CartAmplify’s recommendation engine uses machine learning and behavior analytics to provide personalized product suggestions. It considers factors such as user browsing history, purchase behavior, and popular trends to display the most relevant items.

How to Create Recommendations #

1. Navigate to the Recommendations tab in the left menu of the dashboard and click New Recommendation #

2. Customize Settings #

• Add Recommendation name
• Select store language for which you make recommendations
• Adjust widget placement: Homepage/Cart Page/Product Page/Category or Collection Page
• Select Recommendation type:
  
  Similar Products: Based on similar attributes and categories.
  Frequently Bought Together: Analyzes past purchase combinations.
  Other you may like: Displays popular products based on user activity.
  Recommended for you: Tailored suggestions for each visitor.
  Buy it Again: personalized recommendations based on past purchases
  

3. Select Optimization type: #

✦ Click-Through Rate (CTR) Optimization: Focuses on maximizing the number of users who click on a recommendation or search result. Best for:

• Increasing engagement with recommended products.
• Improving discovery of new or trending items.
• Optimizing homepage and category page recommendations.

Example: A recommendation engine prioritizing items that historically receive the most clicks.

✦ Conversion Rate (CVR) Optimization: Focuses on maximizing the percentage of users who make a purchase after clicking on a recommendation. Best for:

• Increasing engagement with recommended products.
• Improving discovery of new or trending items.
• Optimizing homepage and category page recommendations.

Example: A recommendation system favoring products that have a higher  likelihood of being purchased after a user clicks.  

✦ Revenue Per Session (RPS) Optimization: Focuses on maximizing the total revenue generated per user session. Best for:

• Maximizing revenue from each visitor.
• Balancing click-through and conversion optimization.
• Personalized product bundling and upselling.

Example: A system that promotes higher-margin products or bundles that increase the average order value.

5. Set Constraints:   #

• Maxim Products to show in recommendation
• Set Price Ranking: the way products are sorted based on their price in search results or recommendation engines.
• Select diversity type: ensures that users see a variety of products rather than just the most similar ones. This prevents over-personalization and improves discovery. There are two main types:
  
  ✦ Rule-Based Diversity: Uses predefined rules to ensure diversity in recommendations.  How it works: Set rules to limit similar products (e.g., “show at most 2 items from the same brand”) and ensure recommendations come from different categories (e.g., “include at least 1 accessory with main products”).
  Best for: Businesses that want control over recommendations, avoiding overexposure of certain products.
  Example: If a user browses smartphones, the recommendation engine shows 2 phones, 1 case, and 1 wireless charger instead of just phones.
  
  ✦ Data-Driven Diversity: Uses machine learning and customer behavior to optimize diversity dynamically. How it works: Analyzes user preferences, browsing history, and past purchases, Introduces variety based on behavioral patterns (e.g., users who bought laptops often consider accessories), Balances relevance and diversity automatically.
  Best for: Large-scale e-commerce with many SKUs and personalized recommendations that still maintain variety.
  Example: If a user often buys sports shoes, the system dynamically suggests a mix of different shoe brands and complementary items like running gear.
  
  
• Once configured, enable the recommendation engine and track performance.