Skip to main content

Product Recommendations

Product Recommendations help customers find products they are likely to purchase. They generally increase Average Order Value and Conversion Rate by guiding customers to products they are likely to purchase.

Content-Based / Individual Recommendations

Content-Based Recommendations are recommendations that are based on the content of the product. Individual recommendations combine content-based recommendations with individual customer activity to provide recommendations that are personalized to the customer.

Advantages

  • Beneficial recommendations can be provided even when customer activity is low.
  • Performs better than alternatives when the product catalog is vast (breadth).
  • Can be used to provide recommendations for new products.

Disadvantages

  • Generally lower impact than interaction-based recommendations on small product catalogs (but more effective otherwise).

Recently Viewed

Recently Viewed recommendations show products that the customer has recently viewed. This is a great way to help customers find products they have already shown interest in, but potentially forgotten about.

More Like This

More Like This recommendations show products that are similar to the product the customer is currently viewing. This is a great way to help customers find products that they are currently shopping for.

Like (Your) Recent Activity

Like Recent Activity recommendations show products that are similar to the products the customer has recently viewed. This is a combination of personalizing to the customer's activity and finding contextually similar products. This is generally the most effective content-based recommendation and is often used as a backup for "Products for you" recommendations (that primarily use interaction metrics).

Collaborative Recommendations

Collaborative Recommendations are recommendations that are based on the interactions of the customer, use collaborative filtering to find similar customers, then provide recommendations based on what the group has seen success with.

Advantages

  • Generally higher impact than content-based recommendations. Causation is more clear than correlation with regards to customer preferences.
  • Performs better than alternatives when the product catalog is small (Depth).

Disadvantages

  • Requires a large amount of customer activity to provide beneficial recommendations, without which, content-based recommendations are generally more effective.
  • More popular products are recommended more often, which can lead to a "rich get richer" effect, and penalizes newer products.

None have been created as of yet.