Product Recommendations

Product recommendations are dynamically generated using Commerce Cloud Einstein. You can think of these as implicit recommendations―in contrast to explicit recommendations that directly link objects together.

How Product Recommendations Work

Product recommendations require input data from three sources:

Data source Description

Product information stored in your catalog is collected by running a catalog feed against your Production instance.


Order information generated on your storefront is collected by running an order feed against your Production instance.


Clickstream data is automatically collected from your site on a real-time basis using pixel tracking; you don't have to perform any action to collect this data.

The following activities are automatically tracked:

  • Product View
  • Add To Cart
  • Finish Checkout
  • View Recommendation

This data is collected by Commerce Cloud Einstein, which uses machine-learning models to analyze the collected data and generate product recommendations.

Product recommendations take the form of a list of product IDs (IDs of recommended products). Before this list is passed to your storefront, the system applies business rules―called recommenders―that you define in the Configurator tool. The rules enable you to control the composition of the list of product IDs, enabling you to remove some IDs, change the order of IDs, and so on. In other words, the rules enable you to control which products are shown, hidden, promoted, and demoted.

When the storefront renders the final list of recommended products, they are in the optimal order as defined by your rules.

Product recommendations are rendered in a content slot. A developer must define the content slot within an ISML template. After the content slot is created, you can create slot configurations for the slot. Your content slot configurations―whose Content Type must be set to "Recommendation"―control which template is used to render the recommended products. It also controls which business rules should be applied to the list of recommended product IDs.

Implementing Product Recommendations

The following table describes what steps must be performed to use product recommendations, and describes who performs each step:

Step Who performs this step?
Configure and Run a Catalog Feed against your Production instance.


Configure and Run an Order Feed against your Production instance.


Configure Business Rules (Recommenders) using the Configurator tool. Merchant

Modify ISML Templates to create recommendation content slots and render product recommendations.

Create a Content Slot Configuration to control which template and recommender are used to show the product recommendations. Merchant
X Privacy Update: We use cookies to make interactions with our websites and services easy and meaningful, to better understand how they are used. By continuing to use this site you are giving us your consent to do this. Privacy Policy.