Grouping products which share the same basic set of attributes can help you improve your search results.

The case for product variations: an example.

Say you have these items in your feed

* CAMPER mountain boots. (size: 38, color: blue)
* CAMPER mountain boots. (size: 36, color: blue)
* CAMPER mountain boots. (size: 36, color: red)

Those are indeed three different products and when looking for “mountain boots”, these three search results would appear.

However, you may prefer the user obtained just one single CAMPER mountain boots search results with all three variants that is, the red and blue colors as well as the 38 and 36 sizes, available in the facets panel of your search results.

The de-facto standard

Google has come up with a mechanism to do that in the data feed: the item group id. Long story short: all items with the same item_group_id attribute are considered variants of the same item.

So, what happens when I check the “group products by item_group_id” ?

Just that. All products that share same item_group_id attribute are displayed as one product only in the search results, but all their variants (like color or size), if defined as facets, are displayed in the facets panel.

Back to our first example, when the user looks for “mountain boots”, only one “CAMPER mountain boots” result is shown, but “blue” and “red” colors and “38”, “36” sizes are available at the facets panel.

Follow the leader! Marking an item as the “group leader”

If there’s an item with the field df_group_leader set to either “true”, “t”, “1” or “y”, that item will be chosen as the group representative in search results (if there is no other more relevant sort criteria involved).

NOTICE: If you’re using the API, you need to set a pure boolean value for df_group_leader, either true or false (no quote intended 😉 ).

In a nutshell

Check the group products by item_group_id checkbox` at the Configuration -> Data Sources section of your control panel.
Make sure the products that you want to be grouped share the same item_group_id field in your data feed.
You’re good to go!

What exactly is the name of the field again?

Text Format

Grouping items: For product type feeds, you can use both item group id, item_group_id or group_id. The latter will work for every datatype (not just the “product” datatype)
Marking an item as the group leader: You use the df_group_leader field

XML Format

Grouping items: For product type feeds, You can go with <g:item_group_id> or <item_group_id> or even <group_id>. The latter will work for evey datatype (not just the “product” datatype)
Marking an item as the group leader: You need to use df_group_leader

Below you can download some examples to see how it works.

Grouping product variants in the data feed