To unlock browse tracking, insights, and learnings, Constructor requires a hierarchy that maps the groups of products shown as category (browse) listing pages on a production site.

Either the group IDs or group names in the uploaded product catalog data must be an exact match to data within the DOM or on the window object of the browse pages.

Group hierarchy data from the catalog helps our AI to identify patterns in user behavior to improving rankings across both search result pages and browse result pages.

There are many options to define a group hierarchy to transmit to Constructor

Preferred format (recommended)

Please refer to the Using the preferred format guide for complete detail on how to build an item group hierarchy file. In summary, this consists of:

1. An item groups file, where each group has three fields: parent_id,id,name
2. An items (products) file, where each product has a group_ids field
   1. Each group_id must match the id values in the groups file
   2. If a single item (product) belongs to multiple groups, insert a pipe (|) character to denote each entry
      For example: foo|bar|baz

Breadcrumbs

Breadcrumbs can be used to define a group hierarchy:

• For each item, provide a breadcrumbs field that shows the path(s) to that product with a > denoting each step in the group hierarchy
  For example: Shoes>Basketball>Jordans
• If single product belongs to multiple groups, use a pipe (|) character to denote each entry
  For example:Pants>Jeans>Slim|Men>Pants>Jeans

ID pairs

A group hierarchy can also be established through ID pair definitions. This would consist of the following:

• A groups file, where each definition has three fields: parent_id, id, and name
• A products file, where each definition has a product_id field
• A mapping file, where each definition has two fields: product_id, and group_id
  • Each product_id must match the definitions in the products file
  • Each group_id must match the id definitions in the groups file

Our engineers will write custom code to transform any of these three formats into our desired format, unlocking AI result rankings and analytics for your team.

If there is another method to build a group hierarchy using your specific product catalog data, please don't hesitate to make us aware and we will do our best to accommodate.