Churn prediction with machine learning
AssigneeAirship Group, Inc.
InventorsElizabeth Marjory Orr, Gary Todd Johnson, Neel Banerjee
Life of this patent
Estimating the term…
Abstract
Disclosed is a churn prediction system that predicts with a high level of accuracy which users will and which users will not stop opening the app over a 30-day time period. To this end a model is created using historical event data where the churn-related behavior of each user is known. New event data is then applied to the model to determine the likelihood of each user churning in the future. With these prediction scores a user is then qualified as falling into one of three classifications: low-risk, medium-risk, or high-risk of churn.
Claims20
A computer implemented, the method comprising:
- receiving, from a customer device, a device identifier and an application identifier wherein the device identifier is associated with a user device;
- identifying a plurality of other user devices associated with the customer device wherein each of the plurality of other user devices have performed one or more events;
- for each of the plurality of other users devices:
- classifying the other user device as having churned based on a timestamp associated with each of the one or more events performed by the user device,
- monitoring one or more events performed by the other user device for training a machine learning model to predict a likelihood of a user device churning based on a number and a type of action performed by the user within a second time period comprising the cutoff date;
- applying the trained machine learning model for determining the likelihood of the user device churning within a future time period;
- and
- transmitting, to the customer device, the determined likelihood of the user device churning.
A system comprising:
- a processor;
- and
- a memory comprising stored instructions, the instructions executable by the processor or cause the processor to:
- receive, from a customer device, a user identifier and an application identifier wherein the user identifier uniquely identifies a user associated with the customer device;
- identify a plurality of other users associated with the customer device wherein each of the plurality of other user have performed one or more events;
- for each of the plurality of other users:
- classify the other user as having churned based on a timestamp associated with each of the one or more events performed by the user,
- monitor one or more events performed by the other user for training a machine learning model to predict a likelihood of a user churning based on a number and a type of action performed by the user within a second time period comprising the cutoff date;
- apply the trained machine learning model for determining the likelihood of the user churning within a future time period;
- and
- transmit, to the customer device, the likelihood of the user churning.
A computer program product embodied in a non-transitory computer readable storage medium and comprising computer instructions that when executed by a processor causes the processor to:
- receive from a customer device, a user identifier and an application identifier wherein the user identifier uniquely identifies a user associated with the customer device;
- identify a plurality of other users associated with the customer device wherein each of the plurality of other user have performed one or more events;
- for each of the plurality of other users:
- classify the other user as having churned based on a timestamp associated with each of the one or more events performed by the user;
- monitor one or more events performed by the other user for training a machine learning model to predict a likelihood of a user churning based on a number and a type of action performed by the user within a second time period comprising the cutoff date;
- apply the trained machine learning model for determining the likelihood of the user churning within a future time period;
- and
- transmit, to the customer device, the predicted likelihood of the user churning.
Dependencies are inferred from the claim text.