Data Normalization

What is Data Normalization used for?

Data Normalization groups different source spellings under one consistent master value. For example, navy, Navy Blue and dark blue can map to blue; Extra Large, XL and x-large can share a size group.

Nifelo retains the raw source value and adds a normalized shadow value. The source remains traceable while a sales channel can receive a consistent and translated value. This shared workflow applies to every data source.

Normalization does not correct the source

Normalization does not change the original file, feed or store. Use it for valid variants with the same intended meaning. Correct factually wrong product data in the authoritative source instead of hiding it with a mapping.

The current matrix supports color, size and material. Each field has its own matrix, groups and module switch. Availability may depend on the subscription.

Preparation

The matrix reads distinct raw values from active Smart Products. If values are missing or wrong, first check the Smart Field mapping, run an import after mapping changes, inspect Products and verify product-versus-variant level.

Normalization matrix

Select color, size or material. For every raw value the matrix shows its status and selected master group. Search and pagination help with large sets.

Choosing a group is saved and immediately applied to active products in this source whose raw value matches case-insensitively after trimming surrounding spaces. Navy, navy and navy therefore use the same mapping. Different content such as blue/white is not automatically treated as blue.

Choose No group to release a mapping. The raw value remains available, but no normalized value exists for it.

Manage groups

System groups are centrally managed by Nifelo and cannot be renamed or deleted by customers. Where allowed, custom groups belong only to the selected source. Use stable technical keys rather than campaign text. Some fields intentionally use fixed groups to keep exports predictable.

Renaming a custom key updates existing normalized product values that reference it. Afterwards, check translations, raw mappings and channel configurations. Delete a custom group only after moving dependent raw values where necessary; deletion does not undo already exported data.

Group translations

Use the flag tabs to give one technical group key a presentation value for every active project language—for example blue: Blue, Blau, Bleu. The technical key stays unchanged, so one normalization mapping supports every language. Project settings determine which languages are available.

Enable the module

Color, size and material each have an independent switch.

• Enabled: normalized values may be used by downstream processing and exports.
• Disabled: raw values remain leading while mappings are retained.

Enabling does not automatically map unknown values. New raw values introduced by later imports require review.

Sales channels and other tools

Normalization makes a separate normalized value available; the sales-channel mapping determines whether raw or normalized data is exported and in which language. Normalization itself publishes nothing.

Use the source for factual corrections, Smart Fields for source mappings, normalization for spelling and value variants, rules for predictable business logic, and AI only for suitable content that fixed mappings cannot solve.

Recommended workflow

Verify Smart Fields, import current products, review each matrix, map only truly equivalent values, prefer system groups, add stable custom groups only when necessary, translate all active languages, enable the module after the important values are covered, and verify products and the target channel. Review new raw values after later imports.