Value Mapping & Approved Values

Value mapping translates raw supplier values into the standardized vocabulary defined in your workspace schema. Where column mapping handles the names of fields, value mapping handles the contents inside those fields.

Approved Values

For every select or multiselect attribute in your schema, you define a set of approved values — the canonical vocabulary for that attribute. For example, your color attribute might have approved values like black, white, red, navy blue, and green.

Each approved value can optionally have an external PIM code — an alphanumeric identifier used for integration with external systems. For example, the approved value navy blue might carry the PIM code CLR042.

How Value Mappings Are Stored

Value mappings are stored per adapter (supplier), per attribute. The structure is a nested map:

{
  "color": {
    "Midnight": "black",
    "meadow": "green",
    "clr_029": "navy blue"
  },
  "material": {
    "Eiche massiv": "solid oak",
    "MDF/particle board": "MDF"
  }
}

Each key is a raw value from the supplier's feed, and each value is the approved value it maps to in your schema.

Matching Order

When FeedPrep encounters a raw value, it resolves it using the following priority:

Exact match — The raw value matches a mapping or approved value exactly.
Case-insensitive match — The raw value matches when case differences are ignored (e.g., BLACK matches approved value black).
Wildcard match — The raw value matches a mapping that uses wildcard patterns.

What Happens with Unapproved Values

When a raw value doesn't match any mapping or approved value, the behavior depends on the attribute type:

Single-select attributes: The value is nullified — the field is set to empty rather than storing an unapproved value.
Multiselect attributes: Only the approved values are kept; any unapproved values in the list are removed while valid ones are preserved.

In both cases, unapproved values create inbox items for your review. From the inbox, you can decide how to handle each value:

Map to an existing approved value — Create a value mapping so this raw value resolves to an already-approved value in the future.
Approve as a new value — Add it to your schema's approved values list, making it a valid option going forward.

AI-Assisted Value Mapping

FeedPrep's AI can suggest value mappings by analyzing raw values against your approved values list. This is particularly useful when onboarding a new supplier with hundreds of unmapped values.

AI value mapping costs 1 credit per batch of up to 25 values. Suggestions are always presented for your review before being applied.

Strictness Modes

Each attribute can be configured with a strictness mode that controls how the AI generates suggestions:

Mode	Behavior	Best For
Strict	Always maps to an existing approved value. Never suggests creating new values.	Controlled vocabularies where the list is complete (e.g., sizes, standard colors).
Moderate	Prefers mapping to existing approved values, but will suggest a new value when the raw value is clearly distinct from anything in the current list.	Most attributes — a good default balance.
Flexible / Lenient	Willing to suggest new approved values when appropriate, expanding your vocabulary as new supplier data arrives.	Open-ended attributes where new values are expected (e.g., brand names, product categories).