Select an algorithm for your product recommendation block

Choose from Kameleoon’s built-in recommendation algorithms—such as Popular, Similar, Trending, and Recently Viewed—to control which products appear in your blocks.

Kameleoon algorithms determine which products to recommend.

Here is the list of available algorithms:

Algorithm	Description
Popular products	The algorithm returns products sorted by sales rate.
Similar products	This algorithm returns products from the same category while considering product characteristics provided by the API. Kameleoon recommends products priced between 85% and 125% of the current product’s price. If there aren’t enough products to populate the entire slider, the price range expands to 50%–150%. You can apply additional filters based on other product characteristics. Only works on product detail pages.
Products usually purchased with this product	This algorithm returns products often bought with the current product. The calculation period is 1 month for traffic of more than 300,000 visits per month and 3 months for traffic below 300,000 visits per month. Only works on product detail pages.
Interesting for you	This machine learning algorithm uses the rule “visitors who bought the same products as you also bought these products”.
Trending products	This algorithm is similar to “Popular Products”, but Kameleoon considers the sales rate over a much shorter period (24 hours). Therefore, it represents the current trends and demand. You can edit the time frame (24 h) in the settings.
Recently viewed	This algorithm returns products the user recently viewed and hasn’t added to their cart. If the block appears on a product page, the algorithm excludes the product on display from the selection. The calculation period is 14 days.
Products usually purchased with the same shopping cart	This algorithm is similar to “Products usually purchased with this product”, but it applies to all the products in the cart. The calculation period is 1 month for traffic extending 300,000 visits per months and 3 months for traffic below 300,000 visitors per month.
Product usually purchased after this search query	This algorithm analyzes search queries and results, considering what products users purchased who made the same search. This type of block is usually displayed on the search results page, but separate it visually from the search result itself. The calculation period is 1 week for traffic of more than 500,000 visits per month and 2 weeks for traffic below 500,000 visits per month.
Regular purchase	This algorithm calculates an item’s frequency of purchase and forecasts the ideal time to display it to visitors.
New arrivals	This algorithm returns products marked as “new” in the product feed.
Exact products	This algorithm lets you select and display specific products.
Store recommendations	This algorithm displays products with the tag “accessories” in the feed. Only works on product detail pages.
Partially similar products	Similar to “Similar products”, this algorithm returns products with at least one common category/characteristic.
Previously added to cart	This algorithm returns products that the visitor previously added to their cart. The calculation period is 30 days.
Products usually viewed with this product	This algorithm returns products other users viewed after the current product. The calculation period is 1 day. Only works if there are items in the cart.
Next best product	This algorithm returns the best product to purchase after the current product.

You can select an algorithm, set a quantity of product returned by the algorithm, and add filters.

If you set a limit of ten products, but the limit of products in the block is six, the block will display six products.

To learn how to configure advanced rules for your recommendation block, read this article.

Algorithms at work

Below is a list of algorithms that only work under certain conditions. If the conditions aren’t met, the algorithm won’t work.

Only works on the product page

• Similar products algorithm (works when a product ID is available)
• Products usually purchased with this item algorithm
• Store recommendations algorithm

Only works if there are items in the cart

• Products usually viewed with this product algorithm.

FAQ

Can I import my own recommendation algorithm?

No. Kameleoon doesn’t support Bring Your Own Algorithm (BYOA) for product recommendations. Use one of Kameleoon’s built-in algorithms.

Is there a limit to the number of recommendations I can display?

Kameleoon doesn’t enforce a strict limit on the number of products a recommendation block returns. Requesting many products affects algorithm performance and increases response times, so display fewer than 50 products per block. To display larger product sets, use collections.

Can I use a custom field name (like shop_recommend) for the Store recommendations algorithm?

No. The Store recommendations algorithm requires the accessories tag in your data feed. It doesn’t recognize custom names, such as shop_recommend, for manual product relationships. The algorithm only works on product detail pages.

To link products manually, for example to display complementary items, map the related product IDs to the accessories tag.

The algorithm displays only the products linked through the accessories tag. To restrict these results further, apply standard filters to the recommendation block, for example to exclude out-of-stock items from the manual list.

How do I provide data for the Currently on wishlist algorithm?

The Currently on wishlist algorithm needs an explicit signal when a visitor adds an item to a wishlist, because the recommendation engine doesn’t detect wishlist interactions automatically. Call the trackAddToWishList() method from the Activation API to send the event data that populates the algorithm.