Recommendation types define how products are selected and ranked in a Recommendation List. Each type is designed for a specific purpose, such as showing bestsellers, personalized products, or complementary items, depending on where and how it is used.
Recommendation types power the Recommendation Lists that you place on your pages in the Experience area of Elevate.
How recommendation types work
Elevate uses a combination of signals and data to automatically select and rank products for each recommendation type.
- Customer behavior (clicks, purchases, browsing)
- Product data and relationships
- Real-time signals (trends, stock, newness)
Recommendation types by category
To make it easier to choose the right type, recommendation types can be grouped based on how they work:
AI-powered (behavioral recommendations)
These recommendation types are driven by real-time behavior and product performance across your site. They are designed to automatically optimize for relevance and conversion.
Includes:
- Top products
- Alternatives
- Upsell
- Cart
- Add-to-cart recommendations
Personalized recommendations
These are tailored to the individual shopper, based on their own behavior and interactions. They help shoppers continue their journey and revisit products.
Includes:
- Personal
- Recently viewed
- Favorites
Pre-set (editorial) recommendations
These are based on product relationships or assortment logic, rather than behavior. They are typically used to create curated or inspirational experiences.
Includes:
- Style with
- More from series
- Newest products
How recommendation types behave
Use the table below to understand the behavior of each recommendation type before adding it to a Recommendation List. Each column describes a specific characteristic that may affect how results appear on your site.
| Recommendation type | Includes out-of-stock products | Affected by missing image | Affected by Exposure strategy | Affected by Boost / Bury | Fill up slots | Soft deduplication |
|---|---|---|---|---|---|---|
| Top products | Yes* | Yes* | Yes | Yes | Yes | Yes |
| Personal | No | Yes | Yes | Partially | Yes | Yes |
| Alternatives | No | Yes | Yes | Partially | No | Yes |
| Upsell | No | Yes | No | Partially | Yes** | Yes |
| Cart | No | Yes | Yes | Partially | Yes | Yes |
| Style with | No | Yes | No | No | No | No |
| Favorites | No | Yes | Yes | Partially | Yes | Yes |
| Newest products | Yes* | Yes* | No | No | Yes | No |
| More from series | Yes* | Yes* | Yes | Partially | No | Yes |
| Recently viewed | Yes | No | No | No | No | No |
| Add-to-cart recs. | No | Yes | No | Partially | No | Yes |
* Out-of-stock and missing-image products are included but demoted according to the Out of Stock and Bury Missing Image settings in the Elevate application.
** Upsell can return an empty result if no in-stock products exist within the same gender and age groups, although this is very rare.
Partially under Boost / Bury means the algorithm is influenced, but other strong signals, such as personal affinity, may limit the visible effect. Both Boost / Bury and Exposure strategy affect the same algorithms because both work by adjusting a product's entity score.
Fill up slots means the recommendation type will almost always have products to display within any applied filters.
Soft deduplication means Elevate attempts to reduce repeated products across recommendation lists where possible.
Recommendation types explained
Top products
A generic recommendation type that shows high-performing products based on:
- Aggregated visitor behavior
- Sales and interaction data
- Product lifecycle signals (e.g. newness, stock)
- Selected exposure strategy
This is the most flexible recommendation type and can be used in many contexts. It is also the default fallback, but should ideally be used when no more specific type is better suited.
Personal
Generates personalized recommendations based on the current visitor's behavior, such as:
- Recent views
- Clicks on Add to cart
- Purchases
Used when you want to tailor the experience to each individual shopper.
Recently viewed
Displays products the visitor has recently interacted with.
Helps users easily return to products they have already explored. Can also be shown in autocomplete for quick access.
Alternatives
Shows products that are similar to the one currently being viewed.
Used on product pages to help shoppers compare options and reduce bounce.
Upsell
Displays products that are often bought or considered together with the current product.
Used to increase basket size by suggesting relevant additions.
Style with
Shows products that complement the current product, based on a predefined assortment.
Used to create inspiration and curated combinations.
More from series
Displays products from the same series or collection.
Helps shoppers explore related items within a defined group.
Add-to-cart recommendations
Triggered when a product is added to the cart.
Shows products that are commonly bought together with the selected item, optimized for quick decisions in popups or drawers.
Cart
Displays recommendations based on the current contents of the cart.
Used on the cart page to suggest additional relevant products.
Favorites
Shows recommendations based on products the shopper has marked as favorites.
Useful for inspiring users who are revisiting saved items.
Newest products
Highlights recently added products based on release date.
Commonly used on landing pages to showcase new arrivals.
Read more
Best practices for product recommendations - Choosing the right recommendation type often depends on where it is used in the customer journey, read on to see what works best for different page types.