Skip to main content

Strategy

The Personalization API of OMNICOMMERCE allows e-commerce sites to implement personalized recommendations by analyzing the behavior data of each e-commerce user. It can provide the optimal recommendation results based on the location of the recommendation slot on the e-commerce site, the target metrics of the recommendation slot, and the application scenarios. When requesting recommendations, you can select the type of recommendation model that best suits the characteristics and objectives of the e-commerce site. The type of recommendation model can be choosed through the strategy option entered during the recommendation request.

Types of Strategy​

The list of recommendation model types currently available and the strategies that can be selected in the personalization API are as follows:

This strategy offers products that a user is likely to purchase or click on, based on a variety of user information such as product views, purchases, searches, and recent actions, as well as detailed information about products within the e-commerce platform. The recommendation can function both with and without a specified product, according to the recommendation scenario, and it can operate even if there is currently no specified product. Moreover, depending on the recommendation scenario, the results can be filtered by specific categories, brands, price ranges, etc.

Applicable API

Supported Page Types

  • Main home page
  • Category page
  • Product detail page

similar-items​

This recommendation strategy, based on images, identifies and recommends other products that are similar in attributes and appearance to the product currently being considered. To increase the target metric (click-through rate) of this recommendation, it selects the most optimized products for each individual from among the similar ones using AI. The similar-items strategy requires the input of a product ID.

Applicable API

Supported Page Types

  • Main home page
  • Category page
  • Product detail page
  • Search result page

often-viewed-together​

This strategy recommends products that other customers who viewed a specific product have also viewed frequently. The recommendation results can be filtered by product attributes, such as brand and category.

Applicable API

Supported Page Types

  • Product detail page
  • Shopping cart page

frequently-bought-together​

Products that are frequently purchased together with a specific product will be recommended. These recommended products will be personalized for each user. The recommendation results can be filtered based on product attributes (e.g., brand, category, etc.).

Applicable API

Supported Page Types

  • Product detail page
  • Shopping cart page

Using user behavior data within e-commerce (such as purchases, cart additions, wishlist additions, views, etc.), popular products will be recommended. Among these popular products, those that are personalized for each user will be recommended.

Applicable API

Supported Page Types

  • Main home page
  • Category page
  • Product detail page