You are here:
Generate Personalized Promotion Targets by Using Predictive AI
Create a Scoring Framework template configuration to generate personalized targets for your loyalty program members.
Required Editions
| Available in: Lightning Experience |
| Available in: Enterprise, Performance, Unlimited, and Developer Editions with Loyalty Management and Data Cloud |
| User Permissions Needed | |
|---|---|
| To create a Scoring Framework predictive template configuration: | Data Cloud Architect permission set AND Scoring Framework Admin permission set |
Before you generate targets:
- Complete the prerequisite setup.
- Verify that no previous template configuration creation process is running. If a previous instance is running, creating the new template configuration fails.
To generate targets, perform these steps.
- From Setup, use the Quick Find box to find and select Scoring Framework.
- Click New Template Configuration.
-
Configure the new template.
- For Template Configuration Type, select Generate Personalized Promotion Targets.
-
Enter a name for the template configuration.
The name must be 13 characters or fewer.
- Select Train and Deploy to train a model and generate predictions in one run.
-
Click Save & Continue.
The Set Up Configuration page opens.
-
Complete the template configuration setup.
- Set up data for training and scoring.
- If necessary, define a target variable, select features, and scoring thresholds.
-
If necessary, configure debug settings.
For a full option-by-option breakdown, including recommendations and caveats, see Set Up Data for Training and Scoring Options for Personalized Promotions.
- Save the template configuration.
- Activate the template configuration to start training and scoring.
After you activate the template configuration, use App Hub or App Install History to monitor progress. We recommend that you use App Hub because it shows detailed backend status messages and logs.
For a list of what gets created, see What Happens When Personalized Promotion Template Activation is Complete.
If you need predictions for new customers, rerun your scoring batch data transform regularly to ensure coverage at enrollment. For large datasets, segment members by enrollment date and run predictions for each segment.
