Default Setup

Homepage

Personalized recommendations are displayed in a carousel layout on the homepage at the very top of the page (often after the primary banner).
Last seen recommended products (products a user has seen before, including previous visits) are displayed in a carousel at the bottom of the page.

Product Detail

Product alternatives are shown after the product in detail in a carousel.
Personalized recommendations that other users like are located at the bottom of the page in a grid layout.
The Last seen products are displayed in a carousel at the bottom of the page.

Pre-cart

Personalized cross-sell and upsell recommendations are displayed in a carousel. This is often realized by a modal dialog or a separate page.

Cart

Cross-sell recommendations in a carousel.

Success Page

Personalized recommendations are shown in a grid.

Error Page

Personalized recommendations are shown in a grid.

Please see how to affect placement and exact position of Recommendation Widgets through Recommendation Anchors

Recommendation Layouts

There are three types of recommendation widget layouts. Layouts are defined for several viewports. The following example shows Desktop and Mobile viewports.

Carousel

Desktop

Mobile

Carousel List

Displays multiple product carousels in a vertical list. Each carousel contains items belonging to a specific category.

Desktop

Mobile

Grid

The grid can be configured with a specific number of columns and rows per viewport. A show more button is also present to display more recommendations.

Mobile

Desktop

Configuration

Design

Widget design can be configured to be close to the original website look and feel. The following can be configured:

Recommendation title
Product tile
Fonts
Primary/secondary colour
Paddings

Scenarios

Scenarios are trained recommendation models with the given configuration.

scenario	description
last_seen	Show recently seen items; also, items from previous visits are included.
personalised_similar_items	Content-based recommendation showing item alternatives. Signal/features can be customized for the given website.
recommended_for_you	Personalized recommendations based on other visitors' behavior. The principle of the collaborative filtering method is based on neural networks.
quick_picks	Showing cross-sell and upsell products based on history or recent behavior.
bundles	Looking for product bundles based on real visitors' behavior and that are reasonable for business.