Use Predictive Attributes in Campaigns

Once you have generated Predictive Attributes, there are several ways to leverage them to engage your users, both within your mParticle and within other tools in your stack.

Create a Predictive Audience

The first way to leverage Predictive Attributes is to them them as inclusion criteria for an audience, therefore creating a Predictive Audience. To maximize the value of your Predictive Attribute, review the information in the prediction details, and determine the likelihood range that meets your campaign goals and budget considerations.

• Higher likelihood ranges will contain fewer total users with a greater likelihood of conversion. If your primary aim is to maximize the efficiency and ROI of a campaign intending to drive a specific action like driving purchases or membership upgrades, it likely makes sense to target higher likelihood percentiles.
• Lower likelihood ranges will contain more users with a lower likelihood of conversion. If you have a broader objective like increasing brand awareness or driving general user acquisition, targeting users with a slightly lower conversion likelihood that still exceeds the average may be best for accomplishing your goals.

Specify your likelihood range within a Predictive Audience

Once you have determined which user likelihood range to use, you need to manually set it as a Predictive Attribute within a Predictive Audience, or update it on an existing Predictive Attribute within a Predictive Audience. To do this:

• Create or navigate to the Predictive Audience in which you want to apply your Predictive Attribute.
• Add a new Audience criterion.
• Select Users from the dropdown.
• Open the Choose User Feature dropdown and select User Attributes.
• Select your Predictive Attribute from the left-hand dropdown.
• Use the comparison operators to target your desired likelihood range.
• Select Activate to re-calculate your Audience with your Predictive Attribute.

For example, to target customers in your prediction’s Most Likely Users range, you would do the following:

Forward you Predictive Attributes to a partner tool

You can forward Predictive Attributes directly to third-party tools in the same way you would with any other User Attribute. To do this, create a connection between your Cortex feed input and the partner system of choice as an event integration.

Query Predictive Attributes with the Profile API

You can also use the Profile API to query user profiles for Predictive Attributes, then use these attribute values to drive personalized recommendations and content suggestions throughout the entire customer journey.