Skip to main content


         This documentation site is for previous versions. Visit our new documentation site for current releases.      
 

Predicting customer behavior

Updated on September 15, 2022

Address your customers' needs better by predicting customer behavior with predictions. For example, you can determine the probability of a customer accepting an offer or the likelihood of a customer discontinuing a subscription (churn).

Pega Customer Decision Hub

Default predictions

The most common predictions are built into Pega Customer Decision Hub as part of the Next-Best-Action Designer strategy framework, for example, Predict Action Propensity and Predict Treatment Propensity. Predictions are strategies driven by adaptive and predictive models. Predictions are used in decision strategies to predict outcomes based on customer characteristics and historical customer behavior, such as offer acceptance or web banner clicks.

Standard predictions built into Pega Customer Decision Hub
The list of predictions in Prediction Studio includes the Predict Web Propensity prediction and several more.

For more information, see Standard Predictions and adaptive models.

Custom predictions

To extend your next-best-action strategy framework, you can create your own predictions. Prediction Studio provides a range of prediction templates with preconfigured outcomes for customer engagement. For more information, see Creating predictions for customer engagement.

MLOps

Replacing a model in a prediction is possible at any time in the prediction's life cycle, through Machine Learning Operations (MLOps). As a data scientist, you can add and approve the new model in a non-production environment, and start a deployment process to migrate the new model to production.

Have a question? Get answers now.

Visit the Support Center to ask questions, engage in discussions, share ideas, and help others.

Did you find this content helpful?

Want to help us improve this content?

We'd prefer it if you saw us at our best.

Pega.com is not optimized for Internet Explorer. For the optimal experience, please use:

Close Deprecation Notice
Contact us