Prediction Task
Run a prediction task to match users with products they are likely to buy, then build an audience from the recommendations to use in SMS or push notification campaigns.
Create an audience-goods matching task
How it works
A prediction task uses your user attribute dataset and an approved label dataset to generate product recommendations. The model identifies which products each user is most likely to engage with based on patterns matched between the two datasets. After the task completes, you can build an audience directly from the recommendations for targeted campaigns.
Prerequisites
Before you begin, make sure you have:
A user attribute dataset configured in the User Insight module
An approved label dataset available in Model Center
Access to the Prediction Tasks page
Create a prediction task
Go to User Insight > User Marketing > User Insight > Model Center > Prediction Tasks.
-
Click New Task.
Enter the task name, select the audience source and product catalog, set the number of recommended products, then click OK.
-
View the result details after the task completes.
The task details page shows the recommendations and a list of detailed examples. To build an audience from the results, click Create a new audience on the right side of the detailed example.
Enter the audience name, select the product name, and choose the top recommendations to include.
-
Click OK. The new audience appears in the audience management list in the User Insight module.
Click Confirm to go to the audience management page, where you can analyze the audience and set up push notification or SMS campaigns.