Predicting unsubscription of subscribing users
The present disclosure is directed toward systems and methods for generating an un-subscription model and predicting whether a potential customer will un-subscribe from receiving electronic marketing content from a marketing source. For example, systems and methods described herein involve generating a prediction un-subscription model that predicts whether a potential customer is prone to un-subscribe from receiving future communications about a product or merchant in response to receiving a communication for the product or merchant. The systems and methods further involve determining an appropriate action to take with regard to a potential customer based on whether the potential customer is prone to un-subscribe from receiving future communications.
1. In a digital medium environment for communicating electronic content, a method for predicting un-subscription from receiving the electronic content, comprising:
tracking a plurality of interactions of a subscribing user with regard to electronic marketing content from a marketing source;
applying an un-subscription model to the plurality of interactions of the subscribing user to determine a likelihood that the subscribing user will un-subscribe from receiving the electronic marketing content from the marketing source in response to receiving an electronic marketing communication from the marketing source via a first electronic communication channel between the marketing source and the subscribing user by:
identifying correlations between each interaction of the plurality of interactions of the subscribing user and interactions of other users who previously un-subscribed from receiving the electronic marketing content from the marketing source, the correlations comprising a first correlation associated with a first interaction of the subscribing user and a second correlation associated with a second interaction of the subscribing user;
determining weighted interaction values associated with the plurality of interactions of the subscribing user by assigning the first interaction a first weighted interaction value based on the first correlation and assigning the second interaction a second weighted interaction value based on the second correlation, a weighted interaction value of an interaction indicating an effect of the interaction on the likelihood that the subscribing user will un-subscribe; and
generating an un-subscription score by combining the weighted interaction values;
determining whether the un-subscription score exceeds a threshold that the subscribing user will un-subscribe from receiving the electronic marketing content from the marketing source in response to receiving the electronic marketing communication from the marketing source via the first electronic communication channel;
if the un-subscription score does not exceed the threshold, providing the electronic marketing communication to the subscribing user via the first electronic communication channel; and
if the un-subscription score exceeds the threshold, providing the electronic marketing communication to the subscribing user via a second electronic communication channel.
2. The method of claim 1 , wherein tracking the plurality of interactions of the subscribing user with regard to the electronic marketing content comprises analyzing one or more interactions of the subscribing user with emails including the electronic marketing content from the marketing source.
3. The method of claim 1 , wherein tracking the plurality of interactions of the subscribing user with regard to the electronic marketing content comprises analyzing one or more interactions of the subscribing user with a website associated with the marketing source.
4. The method of claim 1 , wherein tracking the plurality of interactions of the subscribing user with regard to the electronic marketing content comprises identifying a recency with which the electronic marketing content from the marketing source has been provided to the subscribing user.
5. The method of claim 1 , wherein:
tracking the plurality of interactions of the subscribing user with regard to the electronic marketing content comprises tracking multiple types of interactions performed by the subscribing user with regard to the electronic marketing content; and
determining the weighted interaction values associated with the plurality of interactions of the subscribing user comprises determining the weighted interaction values for the multiple types of interactions.
6. The method of claim 1 , wherein the first electronic communication channel comprises an email messaging platform, and wherein the second electronic communication channel comprises a text messaging platform, an instant messaging platform, a social media platform, or a website associated with the marketing source.
7. A non-transitory computer readable medium storing instructions thereon that, when executed by at least one processor, cause a computer system to:
track a plurality of interactions of a subscribing user with regard to electronic marketing content from a marketing source;
apply an un-subscription model to the plurality of interactions of the subscribing user to determine a likelihood that the subscribing user will un-subscribe from receiving the electronic marketing content from the marketing source in response to receiving an electronic marketing communication from the marketing source via a first electronic communication channel between the marketing source and the subscribing user by:
identifying correlations between each interaction of the plurality of interactions of the subscribing user and interactions of other users who previously un-subscribed from receiving the electronic marketing content from the marketing source, the correlations comprising a first correlation associated with a first interaction of the subscribing user and a second correlation associated with a second interaction of the subscribing user;
determining weighted interaction values associated with the plurality of interactions of the subscribing user by assigning the first interaction a first weighted interaction value based on the first correlation and assigning the second interaction a second weighted interaction value based on the second correlation, a weighted interaction value of an interaction indicating an effect of the interaction on the likelihood that the subscribing user will un-subscribe; and
generating an un-subscription score by combining the weighted interaction values;
determine whether the un-subscription score exceeds a threshold that the subscribing user will un-subscribe from receiving the electronic marketing content from the marketing source in response to receiving the electronic marketing communication from the marketing source via the first electronic communication channel;
if the un-subscription score does not exceed the threshold, provide the electronic marketing communication to the subscribing user via the first electronic communication channel; and
if the un-subscription score exceeds the threshold, provide the electronic marketing communication to the subscribing user via a second electronic communication channel.
