[FYI] Audience Composition

15 min

Overview

In addition to audiences built from rules, the Audience Composition capability allows you to combine, rank, enrich and split existing audiences into new ones. Imagine you want to create a campaign for CitiSignal members that are not purchasing for a while and create a different experience based on their average order size. This is where an audience Composition can help.

Audience definition

Marketers would like to target:

dormant purchaser: people that have made at least 4 purchase in the last 4 months and have not purchased in the last month;
that are registered to the loyalty program;
that have an high churn risk (score higher than .6);
and split them by "average order size"
- greater that 100€
- less that 100 €

All the building blocks needed to create the audience are already available in AEP. There is no need to create additional elements.

Create a composition

To create an audience:

click on Audiences, in the left rail menu;
click on the button + Create audience, on the top right corner of the window;
select Compose audience and click on Create.

Now you can see the audience canvas:

As first step, we need to select the foundational audiences:

click on the Audience box in the middle of the screen;
click on the button + Add audience on the right rail menu that appeared;
select 2 audiences:
- Luma - Dormant purchasers
- Luma - Registered users

click Add;
on the right rail menu, in the Merge type section, select Intersection.

The first step of the workflow is complete. We have included all the profiles belonging to both audiences in the composition.

Now we want to focus on the profiles that have an high churn risk, so we are going to exclude the profiles with a churn score lower that 0.6.

Click on the + icon in the middle of the screen and select Exclude;

in the left hand menu that appeared, click on the Type dropdown and select Exclude using attribute;
click on the Attribute text field, in the Exclusion rule section;
in the search field type Churn;
select the Churn prediction attribute;
click Select on the top right corner of the window.

in the left hand menu, click on the Operator dropdown and select Less than;
in the Excluded values text field type 0.6.

The canvas shoud look like this:

The second step of the workflow is complete. We have excluded all profiles with a low risk of churn.

Now we want to create 2 different audiences based on the average order size; for this we use the Split function.

Click on the + icon below the Exclude box and select Split;
in the left hand menu that appeared, click on the Type dropdown and select Attribute split;
click on the Attribute text field;
in the search field type avg;
select the avgOrderSize attribute;
click Select on the top right corner of the window.

on the right hand menu, in the Path 1 box, in the Label field type "Less than 100€";
click on the Operator dropdown and select Less than;
in the Values field type 100.

Same process for Path 2:

on the right hand menu, in the Path 2 box, in the Label field, type "More than 100€";
click on the Operator dropdown and select Greater than or equal to;
in the Values field type 100.

As a final step, we need to save the audiences:

click on the Save audience box in the "Less than 100€" branch;
in the Save name field type [your name] Dormant Luma members - avg order size < 100€
- e.g. [angeli] Dormant Luma members - avg order size < 100€
in the Description field type something more descriptive like: "Dormant Luma members with high churn risk and average order size less than 100€."; it will help the users to better understand the logic of the composition.

Repeat the same steps for the other branch:

click on the Save audience box in the "More than 100€" branch;
in the Save name field type [your name] Dormant Luma members - avg order size > 100€
- e.g. [angeli] Dormant Luma members - avg order size > 100€
in the Description field type something more descriptive like: "Dormant Luma members with high churn risk and average order size greater than 100€."; it will help the users to better understand the logic of the composition.

The canvas should look like this:

Save and publish the composition

The composition is automatically saved every time you make a change, with a standard name like "Composition - [timestamp]"

Let's change the name of the composition:

click on the icon beside the title on the top left corner of the canvas

in the Title field type something like [your name] - Dormant purchasers - high churn risk
- e.g. [angeli] - Dormant purchasers - high churn risk
click Save.

This is the name you'll see in the Compositions list.

At the moment the composition is not active, because it's in draft status.

To see the the audiences generated we need to publish the composition:

click on the Publish button on the top right corner of the page.

Now the composition is published. If you click on the arrow icon beside the composition title, you land on the composition and you see the "Published" label in the status column.

The output of the composition are 2 audiences. Click Browse tab to see the list of audiences. In the search box type "Dormant" and you can see 2 new audiences where the origin is "Audience Composition".

Now you can activate the audiences to the destinations to personalise their experience.