What to do if a recommendation has 0% coverage

If you have noticed that a recommendation has 0% coverage, it means that the recommendation was not able to find any matching items in the catalog based on selected event attributes in the recommendation metric.

A recommendation can search items only based on item_id attribute. It means that a data format in events used to generate recommendations must be exactly the same as item_id in the catalog.

Let’s look at this example. Here we have “Popular Right Now” recommendation template that uses purchase_example.item_id attribute to check what products customers are buying.

Currently, the coverage is 0% meaning that the recommendation cannot find any item_ids from Purchase_example.item_id attribute values in catalog test used in the recommendation. Catalog has 2 items, one of them has id Jeans-001-blue.

There is only one purchase_example event in the database currently and it has the following attributes:

As you can see, the item_id attribute of the Purchase_example event is Jeans-001 which is not in the test catalog item_id column. However, Jeans-001-blue, which is stored in the product_id event attribute matches item_id of the item in the test catalog. If you change recommendation settings to use product_id attribute of Purchase_example instead of item_id, the recommendation will be able to find the correct item in the catalog.

In other words, because item_id in the event did not have "-blue" part, the recommendation could not find the exact match in the catalog data.