Personalized recommendations

This type of recommendation is based on the customers’ purchase preferences and behavioral profile. As a result of the analysis of the product feed, the visited pages and the history of transactions, the customers see product recommendations that match their individual preferences on the first pages, so they are not overwhelmed with the whole stock.


Requirements

  • Product feed delivered
  • Integration of events via AI tracking or API
  • Transactional data (minimum 3-month history) periodic transaction updates (one week history)
  • Online data stream including different channels if possible (recommended)

Configuration

Image presents main menu in Synerise app and how to proceed to the process of creating AI recommendation.

In order to start configuration of the recommendation type, you need to follow the path:

Campaigns (1) → Recommendations (2) → Create recommendation (3)

Once you choose Recommendations, you will see the list of all recommendations you have created so far. Each recommendation is marked with a tag that indicates the status of the created recommendation (either as draft or active). Recommendations can be edited or deleted at any time.

Click Create recommendation (3) to start the process of configuration.


Choose recommendation type

This is the place where you can choose the kind of recommendation type you want to create. In this case, choose personalized recommendations.

Image presents the editor of AI recommendation and the possible choices of AI recommendation types which take form of squares.

  1. Name - Click the pencil icon to assign a name to the recommendation you’re creating.
  2. Recommendation types - You can choose from six types of recommendations: cross sell, similar products, personalized recommendations, visual similarity, top products, last seen.
Note: Some recommendation types may not be available due to your account configuration. Inactive recommendations will be marked with a grey color.

Select number of products

Image presents the editor of AI recommendation, the section presents a button to switch on or off the product recommendations from the customer’s shopping cart and two text fields in which you can choose minimum and maximum number of products to display.

Once you’re done with the basic settings, choose how many products you want to display.

  1. Exclude product from the cart - Turn it on to avoid displaying the products which the customer has added to the cart.
  2. Min. & max. amount - Determine the range of the products you want to display. This way you can decide how much space you want to dedicate to the recommendations. Then, set the filters that will show a narrower range of products.

Image presents list of five categories of filters and circle buttons next to them that enable adjusting the settings of the filters.

  1. Brand filter - This filter lets you limit the range of products included in the recommendation to a particular brand. You can include (determine what kind of products which meet the filter conditions will be included in the recommendation) and exclude (determine what kind of products meet the filter conditions will be excluded from the recommendation) specific brands according to your preferences. After you make your selection, you will see a dropdown list with 3 possible options:

    • Skip - This allows you to intentionally omit this part.
    • Custom - Results will be filtered by brands specified in include/exclude field. If you choose this option, you will be able to choose brands which are available in the product feed from the dropdown list.
  2. Category filter - This filter lets you restrict the range of products in recommendations to a particular product category. You can include (determine what kind of products which meet the filter conditions will be included in the recommendation) and exclude (determine what kind of products meet the filter conditions will be excluded from the recommendation) specific categories according to your preferences. For the sake of the clarity, let’s consider the example: you want to recommend products available in the category in which TV set is placed. TV set is placed in the main category Electronics, and categories in general can take the structure of a tree. In this case, TV set is in Electronics > RTV > TV, so:

    • Skip - This allows you to intentionally omit this part.
    • Top - It concerns products set in the highest subcategory in the hierarchy (in this case it will be Electronics).
    • Parent - It concerns products set in the higher and the highest subcategory (in this case it will be Electronics and RTV).
    • Select - Results will be filtered by categories given by respective include/exclude field.
  3. Price filter - This filter lets you restrict the shown products in the recommendation to products within a specific prince range. You can include (determine what kind of products which meet the filter conditions will be included in the recommendation) and exclude (determine what kind of products meet the filter conditions will be excluded from the recommendation) specific product prices according to your preferences.

    • Skip - This allows you to intentionally omit this part.
    • Amount results - Determines the acceptable price difference between the product being viewed and the recommended one (e.g. from -25$ in relation to the price of the currently viewed product to 40$ ).
  4. Discount filter - This filter lets you restrict the shown products in the recommendation to the discounted products.

    • Select basis - Specify the basis on which the products will be filtered.

      • Percentage - If you choose this option, the discount value will be expressed as a percentage. As a result, you will choose products which are cheaper by a certain percentage.
      • Amount - If you choose this option, the discount value will be expressed as a number. As a result, you will choose products which are cheaper by a certain amount of money.
    • Select type - Additionally, specify the type of products which will be shown in recommendations.

      • Absolute - You can choose two options. The first option is to choose percentage as a basis and absolute as a type. In this case, the range of the shown products will include those which discount value varies, for example, from 0 to 10%. The second option is to choose amount as a basis and absolute as a type. In this case, the range of the shown products will include those which discount value varies, for example, from 0 do 100$.
  5. Gender filter - This filter lets you restrict the shown products in the recommendation to the products intended for a particular gender. You can include (determine what kind of products which meet the filter conditions will be included in the recommendation) and exclude (determine what kind of products meet the filter conditions will be excluded from the recommendation) a specific gender according to your preferences.

    • Male - Results will present the products specified as relevant for males.
    • Female - Results will present the products specified as relevant for females.
    • Unisex - Results will be filtered by products with unisex gender.
  6. Attribute filter - This filter lets you restrict the shown products in the recommendation to the products that contain specific attributes in the product feed. You can include (determine what kind of products which meet the filter conditions will be included in the recommendation) and exclude (determine what kind of products which meet the filter conditions will be excluded from the recommendation) specific brands according to your preferences. Additionally, you can choose range of attributes and indicate their values.

    • Current - Results will be filtered by the same attribute of a product given to recommendation, where name field is the name of custom attribute. If you choose this option, you don’t have to perform any additional actions.
    • Custom - Results will be filtered by attribute value given.

Adjust additional settings

The last step is to adjust the additional settings. This part will help you with choosing the optimal set of products and put them in the most favorable order. Image presents list of two pieces of settings, each consist of a dropdown list and a slider below.

  1. Boosting - Now your products are selected by means of the filters. You can increase the chances for conversion by selecting the best of the best products. You can choose from 5 types of metrics. Then use the slider to determine how active you want this option to be.

    • Sold items count in the last 30 days - The system will choose products which meet the conditions specified in the filters and it will consider the number of sold products in the last 30 days. Then the most recommended products will include those which were sold most frequently during the last 30 days.
    • Sold items value in the last 30 days - The system will choose products which meet the conditions specified in the filters and it will consider the value of the transactions in the last 30 days. Then the most recommended products will include those which were sold for the highest price during the last 30 days.
    • Page visit count in the last 30 days - Apart from the configuration of filters, the system will consider the number of visits on the website of the product. Then the most recommended products will include those which were viewed on the website the most frequently during the last 30 days.
    • Conversion percent in the last 30 days - Apart from the configuration of filters, the system will consider the conversion rate in the last 30 days.
    • Conversion percent after clicking in the recommendation in the last 30 days - The system will consider the filter settings and the conversion percent of the products which have been offered by means of recommendations. Then the most recommended products will include those which were recommended and bought during the last 30 days.
  2. Sorting - Now your products are selected by means of the filters. You can increase the chances for conversion by selecting the best order of the products. You can choose from 5 pieces of metrics.

    • Sold items value in the last 30 days
    • Sold items value in the last 30 days
    • Page visit count in the last 30 days
    • Conversion percent in the last 30 days
    • Conversion percent after clicking in the recommendation in the last 30 days

Save

The next step you can take is to either: save and activate or save your recommendation as a draft.

Image presents list of the bar with three buttons that enable preview of recommendations, saving a draft and saving active campaign.

  1. Preview - You can check the preview of recommendation offer for an individual customer.
  2. Finish later - If you don’t have enough time to finish the configuration of the recommendation, you can save it as a draft by clicking this button.
  3. Save - If you are ready with the recommendation, you can activate it by clicking the button.

Preview settings

Image presents the searcher and a dropdown list consisting of customer attributes.

Once you click Preview, you will be able to select customer and see what kind of products will be chosen by the AI-driven system based on the settings you configured.

To see the preview of products chosen for a particular user, find the user.

You can search for users by the attribute assigned to them, for example, you can search for users by the tags assigned to them, email address, phone number, UUID or any other attributes assigned to users.

If you use the searcher, next time the system will suggest your recent choices.

Image presents four product recommendation connected with retail.

These are product recommendations prepared by the system based on the settings adjusted while creating a recommendation.


Add the recommendation in the following campaigns:

  • Email
  • Dynamic content
  • Webpush
  • Mobile push

You can add recommendations to emails, webpushes, mobile pushes and on the website as a dynamic content. Decide what kind of campaign you want to incorporate the recommendation to.

Regardless of the campaign type you have chosen, proceed to the content of the campaign (let it be an email). You can create a template from scratch or use an existing one. After you make your selection, continue to the editor.

Image presents the editor divided into two parts. The left part is a preview, right part is a section of three text field in which you can paste the HTML, CSS or JS code. Once you get to the editor, choose + Inserts and find the AI suite.

Image presents the pop-up window with opened AI folder, there is a list of AI recommendation, chosen recommendation opens a section for code which you can copy to clipboard or get HTML code.

  1. AI suite - After you click this option, you will find all recommendations you have created so far.
  2. Select one recommendation. You can add as many recommendation as you like, however, you can choose only one at a time. Once you select it, you can see its code.
  3. Click the icon to copy the code, which you need to paste to the text field intended for the HTML code. As a result, you will get raw product recommendations, which is a default template prepared in the application that is ready to be used.

  4. Get HTML code - Click to receive HTML code of product recommendations which you can later edit according to your preferences.

Image presents the pop-up with HTML code.

As a result you get the code which you need to copy to the clipboard so you can paste it later to the section with HTML code.

Image presents the editor divided into two parts. The left part is a preview, right part is a section of three text field in which you can paste the HTML, CSS or JS code.

Edit the content and confirm that you want to use it in the campaign when you’re done.

If you’re ready with your campaign it can be sent to your customers and the basic statistics can be checked on the list of campaigns.


😕

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.