Einstein Recommendations Catalog

A catalog includes all fields that contain attributes used to make Einstein Recommendations. Einstein Recommendations filters assets based on these attributes. For instance, you can define assets in the recommendation pool based on a product’s color or size, or an article’s author and subject matter.

Einstein Recommendations requires a minimum number of attributes, which varies depending on the application or specifications.

Catalog functionality is used for multiple purposes in Einstein Recommendations. Consider these areas as you decide which fields to add to a catalog.

• Support Collect: Catalog data that’s passed from Collect ties to the catalog in the data via the Product Code (item) and SkuID (unique_id). These fields must match the Collect data to function properly.
• Build User Profiles: When collecting user data, the catalog takes the view ID and uses the tagged fields to build the profile.
• Support Recommendation Generation: Any data from the catalog can be used to build the recommendation.
• Locale and Language Support: If any attributes vary by a locale or language, add the field to the catalog.
• Support Attribute-Based Rules: Data that is in the catalog can support rule generation. For example, if you want only men’s clothes shown in a particular recommendation, a field like “gender” could support this rule. If a data type isn’t in the catalog, you can generate a rule for it.

Marketing Cloud Engagement accepts these types of catalogs.

• Product: Products that are available for purchase.
• Content: Articles, blog posts, videos, or other types of content that are available for view.
• Banner: Image files, such as hero graphics, calls to action, or offers.

You can import one file for each catalog type.

• Einstein Recommendations Catalog Field Definitions
  Catalog fields store your catalog’s asset attributes for Einstein Recommendations. These attributes are both required and recommended catalog fields. You can also add custom attributes. Catalog field values can contain up to 1,024 single-byte characters or 512 double-byte characters.
• Reserved Attributes
  See a list of attributes that are reserved for specific use in Einstein Recommendations. You can’t use these attributes to store other values.
• Add an Attribute to Einstein Recommendations
  An attribute describes a product or content characteristic, such as color, dimension, brand name, subject, or author.
• Configure Custom Attributes for Einstein Recommendations
  You can create custom attributes to use with your recommendations. Unmapped fields are automatically used to make recommendations. For streaming updates, Custom Attributes Configuration is available after Marketing Cloud Engagement starts receiving data from your site.
• Create a Catalog Field Mapping
  To ensure that catalog heading names are mapped to the required application field names, create a catalog field mapping in Einstein Recommendations.
• Import a Catalog
  Import a catalog to use in Einstein Recommendations by uploading a delimited text file. The first row of data must contain the column headings, and there must be at least two additional rows of catalog data. We recommend that you update your catalog daily to reflect the changes on the site by the next business morning. For improved reliability, file size is limited to 2 MB when using the catalog import option.
• Catalog Summary
  The Einstein Recommendations catalog summary includes the type of import used and the currently selected reminders interval. The file history table in the Status Console includes the date, time, and status history of the catalog uploads.
• Catalog FAQ
  Get the answers to frequently asked questions about the Einstein Recommendations catalog feature.