Overview
The Data Extractor turns unstructured product data, like values buried in multi-value fields like Tags, into clean, separate fields. Once extracted, these fields can be used across Nosto: as facets and searchable fields in Search & Categories, or as filtering and grouping criteria in Recommendations and Dynamic Bundles. You define an extraction rule once, and the fields become available everywhere they are supported.
The Problem It Solves
Product data is often not structured the way merchandising tools need it to be. Common examples:
Product attributes are stored in Tag1 or other multi-value fields together with other attributes
Attribute values come as a string like Field:Value1,Value2 or "Value1,Value2"
Attribute values come as a JSON array like ["Value1", "Value2"]
A tag like material:wool carries real merchandising value, but as long as it sits in a generic Tag field, it cannot be used as a precise rule. The Data Extractor pulls these attributes out into their own fields, so "material" becomes something you can filter, facet, and merchandise by.
Use cases:
Create more facets
Merchandise by specific attributes, e.g. promotion type
Add as Searchable Fields for better search relevance
Filter and group Recommendations by extracted attributes
Define Dynamic Bundles by extracted attributes
Build custom frontend template elements
Setting Up an Extractor
Decide which attributes you would like to extract into a separate data field.
Go to Catalog Explorer → Settings → Data Extractor.
Give the new extractor a specific name, e.g. "Attributes from Tag1".
Select the type of your Data Extractor based on the field's data structure.
Select the field where your desired attributes are placed now.
(If needed) Set a separator: the character that separates the attribute's name from its value. For example, if the attribute is Style:Casual, your separator is :
By default, the extractor picks up every attribute it finds in the source field. That is usually the best choice: the extra fields simply sit there until you need them, at no cost. If you want to extract something more specific instead, two optional filters narrow it down:
Name Filter extracts only the attribute you name. Set it to Style, and the extractor ignores everything else in the field.
Value Filter leaves out specific values. Add Casual to the list, and the extractor picks up every Style value except that one.
Once you save your extractor, Nosto processes your catalog and extracts the new fields. This can take up to 24 hours. The sections below show how the fields become available in each area.
Verifying Your Extracted Fields in Product Intelligence
Before building rules on top of your new fields, you can check what was actually extracted. Open a product in Product Intelligence: the product detail view shows all extracted fields and their values for that product. If something looks off, adjust your extractor settings before going further.
Using Extracted Fields in Recommendations & Dynamic Bundles
In Recommendations and Dynamic Bundles, no extra setup is needed. As soon as the extraction has run, the fields show up automatically in their own Extracted Fields section, right next to Custom Fields.
In these areas, extracted fields support the same options as Custom Fields:
Recommendation Filters: use extracted fields in both include and exclude rules, with the same conditions you know from Custom Fields, e.g. matching or differing from the displayed product.
Grouping: group recommendation results by an extracted field, the same way you would by a Custom Field.
Dynamic Bundles: use extracted fields as bundle attributes when defining what belongs in a Bundle.
Using Extracted Fields in Search & Categories
Search & Categories works on indexed data, so your extracted field needs to join the index before you can use it:
Go to Search & Categories → Settings → Indexed Fields and add your new field.
The next re-index picks it up. Re-indexing runs 4 times a day - see "Data Sync Frequency" for details.
Your field can now be used in facets, and as searchable fields.
You can also use your field in Merchandsing Rules targeting Search, Collections & Recommendations.
Summary
The Data Extractor turns attributes buried in Tags and similar multi-value fields into clean, separate fields, defined once in Catalog Explorer → Settings → Data Extractor.
Product Intelligence shows the extracted fields and values per product - the quickest way to verify your extraction works as expected.
Recommendations & Dynamic Bundles: extracted fields appear automatically once extraction has run - no extra steps - in their own Extracted Fields section, with the same options as Custom Fields.
Search & Categories: add the field under Indexed Fields, and the next re-index (4 times a day) picks it up for use in facets, merchandising, and search.