Standard Predictions and adaptive models

Updated on September 15, 2022

The most common Predictions are built into Pega Customer Decision Hub as part of the Next-Best-Action Designer strategy framework.

Pega Customer Decision Hub

These Predictions provide the propensities for inbound and outbound channels:

Prediction	Channel	Response timeout	Response labels	Outcome	Adaptive model
Predict Web Propensity	Web	30 minutes	Clicked NoResponse	Propensity to Click	Web_Click_Through_Rate
Predict Mobile Propensity	Mobile	30 minutes	Clicked NoResponse	Propensity to Click	Mobile_Click_Through_Rate
Predict Inbound CallCenter Propensity	CallCenter	No timeout	Accepted Rejected	Propensity to Accept	CallCenter_Click_Through_Rate
Predict Inbound Retail Propensity	Retail	30 minutes	Accepted Rejected	Propensity to Accept	Retail_Click_Through_Rate
Predict Inbound Other Propensity	Other	30 minutes	Clicked NoResponse	Propensity to Click	Other Inbound Click Through Rate
Predict Inbound Default Propensity	Any channel other than CallCenter, Mobile, Retail, or Other	No timeout	Clicked NoResponse	Propensity to Click	Default_Inbound_Model

Prediction	Channel	Response timeout	Response labels	Outcome	Adaptive model
Predict Outbound Email Propensity	Email	Three days	Clicked NoResponse	Propensity to Click	Email_Click_Through_Rate
Predict Outbound SMS Propensity	SMS	Three days	Clicked NoResponse	Propensity to Click	SMS_Click_Through_Rate
Predict Outbound Push Propensity	Push	Three days	Clicked NoResponse	Propensity to Click	Push_Click_Through_Rate
Predict Outbound CallCenter Propensity	CallCenter	No timeout	Accepted Rejected	Propensity to Accept	Call Center Accept Rate Outbound
Predict Outbound Retail Propensity	Retail	No timeout	Accepted Rejected	Propensity to Accept	Retail Click Through Rate Outbound
Predict Outbound Default Propensity	ATM, Direct Mail, IVR	Three days	Clicked NoResponse	Propensity to Click	Default_Outbound_Model

Additionally, Pega Customer Decision Hub includes the following special Predictions:

Predict Action Propensity: The Predict Action Propensity Prediction is a substrategy in the ActionLevelPropensity strategy. This Prediction includes the Omni Adaptive Model that calculates the propensity for each action available to a customer across eligible channels.
Predict Treatment Propensity: The Predict Treatment Propensity Prediction is a substrategy in the TreatmentLevelPropensity strategy. This Prediction does not include any models directly, but includes multiple Prediction shapes that reference the inbound and outbound channel Predictions listed in the preceding tables. This strategy separates treatments based on channel to ensure that the correct model is applied.
Paid Optimization Settings: The Paid Optimization Settings Prediction is an empty substrategy in the Predict Treatment Propensity Prediction. It does not contain any adaptive models, because for the Paid channel (outbound), the system does not get any response feedback, so the models could not learn. The Paid Optimization Settings substrategy is a placeholder, in which you can implement your own modelling for the Paid channel, typically, by using another channel in the framework as a proxy for the propensity.

Have a question? Get answers now.

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

Visit the Support Center