Commerce Cloud Einstein Site Recommendations

Einstein Site Recommendations predicts the most relevant products to promote to each individual shopper based on your recommendation specifications.

After you create a site recommendation, it is added to your list of recommendations. You can clone and delete listed recommendations and view recommendation history.

Recommenders

Commerce Cloud Einstein has different types of product recommendations that can be placed on specific pages, and different strategies to utilize them.

Where to Place Predictive Recommendations

Where you place recommendations affects the shopping experience. Some recommenders only work on specific pages, and some work on any page.

ISSlot Context Storefront Page Product to Product Products in a Category Products in all Categories (No Anchor Required) Recently Viewed (No Anchor Required)
Global Home page    
Global Category Landing page    
Category Category Landing page    
Global PDP    
Global Cart  
Global Wishlist    
Global Checkout    
Global My Account    
Global My Recommendations    
Global 404 Error    

Recommender Types

Consider these recommender types and their strategies when deciding where to place a recommendation. Available strategies change based on the type of recommender.

Type Description Anchor Expected Typical Placement Available Strategies Default Primary and Secondary Strategies
Product to Product Given a product or list of products, similar or related affinity products are recommended. product-id PDPs
  • Product affinity algorithm
  • Customers who viewed also viewed
  • Customers who viewed ultimately bought
  • Customers who bought also bought
  • Real-time personalized
  1. Customers who viewed also viewed
  2. Product affinity algorithm
Products in a category Given a category, products from within that category are recommended. category-id Category pages
  • Recent top sellers
  • Recent most viewed
  • Real-time personalized
  1. Real-time personalized
  2. Recent top sellers
Products in all categories Products from across all categories are recommended. None
  • Home page
  • My Account page
  • My Recommendations page
  • Footer
  • Cart
  • Mini-cart
  • Wishlist
  • Checkout
  • Recent top sellers
  • Recent most viewed
  • Real-time personalized
  1. Real-time personalized
  2. Recent top sellers
Recently-viewed Products recently viewed by the shopper are shown. None Any page Recently viewed None

Strategies

The first recommendation strategy you select is the primary strategy. You can also have a secondary strategy as a backup in the rare case when the primary strategy doesn’t return any recommendations or has insufficient data to produce high-quality results. Configuring too many restrictive rules can constrain the recommender results. In this case, the secondary strategy provides additional coverage.

We recommend using one strategy that applies to your site's entire customer base and another strategy based on an individual's history to provide personalized experiences for both existing and new customers.

Table 1. Strategies That Use Your Site's Entire Customer Base
Strategy Anchor Expected Result
Customers who viewed also viewed product-id View-to-view correlations
Customers who viewed ultimately bought product-id View-to-buy correlations
Customers who bought also bought product-id Buy-to-buy correlations
Recent top sellers category-id or none
  • Top-revenue products

    Products within a specified category or from all categories when a category is not specified are recommended.

  • Products in all categories

    If the shopper's location and device type are available, products with the highest trailing seven-day revenue among shoppers visiting from that location on that type of device are recommended. If the shopper's location and device type are not available, products with the highest trailing seven-day revenue among all shoppers are recommended.

  • Products in a category

    If the shopper's location and device type are available, products with the highest life-to-date revenue among shoppers visiting from that location on that type of device are recommended. If the shopper's location and device type are not available, the products with the highest life-to-date revenue among all shoppers are recommended.

Recent most viewed category-id / none Most viewed products within a specified category or from all categories when a category is not specified are recommended.

Maximum number of recent most-viewed products is 10.

Table 2. Strategies That Use Individual Customer Experience and History Blended with Your Site's Entire Customer Base
Strategy Anchor Expected Result
Product affinity algorithm product-id Model-generated affinity recommendations based on the purchase history of the entire customer base are used.
Real-time personalized None The highest ranked products for a specific user based on the user’s recent browsing history are returned. The most recent four products that the user is most likely to be interested in viewing next are shown.

Create a Recommender

When you create a product recommender, you specify the type of recommendation and the strategy you want to use. You can also add rules to fine-tune the recommendation.

  1. On the Configurator home page, click the Site Recommendations tab.
  2. From the Sites Recommendations page, click New Recommender.
    1. Name the recommender.
      Use only letters, hyphens, and underscores, up to 100 characters. Spaces are not allowed.
      Note: After creating a recommender, you can't change its name.
    2. Choose a recommender type.
    3. Click OK.
  3. Configure the general settings.
  4. Add up to three strategies.
    1. Select a strategy.
    2. Select a category to filter with, if necessary.
    3. Click Add Strategy.
    4. Use the arrows to order the strategies. To remove a strategy, click the trashcan icon.
  5. (Optional) Add up to 30 rules to fine-tune which recommendations shoppers see.
    1. Click Add Rule.
    2. Select an action, criteria, and values.
    3. Use the arrows to order the rules. To remove a rule, click the trashcan icon.
  6. Preview your recommender before saving by selecting categories.
  7. Click Save.
    The Recommender is added to the Site Recommendations list.