8. The non-transitory computer readable medium of claim 7 , wherein the instructions, when executed by the at least one processor, cause the computer system to track the plurality of interactions of the subscribing user by:
analyzing one or more interactions of the subscribing user with a web site associated with the marketing source;
analyzing one or more interactions of the subscribing user with a subset of electronic marketing content from the marketing source sent via the first electronic communication channel; and
identifying a frequency with which the electronic marketing content been provided to the subscribing user.
9. The non-transitory computer readable medium of claim 7 , wherein:
assigning the first interaction the first weighted interaction value comprises assigning the first interaction the first weighted interaction value further based on a first type of interaction associated with the first interaction; and
assigning the second interaction the second weighted interaction value comprises assigning the second interaction the second weighted interaction value further based on a second type of interaction associated with the second interaction.
10. The non-transitory computer readable medium of claim 8 ,
further comprising instructions that, when executed by the at least one processor, cause the computer system to:
determine a first threshold value associated with a first probability that the subscribing user will un-subscribe from receiving the electronic marketing content; and
determine a second threshold value associated with a second probability that the subscribing user will un-subscribe from receiving the electronic marketing content,
wherein the second probability is different from the first probability,
wherein the instructions, when executed by the at least one processor, cause the computer system to determine whether the un-subscription score exceeds the threshold that the subscribing user will un-subscribe from receiving electronic marketing content by determining whether the un-subscription score exceeds the first threshold value or the second threshold value.
11. The non-transitory computer readable medium of claim 7 , further comprising instructions that, when executed by the at least one processor, cause the computer system to determine the threshold that the subscribing user will un-subscribe from receiving the electronic marketing content from the marketing source utilizing a lift chart.
12. A system comprising:
at least one processor; and
at least one non-transitory computer-readable storage medium storing instructions thereon that, when executed by the at least one processor, cause the system to:
track a plurality of interactions of a subscribing user with regard to electronic marketing content from a marketing source;
apply an un-subscription model to the plurality of interactions of the subscribing user to determine a likelihood that the subscribing user will un-subscribe from receiving the electronic marketing content from the marketing source in response to receiving an electronic marketing communication from the marketing source via a first electronic communication channel between the marketing source and the subscribing user by:
identifying correlations between each interaction of the plurality of interactions of the subscribing user and interactions of other users who previously un-subscribed from receiving the electronic marketing content from the marketing source, the correlations comprising a first correlation associated with a first interaction of the subscribing user and a second correlation associated with a second interaction of the subscribing user;
determining weighted interaction values associated with the plurality of interactions of the subscribing user by assigning the first interaction a first weighted interaction value based on the first correlation and assigning the second interaction a second weighted interaction value based on the second correlation, a weighted interaction value of an interaction indicating an effect of the interaction on the likelihood that the subscribing user will un-subscribe; and
generating an un-subscription score by combining the weighted interaction values;
determine whether the un-subscription score exceeds a threshold that the subscribing user will un-subscribe from receiving the electronic marketing content from the marketing source in response to receiving the electronic marketing communication from the marketing source via the first electronic communication channel;
if the un-subscription score does not exceed the threshold, provide the electronic marketing communication to the subscribing user via the first electronic communication channel; and
if the un-subscription score exceeds the threshold, provide the electronic marketing communication to the subscribing user via a second electronic communication channel.
13. The system of claim 12 , wherein determining the weighted interaction values associated with the plurality of interactions of the subscribing user by assigning the first interaction the first weighted interaction value based on the first correlation and assigning the second interaction the second weighted interaction value based on the second correlation comprises:
determining that the first interaction of the subscribing user is more frequent than the second interaction of the subscribing user among the interactions of the other users who previously un-subscribed; and
associating the first weighted interaction value with the first interaction that is higher than the second weighted interaction value associated with the second interaction based on determining that the first interaction is more frequent than the second interaction among the interactions of the other users who previously un-subscribed.
14. The system of claim 12 , wherein the first electronic communication channel comprises an email messaging platform, and wherein the second electronic communication channel comprises a text messaging platform.
15. The system of claim 12 , wherein the instructions, when executed by the at least one processor, cause the system to track the plurality of interactions of the subscribing user with regard to the electronic marketing content from the marketing source by:
analyzing one or more interactions of the subscribing user with emails including the electronic marketing content from the marketing source; and
analyzing one or more interactions of the subscribing user with a web site associated with the marketing source.
16. The method of claim 1 , further comprising:
determining that the first correlation corresponds to a higher correlation than the second correlation.
17. The method of claim 16 , wherein determining that the first correlation corresponds to the higher correlation than the second correlation comprises determining that the first interaction is more frequent than the second interaction among the interactions of the other users who previously un-subscribed.
18. The method of claim 16 , wherein determining that the first correlation corresponds to the higher correlation than the second correlation comprises:
determining that the first interaction corresponds to a first set of interactions of the other users, the first set of interactions having a first proximity to an un-subscription of the other users and the first proximity indicating that the first set of interactions occurred just prior to the un- subscription of the other users;
determining that the second interaction corresponds to a second set of interactions of the other users, the second set of interactions having a second proximity to the un-subscription of the other users; and
determining that the first proximity is less than the second proximity indicating that the first set of interactions occurred more closely in time to the un-subscription of the other users than the second set of interactions.