Personalized homepage with section recommendation

With Section recommendations, you can create a custom page with recommended products divided into sections. These sections display items based on a chosen item attribute and additionally they are automatically adjusted to match the customers’ preference. This type of recommendation lets the customer have a product overview, with the items relevant for them, chosen by the AI-based models.

In this use case, you can find an instruction on creating a recommendation for a retail website that consists of 5 rows of categories with 4 items for each category type. The categories of the products are represented in the following way: “Clothing > t-shirts and tops > t-shirts > long sleeve”. However, we would like to present the section with a smaller granularity of categories, for example, “Clothing > t-shirts and tops”.

Prerequisites

• The items feed must be provided.
• You must configure the AI engine:
  • Enable the Section page model.
  • In the Attributes for distinct filters section, you must select the attributes which will be available for use in the Item attribute field when creating a section page recommendation. In the Item attribute field you will choose the item feature (for example, brand) based on which the items will be selected for a slot in the recommendation. If this step is skipped, the field will remain empty, making it impossible to complete the recommendation.

Optionally, you can provide metadata catalog.

Create a recommendation

1. Go to Communication > Recommendations > Add recommendation.
2. In the top left corner, enter the name of your recommendation.
3. In the Type & Items feed section, click Define.
4. From the Items feed dropdown menu, choose the provided feed.
5. Choose the Section recommendation type.
6. Under the section types, from the dropdown list, choose your Metadata catalog.
7. Click Apply.
8. In the Items section, click Define.
9. Click Add slot.
10. Set the Number of sections option to 5.
11. Set the Number of items per section option to 4.
12. From the Items attribute dropdown, choose the category attribute.
13. In the Category level input area that appears, define the category level as a numeric value, in our case 2.
    Tip: If your products categories have a X > Y > Z structure, level 0 will be X > Y > Z. Level 1 will be X > Y and so on. Here, you are defining how granular the category recommendations will be. For example, if you sell shoes, you will have a Outdoor > Sport > Running category and a Outdoor > Sport > Football category. If level 0 is provided, both categories can be recommended. If level 1 is provided, Outdoor > Sport category will be recommended to the customer.
14. Click Apply.
15. In the top right corner, click Save.

What’s next

You can display the recommendation on your home page, use dynamic content.

1. Go to Communication > Dynamic content > New dynamic content.
2. In the body of the dynamic content, use the recommendation insert.
   
   Click here to display the insert

   {% recommendations3 campaignId=aIEkwqXw0wLB %}
       {% for slot in slots_products3 %}
           {% for row in slot.rows %}
   
               <!-- Variable metadata which allows you to access the metadata catalog -->
               {{ row.metadata.itemId }}
               {{ row.metadata.imageLink }}
   
               <!-- Iterates through items in a given section -->
               {% for item in row.items %}
                   {{ item.itemId }}
                   {{ item.title }}
               {% endfor %}
   
           {% endfor %}
       {% endfor %}
   {% endrecommendations3 %}

   Note: Read more about how to use recommendation in inserts here.
3. Add CSS and/or HTML to the dynamic content.
4. Define the rest of the settings.

Check the use case set up on the Synerise Demo workspace

You can check the recommendations which let you implement described basic AI Search on your website.

If you don’t have access to the Synerise Demo workspace, please leave your contact details in this form, and our representative will contact you shortly.

Read more

• Recommendations