Automatic personalized email triggers
A method including training a logistic regression model to create a trained model to provide probabilities of users clicking on emails of one or more email campaigns within each of multiple different time periods. Input predictor variables of the logistic regression model include (i) user feature data including personal user features and online activity history for users in the multiple different time periods and (ii) email feature data including sent times and item category data for multiple different emails in the one or more email campaigns. Output dependent variables of the logistic regression model include responses by the users to the one or more email campaigns. The method also includes triggering sending a first email of the one or more email campaigns to a first user of the users at a selected time period based at least in part on the trained model. Other embodiments are disclosed.
1. A system comprising:
one or more processors; and
one or more non-transitory computer-readable media storing computing instructions that, when executed on the one or more processors, cause the one or more processors to perform operations comprising:
training a logistic regression machine-learning model to create a trained model to provide respective probabilities of users clicking on emails of one or more email campaigns within each of multiple different time periods, wherein input predictor variables of the logistic regression machine-learning model comprise (i) user feature data comprising personal user features and online activity history for users in the multiple different time periods and (ii) email feature data comprising sent times and item category data for multiple different emails in the one or more email campaigns, wherein the personal user features comprise, for each of the multiple different time periods, a location of a user of the users, a brand affinity of the user, and a price affinity of the user, wherein the email feature data further comprise, for each email of the multiple different emails in the one or more email campaigns: a category of one or more items featured in the email, one or more brands of the one of more items featured in the email, and at least one of one or more discount values of the one or more items featured in the email or one or more special promotional values of the one or more items featured in the email, and wherein output dependent variables of the logistic regression machine-learning model comprise responses by the users to the one or more email campaigns; and
triggering sending a first email of the one or more email campaigns to a first user of the users at a selected time period based at least in part on the trained model.
2. The system of claim 1 , wherein the online activity history for each of the users comprises at least one of: add-to-carts, purchases, search activity, or item views in a category context.
3. The system of claim 1 , wherein the online activity history for a user of the users comprises, for a last online session of the user, at least one of: a time of the last online session, a cart state after the last online session, or a last action performed by the user in the last online session.
4. The system of claim 1 , wherein the multiple different time periods comprise a first time period for a past 60 days, a second time period for a past 30 days, a third time period for a past 7 days, a fourth time period for a past 1 day, and a fifth time period for a current online session.
5. The system of claim 1 , wherein the operations further comprise:
generating the user feature data for the users.
6. The system of claim 1 , wherein each email of the email campaigns comprises one or more hyperlinks to one or more items.
7. The system of claim 6 , wherein the operations further comprise:
generating the email feature data for the emails in the one or more email campaigns.
8. The system of claim 1 , wherein triggering sending the first email to the first user at the selected time period comprises:
selecting the selected time period from among the multiple different time periods; and
generating a probability score that the first user will click on an email for each of the emails of the one or more email campaigns at the selected time period, using the logistic regression machine-learning model.
9. The system of claim 8 , wherein triggering sending the first email to the first user at the selected time period further comprises:
determining a maximum probability score from among the probability scores across the emails of the one or more email campaigns for the selected time period, the first email corresponding to the maximum probability score; and
triggering sending the first email to the first user at the selected time period when the maximum probability score exceeds a predefined threshold probability score.
10. The system of claim 8 , wherein triggering sending the first email to the first user at the selected time period further comprises:
triggering sending the first email to the first user at the selected time period when the probability score that the first user will click on the first email at the selected time period exceeds a predefined threshold probability score.
11. A method implemented via execution of computing instructions at one or more processors, the method comprising:
training a logistic regression machine-learning model to create a trained model to provide respective probabilities of users clicking on emails of one or more email campaigns within each of multiple different time periods, wherein input predictor variables of the logistic regression machine-learning model comprise (i) user feature data comprising personal user features and online activity history for users in the multiple different time periods and (ii) email feature data comprising sent times and item category data for multiple different emails in the one or more email campaigns, wherein the personal user features comprise, for each of the multiple different time periods, a location of a user of the users, a brand affinity of the user, and a price affinity of the user, wherein the email feature data further comprise, for each email of the multiple different emails in the one or more email campaigns: a category of one or more items featured in the email, one or more brands of the one of more items featured in the email, and at least one of one or more discount values of the one or more items featured in the email or one or more special promotional values of the one or more items featured in the email, and wherein output dependent variables of the logistic regression machine-learning model comprise responses by the users to the one or more email campaigns; and
triggering sending a first email of the one or more email campaigns to a first user of the users at a selected time period based at least in part on the trained model.
12. The method of claim 11 , wherein the online activity history for each of the users comprises at least one of: add-to-carts, purchases, search activity, or item views in a category context.
13. The method of claim 11 , wherein the online activity history for a user of the users comprises, for a last online session of the user, at least one of: a time of the last online session, a cart state after the last online session, or a last action performed by the user in the last online session.
14. The method of claim 11 , wherein the multiple different time periods comprise a first time period for a past 60 days, a second time period for a past 30 days, a third time period for a past 7 days, a fourth time period for a past 1 day, and a fifth time period for a current online session.
15. The method of claim 11 further comprising:
generating the user feature data for the users.
16. The method of claim 11 , wherein each email of the email campaigns comprises one or more hyperlinks to one or more items.
17. The method of claim 16 , further comprising:
generating the email feature data for the emails in the one or more email campaigns.
18. The method of claim 11 , wherein triggering sending the first email to the first user at the selected time period comprises:
selecting the selected time period from among the multiple different time periods; and
generating a probability score that the first user will click on an email for each of the emails of the one or more email campaigns at the selected time period, using the logistic regression machine-learning model.
19. The method of claim 18 , wherein triggering sending the first email to the first user at the selected time period further comprises:
determining a maximum probability score from among the probability scores across the emails of the one or more email campaigns for the selected time period, the first email corresponding to the maximum probability score; and
triggering sending the first email to the first user at the selected time period when the maximum probability score exceeds a predefined threshold probability score.
20. The method of claim 18 , wherein triggering sending the first email to the first user at the selected time period further comprises:
triggering sending the first email to the first user at the selected time period when the probability score that the first user will click on the first email at the selected time period exceeds a predefined threshold probability score.