Predictive scores are primarily used to help identify customers who are most likely or least likely to convert, given a specific marketing goal. Once computed, predictive scores are available on the customer profile and can be used for segmentation and syndication, like any other customer attribute. You can also use predictive scores to filter customers in a campaign.
When you create a segment, all the predictive scores that you set up are available for you to use.
In the following example, a ‘High likelihood to buy’ segment is created for customers with a 70%-100% predictive score for purchase.
For more examples, see the Segmentation documentation.
Predictive scores and affinities in templates
When you select a preview user for a template, you can view the information for that customers on the Data tab of the template. The information for the customers includes user_metrics that contain information about customer affinities and predictive scores for the customers. You can use this information in your templates.
For example, you can use catalog affinities to determine which brand of sunglasses to promote in the email - Oakley or the new eco friendly Bird.
In the following example for an email template, the Display Condition feature is used to offer a discount of 25% to customers to upgrade their membership if they have a 75% or higher predictive score for customers likely to upgrade. A discount of 35% is offered to customers if they are less likely to upgrade.
Filter on predictive scores in campaigns
Predictive scores are available on the customer profile like any other customer attribute. In a campaign trigger, you can filter customers based on the predictive scores.
In the following example, an email will be sent to customers that have a predictive purchase score greater than 75%.