Page tree
Skip to end of metadata
Go to start of metadata


Interaction Studio Classic Only

This use case is for customers using Interaction Studio Classic (formerly Evergage Classic) ONLY. For customers using the Interaction Studio 'Campaigns and Templates' application, refer to the Use Case Library instead.

SmartBundle is an Einstein Recipe ingredient that makes it possible for you to configure recommendation recipes so that a “bundle” of items is suggested from multiple, predetermined product or content categories. Using SmartBundle, you can control what is recommended to visitors while they are engaged with particular products and content items, eliminating the need for merchandisers to spend time manually curating “complete the look” campaigns.  This machine learning algorithm looks at products purchased together and scores them against a shopper’s affinities towards brand, color, category in order to make the most relevant recommendation for that shopper to “complete the look”.


Category Extend Personalization/Machine Learning Vertical Retail, eCommerce
Topic Instruct machine-learning algorithms to specifically promote items from multiple, select categories ID #237
Sections in this Article


For example, suppose a shopper is looking at a blouse or shirt. You can configure the recommendation recipe to suggest items that could work with that top to “complete the look” - like pants, skirts, accessories, or shoes.  SmartBundle understands the shopper's style, category and brand affinities and recommend complementary items which are chosen by the algorithm and boost based on visitor's affinities.

When providing options to “complete the look,” SmartBundle will automatically recommend flats for the shopper more affine to flats, heels for the shopper more affine to heels, and so on. This automates the curation of “complete the look," but more importantly, it speaks to the individual shopper’s preferences and truly personalizes every step of the journey. By suggesting additional items that the shopper could purchase - and reducing the effort it takes to find those items - the shopper is more likely to purchase the additional items.


No segment is required for this play.


A typical goal you should measure when using this machine learning algorithm for merchandising is an increase in AOV. 



Here is a checklist of what you need to do in Interaction Studio to create this play on your own site:

  1. Determine placement and behavior. For example, will SmartBundle work best on the cart page, within a checkout modal, as part of an exit intent message, or within the PDP? What behaviors should trigger inclusion of SmartBundle? Does customer simply view an item detail? Does this augment or replace any other recommendations campaigns on the page?

  2. Associate categories within the Interaction Studio catalog

  3. Recipe

    1. Create a recipe based on the SmartBundle base ingredient

    2. Add inclusions or boosters based on placement and behavior

  4. Item template

    1. Ensure you have an item template that will match the look and feel for your placement

  5. Create a new campaign within Interaction Studio’s Visual Editor

  6. Include the appropriate messaging and Item Template

  7. In the message settings for each experience, ensure the item template is set for Promoted Content using the SmartBundle recipe you’ve created

  8. Test and launch as you would any other campaign