User Documentation
API Clients & Integrations
Scenario Recipes
Introduction
Recombee is a Recommender as a Service with easy integration and powerful Admin UI.
The Recombee recommendation engine can be applied to any domain that has a catalog of items and is interacted by a large number of users. Applicable to web and mobile apps, Recombee improves user experience by showing the most relevant content for individual users.
How it works
?
?
?
?
?
?
?
Item Properties describe the items that you want to recommend.
Example of properties
- Categories
- Text description
- Images
- Labels
- Genres
- Expiration date
- Destination
- Geo location
- Age restriction
- and more
User Properties (optional) describe your users.
Example of properties
- Gender
- Language
- Age
- and more
Interactions between Users and Items are the most important data for the recommender system.
Supported Interactions
- Views
- Purchases
- Cart Additions
- Bookmark
- Ratings (Likes)
- View Portions
Recommendation Type
The type can be for example recommendation for a particular user, recommendation of related items or personalized full text.Recommendation Logic
Desired behavior of the recommendation models.Business Rules (Filters & Boosters)
Influence the outcome of the recommendations by applying custom Filters and Boosters.Scenarios
Each Scenario defines a particular application of the recommendations at your website, mobile app or in emailing.
Item properties describe the products or content that you want to recommend to the users. You can define various types of properties that are analyzed by the content-based recommendation models. In addition to that, you can also use the data in your Filter or Booster rules. Data can be delivered using a Catalog Feed or the Recombee API.
User properties (optional) contain data about your users. Data can be delivered using the Recombee API.
Interactions between Users and Items are the most important data for the recommender system, as the collaborative filtering models learn various relations from them. You can send them using the No-Code Widget, Client-Side API, Server-Side API, or Segment Integration.
Scenarios define a particular application of a recommendation at your website, mobile app, or email. They may be homepage-featured, watch-next, read-next, similar-products, emailing-after-purchase, or others. Each Scenario is of a particular recommendation type. The type can be for example: a recommendation for a particular user, recommendation of related items, or a personalized full-text search. You can also apply various settings to each Scenario.
Recommendation Logic specifies the desired behavior of the recommendation model. We have prepared different Logics for various use-cases that you can choose from.
Business Rules are used to filter or boost items based on their properties. We provide a library with frequently used Rules that you can apply out-of-the-box - and start using immediately. You can also specify your own Rules using our intuitive Recombee Query Language (ReQL).
Constraints allow you to configure how items with shared conditions may repeat within the recommendation results.
You can integrate recommendations using our No-Code widget, our Widget SDKs, or API SDKs in the programming language of your choice.
Where to use
Recombee can be utilized in various domains:
- Video
- Articles, News, Media
- E-commerce
- Music, Podcasts
- Marketplaces
- Real estate
- Job boards, HR, Networking
- Travel, Trips
- and many more
Thanks to our simple integration process, you will be able to provide content, product, or specialized personalization on your website, web app, mobile app, or emailing in a short time.
Personalized Search, Homepage, Product/Content detail page, Push notifications, recommended content in Read next or Watch next - those are only a few examples of how you can apply Recombee personalization to improve UX and loyalty of your users.
- Personalized search
- Featured on homepage
- Newsletter
- Read next
- Watch next
- Similar products
- Similar content
- Other also bought
- Push notifications
- etc.
