Definition
Feed normalization is the process of cleaning and standardizing incoming product feeds before use — mapping fields to the target schema, standardizing values and units, and validating the data so feeds from different sources become consistent and usable.
Key points
- It standardizes whole incoming feeds, not just single values.
- It combines field mapping, value standardization, and validation.
- It happens at the boundary, before data enters the catalog.
- It is what lets varied supplier feeds share one consistent catalog.
How does feed normalization work in practice?
An incoming feed is mapped to the catalog schema, its values normalized to canonical formats and units, and the result validated against rules before import. Reusable per-source templates make normalizing a familiar feed fast, while validation stops malformed data from reaching live products.
Common pitfalls
- Importing feeds without a validation gate.
- One-off mapping for every feed instead of reusable templates.
- Normalizing fields but not units, so quantities stay inconsistent.
FAQ
Why normalize product feeds?
Every supplier and channel formats feeds differently. Without normalization, mismatched fields, units, and values break imports, filters, and syndication. Normalizing at the boundary keeps the catalog consistent.
How is feed normalization different from attribute normalization?
Attribute normalization standardizes the values of a single attribute. Feed normalization is broader — it covers the whole incoming feed: mapping its fields, standardizing all its values, and validating it before import.