Creating Product Recommendation Campaigns

You are here:
< All Topics

CanisHub has around 10+ recommendations models that can be used to personalize different page types (Homepage, PDP, PLP or Cart Page)

Product recommendations is part of personalization strategy that starts with page type. You need to first decide which page you want to personalize. Based on page type, product recommendation model and strategies will change.

Let us see what are the recommendation models that you can implement of different page types.

Homepage Personalization

  • User Interest (Recommended):
    • Placement: Use this recommendation model in the first half of the page.
    • Priority: High
    • Benefit: This model is very effective to target repeat visitors. It will only populate if the system is able to detect the interest of a visitor and will show products matching it, making their shopping journey very easy.
  • Recently Viewed:
    • Placement: Use this recommendation model in the second half bottom of the page.
    • Priority: Low
    • Benefit: This model is very effective when you have more than 500 products. It is commonly used in fashion industry where customers look into multiple products before making a decision to buy a single product. CanisHub, stores the each visitors browsing history, so if the customer comes back, we will be able to populate this model very well.
  • Trending by Order/Bestseller:
    • Placement: Use this recommendation model in the first half but below the user interest model
    • Priority: Medium
    • Benefit: This model is effective to combat cold start problem of recommendation, where you show the top trending or bestselling products of your store to new visitors.
  • Trending by First Order:
    • Placement: Use this recommendation model in the first half of the page
    • Priority: High
    • Benefit: Trending is a second model that helps you solve cold start on homepage. Instead of showing what is bestseller you can also show what is best selling among new visitors. This way, you can show products that new visitors are confident of buying for the first time. There is an estimate overlap of 40-50% with the best seller, however, this is a much more sophisticated model than bestseller.

Product Page (PDP) Personalization

  • Personalized Related Products
    • Placement: Use this recommendation just below the frequently bought together to upsell related or relevant products
    • Priority: High
    • Benefit: Personalized related products is a proprietary model by CanisHub. It allows you to show related or similar product that the visitor is viewing but sorted in the order based on that visitor browsing history. It considers the interest of the visitor and sort the entire similar product catalog for better and relevant targeting.
  • Frequently Bought Together
    • Placement: Use this recommendation model on the page just below the main product
    • Priority: High
    • Benefit: People also bought or Frequently bought together helps you improve your cart value. Use this model just below the main product page to show what other products people generally buy with the main product. Example: With Mobile Phone, people might also buy phone cover and screenguard. The system based on the order history finds the products that usually sells with the viewed product.
  • Similar Products by Categories
    • Placement: Use this recommendation model as an alternative to Personalized Related Products
    • Priority: Low
    • Benefit: If you want to arrange the order of sorting based on your categories on your site than use this model. It will allow you set the priority between product type, collection and vendor. It benefits you if you want personalization to behave in a certain way and want your recommendations to consist of a specific combination.

Cart Page Personalization

  • Customer Also Bought
    • Placement: Use this recommendation model just below the cart table
    • Priority: High
    • Benefit: Customer also bought helps you to achieve higher AOV. It basically give options to visitors to choose from what people mostly buy after buying the product that is already in the cart. It should be avoided on stores where multiple purchases or per order content is low.
  • Frequently Bought Together
    • Placement: Use this recommendation model in the below or as an alternative to customer also bought
    • Priority: Medium
    • Benefit: This model helps you to show products that are bought with the product already added to cart. Mostly, this model is used on product detail page, however, for many stores it also works wonder on cart page. You can try this option on cart page too.

Product Listing (Collection) Page (PLP) Personalization

  • Trending in Collection / Category
    • Placement: Use this recommendation model above the category or collection listing page
    • Priority: High
    • Benefits: This model works best when the collections have more than 100 products. This will show the best seller for each category or collection that the visitor is looking into. This increases the CTR by at least 2.3x times from collection page to product page.
  • Recently Viewed Products
    • Placement: Use this recommendation model in the sliding drawer widget than as an embedded solution
    • Priority: Medium
    • Benefit: This model will help you navigate the user to recently viewed product easily on the collection page. Mostly, customers on collection page are trying to find the relevant products. Half of the time, they want to go back to products they have already seen. With sliding drawer, this can be easily achieved without distracting the user from finding the product from the listing page.

Final Tips: No one size fits all. Similarly, no one recommendation model fits all store. Therefore, keep trying with multiple recommendation model every 30 days to see the impact and measure the CTR. In case of any further consultation, please reach out to us at contact@canishub.com and we will help you chalk out a recommendation plan for your store for free.

Table of Contents