Creating recommendations

Once the model training is completed, you can create a recommendation. The recommendations you create and activate will not be visible until you indicate how and where the recommendations are to be displayed.

Distributing recommendations


You can use the ID of the recommendation and put it in the form of an insert in other types of communication such as:

  • dynamic content - this way you can show the recommendations on your website,
  • email - this way you can send out recommended items through emails,
  • web push - this way you can send out recommended items in notifications on your website.
  • mobile application - you can use documents to build your own mobile app and show the recommended items.

You may want to read


Requirements


  • You must configure AI engine.
  • You must be granted a set of user permissions that allow you to access the Communication module, to create, to edit and execute recommendations.

Creating recommendation


  1. Go to Communication icon Communication > Recommendations v2 > Add recommendation.
  2. Enter the name of the recommendation (it is only visible on the list of recommendation).

Select feed and recommendation type


Start with indicating the feed from which items will be sourced to recommendations and the type of recommendations to be displayed.

Type & Source section
The Type & Source section
  1. In the Type & source section, click Define.
  2. From the Catalog dropdown list, select a trained product feed.
  3. Below, select the recommendation type. If the recommendations are greyed out, it means the AI engine is not trained yet.
    Note: To get information about recommendation types, click here.
  4. Confirm the settings by clicking Apply.

Configure item settings


The next step is to define the number of items to be displayed in the recommendation and narrowing down the scope of displayed items using filters. Because each recommendation is different for every user, you can’t indicate exactly the items to be shown in the recommendations. However, you can get the preview of the recommendation for any customer.

A blank form for configuring recommendation campaigns
A blank form for configuring recommendation

Define the number of displayed items

  1. In the Items section, click Define.
  2. In the Minimum field, enter the lowest number of items displayed in the recommendation frame.
  3. In the Maximum field, enter the highest number of items displayed in the recommendation frame.
  4. If you don’t want to display the items a customer has already purchased, switch the Exclude already bought products toggle on.

Select conditions of displaying items

  1. To narrow down the scope of the items displayed in the recommendation frame, click Define filters.
    1. Select one of the filter creators:
      • visual builder
      • IQL query wizard
        Elastic filters explained

        Apart from selecting the items to be displayed in the recommendation frame, the elastic filter supplements the recommendation frame if it’s not entirely filled up with the items. For example, if you select to display up to 10 items, and you have only 5 items that meet the conditions to be displayed in the recommendation frame, then the empty slots in the frame are replaced with personalized items.

        Static filters explained

        When you use a static filter, it shows the fixed number of items that match the conditions of the filter. If there is not enough items to fill in the recommendation frame (the number you entered in the step 3), the recommendation frame is not displayed at all.

WARNING: We don’t recommend using Elastic filters or Static filters at the same time.

Increase the recommendation variety

  1. To manage the variety of items in the recommendation frame, in the Distinct filter section, click Define filters.

    Important: This filter can only use attributes defined in configuring the recommendation engine.

    1. From the list, select an attribute.
    2. In the Max no. of occurrences, define the number of items with the same value of the attribute (for example, a brand) that can be displayed in the recommendation frame.
      Result:
    A blank form for configuring recommendation campaigns
    Attributes defined in the distinct filters

    Explanation: In the example in the screen, a recommendation frame displays items of various brands (a brand cannot be repeated more than once), various colors (each item in the recommendation frame is in different color), and only 3 items in the frame can be taken from the lowest category level.

    Distinct filter explained

    Distinct filters allow you to increase the variety of items displayed in the recommendation frame. You can define exactly the allowed number of items that share the same attribute value to be shown, for example, a number of items that have the same brand, color, shape, category, and so on.

    • For all recommendation types except for Last seen, the engine considers up to 1000 items with the highest score that match the recommendation type. For example, if you selected Cross-sell type, the engine analyzes up to 1000 items that match the cross-sell recommendation type, and then selects the number of items you chose to display in the recommendation frame.
    • For the Last seen recommendation type, the engine considers the last 100 page visit events. Based on the data from these events, the engine selects the number of items you chose to display in the recommendation frame.

Define the boosting factors

You can influence the arrangement of items in the recommendation, which is by default arranged according to a score given by the model (the type of recommendation selected). You can boost the items in the recommendation by means of two factors:

  • Boosting - It allows you to select the top items according to the chosen metric out of the items which meet the conditions defined in the elastic/static/distinct filters.
  • Sorting - It allows you to define the order of the items in a particular order.
  1. In the Additional settings section, select the option:

    • Boosting the score of your items:
      1. Choose one of 5 metrics.
      2. Determine how much you want the metric to influence the score of the items by using a slider.
    • Sorting the order of items according to a metric:
      1. Choose one of 5 metrics.
      2. Determine how much you want the metric to influence the score of the items by using a slider.
    Metrics explained

    • Sold items count in the last 30 days - The system favors the items that were sold in the highest quantity in the last 30 days.
    • Sold items value in the last 30 days - The system favors the items which summed value, that is how many were sold but also how much they cost, is the highest in the last 30 days.
    • Page visit count in the last 30 days - The system favors your popular items, based on the amount of times the items’ page was visited in the last 30 days.
    • Conversion percent in the last 30 days - The system favors those items which conversion rate in the last 30 days is the highest.
    • Conversion percent after clicking in the recommendation in the last 30 days - The system favors those items which conversion rate after clicking in the recommendation is the highest in the last 30 days.
    • Sold items count in the last day - The system favors the items which were sold most frequently the day before.
    • Sold items count from the same weekday last week - The system favors those items which were sold most frequently during the same weekday, a week before.

  2. Confirm the settings by clicking Apply.

  3. To save the recommendation:

    • as a draft, click Finish later
    • and activate the recommendation, click Save.
      Once you activate the recommendation, you can use it as described in the Distributing recommendations section.
😕

We are sorry to hear that

Thank you for helping improve out documentation. If you need help or have any questions, please consider contacting support.

😉

Awesome!

Thank you for helping improve out documentation. If you need help or have any questions, please consider contacting support.