All recommendations are not the same in nature, and businesses need to understand the different types of recommendations to provide the best possible experience. By understanding the context and strengths of each recommendation type, businesses can deliver the right personalization experience that best suits their needs.
Types of Recommendations
MoEngage has two categories of recommendations.
Basic Recommendations
These recommendations are based on rules and offer the following models for setting up recommendations. You can learn more about the basic recommendations in this article.
- Item Attributes
- User actions
Advanced Recommendations
These recommendations utilize sophisticated machine learning and deep learning algorithms, incorporating the following distinct models. You can learn more about the advanced recommendations in this article.
- Recommended items (user personalized items)
- Similar items
- Frequently viewed together items
- Frequently bought together items
Getting Started
Delivering item recommendation experience to consumers is a set of processes, which can be classified into the following categories.
The setup consists of the following steps:
-
Recommendation settings: These configurations enable the recommendations to comprehend your data points, user behavior, and generate personalized results based on the specific requirements of your use case.
- Product catalogs: A catalog contains a list of all business offerings and respective information. Campaigns are personalized with item attributes of recommended items from the catalog.
- Map the User actions: The actions performed by users on the app or website need to be mapped to standard MoEngage events for MoEngage to understand the items interacted while performing those custom actions. For example, when a user adds an item to the shopping cart, MoEngage needs to know which event contains this information to track that event and store the item attributes associated with the event.
- Create Recommendation: The recommendations are defined by the user to select items from a product catalog based on a user action or catalog attribute. After you create the recommendations, you can use them in different types of campaigns.
- Campaign personalization: Regardless of the type of recommendation, the process of personalizing it in the campaign content remains the same. You can personalize content with one or more recommendations in the same campaign under content steps. You can also create additional content rules with flexible Jinja programs.