Predictive Winback (EDP)
Overview
The Predictive Winback Journey is a more sophisticated variant of the standard Winback Journey. Predictive Winback incorporates the use of Cheetah Digital's Machine Learning (ML) Engine. The ML Engine is a highly flexible implementation of cutting-edge machine learning tools, built natively on the Cheetah Digital application ecosystem.
The goal of the Predictive Winback journey is to predict potential disengagement by your customers, and to preemptively entice them to re-engage with your brand before they progress toward disengagement.
The following diagram depicts the basic flow for the Predictive Winback journey.
The Predictive Winback journey begins with the execution of a Propensity to Unengage Machine Learning model. This model assigns a score to each consumer representing a percentage of how likely that individual is to unengage with your brand or product within a specified time horizon. After scoring all of your consumers, Journey Designer utilizes a scoring threshold to determine which consumers get added to this journey. The threshold and time horizon are both fully customizable to meet your marketing strategy. For example, you could configure the platform to add consumers who have at least a 70% likelihood of unengaging with your brand in the next three months.
After a consumer is added to the Predictive Winback journey, the platform contacts them via an email message, trying to entice them to reconnect with your brand. The journey will then wait to see if they confirm their desire to remain connected.
If the consumer does not opt-in, the journey splits consumers into two separate tracks. Consumers identified as "high value" progress down the high value path of the journey. Conversely, other consumers progress down the standard path of the journey. The business rules used to define "high value" are very flexible, and can be customized to meet your specific requirements. A common use case is to look at purchase history, such as when and how often the consumer has made a purchase, the purchase amount, or total lifetime purchases.
In the high value path, the journey grants the consumer a special promotion, sends the promotion details via an email message, then waits. In the standard path, the platform follows the same contact strategy, but presents the consumer with a standard promotion. For example, you could send your high value consumers an exclusive 20%-off coupon code, whereas consumers in the standard path get a 10%-off coupon code.
If the consumer doesn't respond to the promotional email, the platform follows up with a final "last chance" email message that will hopefully entice the consumer to re-engage. The platform again waits to see if the consumer responds.
If the consumer still doesn't reply after all email messages have been sent and the final wait period has elapsed, the platform updates an attribute in your database to indicate that this consumer has lapsed. The consumer is removed from the journey, and the journey ends.
If at any point during the journey, the consumer confirms that they want to remain opt-ed in, the platform will record that the goal was achieved, and then remove the consumer from the journey.
For additional details on this standard journey, please see: