Affinity Segments

Affinity Segments are groups of users created based on the user performing an action with a specific attribute value. The affinity segment groups user based on their personalities, lifestyles, beliefs, and choices. The characteristics are used to predict how customers respond to your campaigns.

Analyze all the characteristics and find the dominating behavioral traits to determine the primary characteristic of the user. The affinity segment lists the users who have the dominating characteristic.

 

Affinity segment provides: 

  • Users who are playing Madonna's song the most number of times, compared to all the other singers in their played songs.
  • Users who are booking a ride from 10 AM to 12 PM predominantly, compared to all other hours of the ride bookings.
  • Users who have booked hotels at least 30% of the time on a travel website in all of their bookings.

  • Users who have taken maths class at least 60% of the time on an education app in all of their classes.

  • Top 10% of the users who have read the political news.

  • Bottom 20% of the users who have purchased the Milk from the grocery app

 

To create an Affinity Segment,

  1. Navigate to MoEngage Dashboard > Segment on the sidebar
  2. Click on Create Segments.
    The Create Segments page is displayed.

    Access_affinity.gif

  3. Click User Affinity and select the event from the drop-down list.
  4. From the drop-down list, select one of the following affinity filters:
      1. Predominantly
        For more information, refer to Predominantly.
      2. Minimum of
        For more information, refer to Minimum of.
      3. Most no. of times
        For more information, refer to Top users or Bottom users.
      4. Least no. of times
        For more information, refer to Top users or Bottom users.
  5. Select the attributes used with the affinity filter.
  6. Click Show Count to show the count of users in the affinity segement.

 

Affinity Filters

Predominantly

Segment users performing an event with selected attribute values most of the time with respect to all the other attribute values.

A user is playing 10 songs. Out of 10 songs 4 songs of Rihanna, 3 Songs of Enrique, 2 songs of Adele, and 1 Song of Madonna are played. For this user, the predominant singer is Rihanna.
The number of songs played is different for different users, the predominant singer for any user is the most number of songs played by that singer.

To segment users, who have executed an action (event) predominantly with an attribute value: 

  1. In the User affinity tab select event
  2. Select predominantly, then select attribute and filter condition
  3. Select the value that should be predominant.

In the above example, the event is Song Played, the attribute name is Artist Name and Rihanna is the value. The users who have played songs of Rihanna predominantly in all of their songs are listed.

 

Predom.gif

 

Multiple affinity attributes can be selected. A combination of these attribute values for an event is considered as one value. Among all these values, the predominancy is calculated.

In the example, along with the singer (Rihanna), the platform (Android) is also selected. The singer name + platform is a single value (Rihanna_Android) and is predominant across all other values (Rihanna_iOS, Rihanna_Web, Enrique_iOS, Enrique_Android, and so on.)

 

Predo_Multi.png

 

Normal event attribute filters can be used to filter out the desired events. On the filtered events, the Predominantly calculation is performed.

Prod_Filter.png

 

Predominantly filter is applicable to the following scenarios:

  1. Users who are purchasing category Trending Tshirt most number of times, compared to all the other categories in all of their purchases.

  2. Users who are purchasing for order value for 10,000 or more predominantly, compared to all of the order values.

  3. Users who are purchasing groceries between 1st to 5th of the month predominantly, compared to all other grocery orders in the month.

  4. Users who are booking a table for Saturday or Sunday in a restaurant predominantly, compared to all other days of the week.

 

Minimum of

Segment users performing an event with selected attribute value for at least a set percentage of the times with respect to total attributes values.

For example, a user is playing 10 songs. Out of 10 songs,  3 songs of Rihanna, 3 Songs of Enrique, 2 songs of Adele, and 2 Song of Madonna are played. For this user, there are 2 singers' songs played > 25% of the time that is of Rihanna and Enrique.
The number of songs played is different for different users, Minimum of filters the users who have played the song for at least a set percent of the time with the selected singer.

To segment users, who have executed an action (event) at least for a set percentage of times with an attribute value: 

  1. In the User affinity tab select event, then select For a minimum of & percentage
  2. Select attribute and filter condition
  3. Select the value that should be a minimum of the attribute.

In the example, the event is Song Played, the attribute name is Artist Name is Enrique and the minimum value is 25% of the times. The users who have played songs of Enrique at least 25% of the time in all of their songs are listed.

 

Minimumof.gif

 

In the example, along with the singer (Enrique), the Day of the Week (Monday) is also selected. The singer name + Day of the Week is a single value (Enrique_Monday) and is a minimum set percent across all other values (Enrique_Tuesday, Enrique_Wednesday, Adele_Monday, Adele_Tuesday, and so on.)

 

Min_Multi.png

 

Multiple affinity attributes can be selected. A combination of these attribute values for an event is considered as one value. Among all these values, the minimum percentage is calculated.

Normal event attribute filters can be used to filter out the desired event. On the filtered events, the Minimum of calculation will get executed.

 

Min_Filter.png

 

Minimum of filter is applicable to the following scenarios: 

  1. Users who are booking a ride from 10 AM to 12 PM at least 60% of the time with respect to all other ride bookings.
  2. Users who are purchasing groceries between 1st to 5th of the month at least 70% of the time with respect to all grocery orders.

  3. Users who are booking a table for Saturday or Sunday in a restaurant at least 50% of the time with respect to all table booking.

 

Top Users and Bottom Users

Segment users who are in the top or bottom percentile in terms of performing an event. 

Top and bottom users are determined by the number of times they perform the selected event with respect to the other users.

For example - Segment top 10% of the users who have read the news. Segment bottom 20% of the users who have purchased from the grocery app.

To segment users, who are in the top percentile or bottom percentile of the users to execute an action or event:

  1. In the User affinity tab select event, then select Most no. of times or Least no. of times.
  2. Type the percentage count.

 

TopBot.gif

 

Normal event attribute filters can be used to filter the desired events. The top or bottom set of users are determined on the filtered events.

 

Top_Filt.png

 

 

Was this article helpful?
0 out of 0 found this helpful