months to production launch
Business task
E-commerce was growing faster than the old catalog process. Partner data arrived in different formats, managers reconciled attributes manually, images and variants needed separate checks, and online-store publishing depended on one-off uploads.
My responsibility
I worked on the backend/PIM scope: turning catalog-team requirements into a data model, loading rules, bulk-operation flows, and exchanges with external systems. The focus was not a one-time migration, but a workflow catalog managers could use every day after launch.
Catalog model
Hierarchies, attributes, and relations were designed around product pages, search, filters, and long-term assortment maintenance. The goal was not simply to create fields, but to keep the model flexible for new categories, attribute changes, and ongoing content updates.
Bulk operations
The 100k+ product launch required bulk creation and editing, Excel import/export tools, validation rules, and approval steps. That reduced the catalog team’s dependency on ad-hoc development tasks.
Media and variants
Images were matched to products by barcode, and product variants were grouped into parent cards. Repetitive manual work became part of a controlled process instead of a separate operation before each publication.
Integrations and launch
Partner-data imports and online-store exports were split into separate flows. That made it possible to stabilize incoming data independently from publication and control which changes were pushed to external systems.
Business value
At launch, the catalog already supported 100k+ SKUs, bulk operations, and a clear product path from incoming data to storefront. For managers, this reduced manual coordination; for engineering, it lowered routine requests and clarified data ownership.