Product Recommendation Algorithms

This article provides a breakdown of all the product recommendation algorithms available in Justuno.

At a glance:
Introduction
Algorithms
Filters
Common Use Cases for Algorithms

Last update 2/20/2025

Introduction

Product recommendation algorithms continuously analyze your product data and website visitor behavior to showcase items in your inventory that are most likely to interest and convert visitors. These algorithms are managed in the Commerce section within Justuno and are crucial when integrating a Product Recommendation element into your Justuno designs.

The following algorithms are available by default to all Justuno users:

• Upsell  - Relevant products at a higher price range and within the same category.
• Cross-Sell: Relevant complementary products
• Most Viewed: Most viewed products for each category.
• Most Purchased: Most purchased products for each product category.
• Previously Viewed: Previously viewed products by current customer.
• Complete The Look: Accompanying products from the same collection.
• Current Cart Items: Products placed in the cart by current customer.
• Last Abandoned Cart Items: Products placed in the cart in previous session.

Filters

Additional product filters can be applied within each algorithm to include or exclude specific products:

• Products: These filters allow you to limit the pool of products available for recommendations based on specific products.
• Brands: These filters allow you to limit the pool of products available for recommendations based on specific product brands.
• Categories: These filters allow you to limit the pool of products available for recommendations based on specific product categories.
• Tags: These filters allow you to limit the pool of products available for recommendations based on specific product tags.
• Force Include Products: This filter is not so much a filter but more so a product push.  For instance, while the algorithm might decide to show you 10 products based on all the logic that goes into it, you may want to always still show a particular product in front of them.  A common use is to push a free tester item or gift card as the first couple products and then the relevant dynamic items after that.
• Inventory: These filters allow you to limit the pool of products available for recommendations based on inventory availability.
• Price: These filters allow you to limit the pool of products available for recommendations based on price.
• Discount Products: These filters allow you to limit the pool of products available for recommendations based on the product being on sale.
• Margin: These filters allow you to limit the pool of products available for recommendations based on product margin.
• Date Added: These filters allow you to limit the pool of products available for recommendations based on product added or date available dates.

Common Use Cases for Algorithms

• Upsell: You would want to generally show an upsell recommendation to visitors on a PDP page.  The idea is that they have shown some interest in this product but you may be able to upsell them on a similar product that costs more.  Likewise, you would NOT want to use an upsell on a cart page because the visit has already decided they want to buy the item in that situation and you don't want to make them reconsider.  In that case the Cross-sell is the better choice.
• Cross-Sell: Cross-sell recommendations are far better suited to show based on items in your cart, so I always recommend this algorithm for use on cart pages since the visitor will have items in their cart they are already partially committed to buying and you want to recommend complimentary items to go with them so they add more to their cart.
• Most Viewed: In the scenario when you don't yet have enough purchase data, this data tends to generate much much faster, otherwise the following most purchased algorithm is usually the better choice for these types of recommendations. 
• Most Purchased: If you are trying to move your best sellers this is the algorithm you want.
• Previously Viewed: This one is useful if you wantto present to your visitor items they have already viewed only.
• Complete The Look: This algorithm is great at suggesting one or more items from specific filtered categories (you will want to select the categories from the filters to create groups of categories that go together).  You can get away with not specifying the category filters, although the results may not be quite what you are after as it will try to determine on it's own which categories make sense to group by.
• Current Cart Items: This will list out your visitor's current carted items. The workflow with this design in it should have enrollment triggers set to only allow enrollment if cart total is > 0 in current session to avoid enrolling people who won't have this data on their profile.
• Last Abandoned Cart Items: This will list out your visitor's previous session abandoned carted items. The workflow with this design in it should have enrollment triggers set to only allow enrollment if cart total is > 0 in previous session to avoid enrolling people who won't have this data on their profile.

One more important callout feature you generally want to have on for Upsell and Cross-Sell is the Fill available space with best selling products, this feature is important to fill in more product recommendations when your catalog/order data isn't sufficient enough to make enough recommendations based on the products Ids sent to our api.