Skip to content

Recommendations Modules

Recommendations modules provide different types of product recommendations including similar items, frequently bought together, "complete the look", and personalized picks.

Overview

Recommendations modules help users discover products through various recommendation strategies, each optimized for different use cases. The Marqo pixel automatically collects user interaction data (product views, purchases, add-to-cart events) to power these recommendations, so they work automatically without requiring any manual data collection.

Module Types

Similar Items

Recommend products similar to a viewed product:

{
  "module_type": "similar_items",
  "seed_product": "prod_123",
  "algorithm": "content_based",
  "max_results": 10,
  "diversity": "medium"
}

Frequently Bought Together

Recommend products often purchased together:

{
  "module_type": "frequently_bought_together",
  "seed_product": "prod_123",
  "algorithm": "collaborative_filtering",
  "max_results": 5
}

Complete the Look

Recommend complementary products:

{
  "module_type": "complete_the_look",
  "seed_product": "prod_123",
  "complementary_types": ["accessories", "matching_items"],
  "max_results": 8
}

Personalized Picks

User-specific recommendations:

{
  "module_type": "personalized_picks",
  "user_id": "user_456",
  "algorithm": "hybrid",
  "max_results": 12
}

Best Practices

  1. Choose right module: Select modules appropriate for context
  2. Test performance: A/B test different modules
  3. Monitor metrics: Track click-through and conversion rates
  4. Maintain diversity: Ensure recommendations aren't too narrow
  5. Update regularly: Keep recommendations fresh