Overview

Personalize product discovery with MoEngage Recommendations. Learn how the engine works, when to use Basic versus Advanced models, and how to choose the right one for your goal.

Recommendations let you show each customer the products, content, or offers they are most likely to act on — automatically, and at scale. Instead of promoting the same items to everyone, you can surface the right item at the right moment, whether that is a personalized pick, a relevant alternative, or a timely price-drop alert.

The Recommendations engine covers the full range, from promoting the same curated items to everyone, to identifying relevant items tailored to each individual, inspired by similar users’ behavior or item similarities.

How Recommendations Work

Setting up a recommendation follows four steps. You complete the first once, then reuse it across every recommendation you build.

1. Prepare your data

   Connect an item catalog (your products, content, or listings) and map your user-action events (views, cart adds, wishlist saves, purchases) to MoEngage’s standard events. See Prerequisites.

2. Choose a recommendation type

   Decide between Basic (rules you define) and Advanced (machine-learning models). See Choosing a Model below.

3. Create

   Configure the recommendation with the filters and sorting you need.

4. Test

   Preview the results for a real user, a random sample, or a specific item before you launch. See Test a Recommendation.

5. Use it in a campaign

   Insert the recommendation into a Push, Email, In-App, or other campaign to deliver personalized content to your users.

Basic vs Advanced Recommendations

MoEngage offers two families of recommendations. Most accounts use both, depending on the campaign.

	Basic	Advanced
How it decides	Rules you define	Machine-learning models (Merlin AI)
Best when	You know exactly what to promote	You want personalization to adapt automatically
Effort	You write the filters and sorting	The model learns from behavior
Data readiness	Catalog (and user actions for some models)	Catalog plus enough mapped user-action history to train
Models	Item Attributes, User Actions, Catalog Alerts	Recommended Items, Similar Items, Frequently Viewed Together, Frequently Bought Together, Trending Items, Popular Items

Choosing a Model

Start from your goal and work back to the model.

I want to…	Use	Family
Promote specific products — new arrivals, a brand, seasonal picks	Item Attributes	Basic
Recall items a user showed intent on — cart, wishlist, recently viewed	User Actions	Basic
Alert users to price drops, back-in-stock, or fast-selling items	Catalog Alerts	Basic
Give each user personalized “recommended for you” picks	Recommended Items	Advanced
Show alternatives to an item a user is viewing	Similar Items	Advanced
Cross-sell items often viewed together	Frequently Viewed Together	Advanced
Bundle items often bought together	Frequently Bought Together	Advanced
Surface what’s trending overall, or within a category or region	Trending Items (and Popular Items as a default fallback)	Advanced

Benefits

• Higher engagement and conversion — relevant suggestions keep users active and increase the likelihood of purchase.
• Better cross-sell and upsell — contextual bundles and add-ons lift average order value.
• Reduced churn — the right item at the right touchpoint keeps users coming back.
• Stronger loyalty — personalized experiences make customers feel understood.

Recommendations Across Industries

Recommendations apply to almost any catalog-driven business. A few representative use cases:

Industry	Example use case
Retail	Recommend complementary products, add-ons, or upgrades; re-engage on abandoned carts and wishlists.
Banking and Finance	Suggest loans, insurance, or investments based on a customer’s profile; surface relevant card offers.
Classifieds and Real Estate	Recommend jobs matched to a seeker’s skills, or listings near the user’s location.
Ed-Tech	Recommend courses or certifications based on a learner’s goals and interests.
Entertainment and Media	Recommend shows, articles, or videos based on viewing history and similar users.
Travel	Recommend packages or destinations based on traveler preferences and past trips.
Gaming	Suggest new titles based on genres and games a player has already played.

Next Steps

Before building your first recommendation, confirm your data is ready:

• Connect an Item Catalog.
• Map User Action Events.

Once your data is set up, choose between Basic or Advanced recommendations. Then, explore how to use recommendations in Push, Email, and On-Site Messaging campaigns.