4 Data Processing
INTRO
Here you’ll learn how we handle your data and how it is transferred to the online shop.
We distinguish between two types of data:
Product data – including title, description, brand, GTIN, images, attributes, and accessories
Offer data – including prices, stock levels, warehouse information, VAT rates, and other trade-specific information
To provide our customers with the best possible product information, we use automated processes, some of which are AI-assisted. These enable us to:
Identify the highest-quality information from various data sources
Efficiently detect and prevent pricing and stock errors
Process large volumes of data at scale
The final presentation of your product information in the online shop is influenced by the following factors:
Data quality – the more complete and structured your data is, the more accurate the presentation
Time of submission – the freshness and order of data delivery are crucial
AI optimization – our systems automatically optimize the data to ensure the best possible customer experience
Because this process is fully automated, we can only make limited manual adjustments to product data. Your data is optimized by our AI and may therefore appear in an improved form in the online shop meaning it may differ from the originally transmitted data.
Table of Contents
FROM DATASET TO PRODUCT IN THE SHOP
Assignment
If the GTIN already exists in our system, your product and offer data is assigned to the existing product. We take your product data into account and enrich the existing product with it.
If the GTIN does not yet exist in our system, a new product is created.
Without a GTIN, your data cannot be processed and will therefore be ignored.
Product creation
As with assignment, we also require valid offer data for product creation. In addition, the following product data is required:
Title
Brand
Valid product type assignment
With this information, we can check the specific quality requirements and the legal compliance information.
Product type is assigned using AI. The following data is used as input:
Title
Images
Descriptions
Category path
Important: Due to continuous assortment growth and changes in the market, our structure is constantly evolving. New and existing products may therefore change product type over time. To respond to structural changes automatically, it is important that we receive the most accurate raw data possible as input. We advise against mapping to the Galaxus structure, as it can change regularly.
If the data is incomplete or inaccurate, the product type cannot be assigned.
Product type missing
Without a product type assignment, we cannot check data quality and compliance. The product will remain offline.
Product-type-specific attributes
The requirements for a product of the product type "T-shirt" differ fundamentally from those of the product type "wooden table". This information is therefore crucial to Galaxus’ quality standards.
Clearance & Quality
5 Product Data quality guidelines
A product with only minimal information is not considered purchasable and must pass our quality assurance. The following factors are checked:
Images
Title
Quality attributes
Compliance attributes
If this information is missing, the product cannot be published in the online shop.
Offer data is checked against the following criteria:
Valid prices
Correct stock information
Valid compliance
Images are sorted, filtered (e.g., to remove logos), and cropped using AI.
Product attributes are assigned to our structure using AI. Manual mapping is not required. Attribute updates are applied once a threshold of new attributes is exceeded.