Identifying user segment assignments

• US 10,084,870 B1
• Filed: 09/02/2014
• Issued: 09/25/2018
• Est. Priority Date: 03/10/2011
• Status: Active Grant

• Current Assignee: Amazon Technologies, Inc. (Amazon.com, Inc.)
• Original Assignee: Amazon Technologies, Inc. (Amazon.com, Inc.)
• Inventors: Brundage, Michael L.
• Primary Examiner(s): Smith, Paulinho E
• Application Number:
  US14/474,610
• Time in Patent Office: 1,484 Days
• Field of Search: None
• CPC Class Codes: G06N 5/04 Inference or reasoning models
• CPC Class Codes: G06N 7/01 Probabilistic graphical models, e.g. probabilistic networks
• CPC Class Codes: G06Q 30/00 Commerce
• CPC Class Codes: H04L 67/535 Tracking the activity of the user network monitoring arrangements H04L43/00; recording of computer activity G06F11/34

Abstract

Disclosed are various embodiments for identifying segment assignments of users within segments of interest. Historical user behavior associated with users for whom segment assignments within segments of interest is analyzed. Probabilities associating possible segment assignments within a segment of interest with user behavior are calculated. User behavior of anonymous users and/or users for which segment assignments are unknown can be determined along with a confidence score.

Claims

1. A non-transitory computer-readable medium embodying a program executable in a computing device, the program, when executed, causing the computing device to at least:

identify a plurality of authenticated users for which at least one segment assignment in at least one segment of interest is known based upon a respective user account of individual ones of the plurality of authenticated users;

identify at least one of a plurality of historical user behaviors associated with at least a subset of the authenticated users;

calculate, for the at least one of the plurality of historical user behaviors, a percentage of the authenticated users in each segment assignment associated with the at least one segment of interest that demonstrates the at least one of the plurality of historical user behaviors;

store the percentage in a data store;

identify a behavior associated with an unauthenticated user;

identify one of the plurality of historical user behaviors corresponding to the behavior;

identify a segment of interest associated with the one of the plurality of historical user behaviors;

associate the unauthenticated user with a first segment assignment in the segment of interest associated with a highest percentage of the authenticated users for which the first segment assignment is known who have exhibited the behavior; and

target the unauthenticated user with content associated with the first segment assignment in the segment of interest associated with the highest percentage of the authenticated users for which the first segment assignment is known who have exhibited the behavior.

2. The non-transitory computer-readable medium of claim 1, further causing the computing device to at least calculate a confidence score associated with the first segment assignment of the unauthenticated user.

3. The non-transitory computer-readable medium of claim 1, wherein the at least one of the plurality of historical user behaviors comprises a response rate associated with how frequently a user responds to a plurality of email campaigns.

4. A system, comprising:

at least one computing device;

a data store in communication with the at least one computing device; and

a user behavior application executable in the at least one computing device, the user behavior application causing the at least one computing device to at least;

identify at least one segment of interest among a user population;

identify a plurality of authenticated users from the user population with a known segment assignment in the at least one segment of interest based upon a respective user account associated with individual ones of the plurality of authenticated users;

identify at least one behavior of interest in connection with at least one of the plurality of authenticated users;

calculate a percentage of the plurality of authenticated users in each of a plurality of possible segment assignments associated with the at least one segment of interest who exhibit the at least one behavior of interest;

store the percentage in a data store;

observe a user behavior associated with an unauthenticated user of the at least one computing device;

determine whether the user behavior corresponds to the at least one behavior of interest;

assign the unauthenticated user to a segment assignment in the at least one segment of interest associated with a highest percentage of users who are known to have exhibited the at least one behavior of interest; and

target the unauthenticated user with content that is relevant to the segment assignment within the at least one segment of interest.

5. The system of claim 4, wherein the at least one behavior of interest is retrieved from historical behavior data associated with a plurality of tracked user sessions stored in the data store.

6. The system of claim 5, wherein the at least one behavior of interest comprises at least a response rate associated with how frequently the at least one of the plurality of authenticated users responds to a plurality of email campaigns.

7. The system of claim 5, wherein the historical behavior data comprises at least one of:

product view history an electronic commerce site, a history associated with a plurality of locations viewed on a map site, or a plurality of interactions on a social networking site.

8. The system of claim 4, wherein the at least one behavior of interest further comprises a search term submitted to a search engine by the unauthenticated user.

9. The system of claim 4, wherein the at least one behavior of interest further comprises a hyperlink activated by the unauthenticated user.

10. The system of claim 4, wherein the at least one segment of interest further comprises at least one demographic category, the at least one demographic category being at least one of:

sex, age, income, race, marital status, familial status, home ownership status, employment status, language, or geographic location.

11. The system of claim 4, wherein the user behavior application further comprises logic that calculates a confidence score associated with the assignment of the unauthenticated user to the segment assignment based at least upon the highest percentage of the plurality of authenticated users who are known to have exhibited the at least one behavior of interest.

12. The system of claim 11, wherein the user behavior application further causes the at least one computing device to at least:

observe a second user behavior associated with the unauthenticated user;

determine whether the second user behavior corresponds to another historical behavior associated with the at least one segment of interest;

assign the unauthenticated user to a segment assignment corresponding to the another historical behavior within the at least one segment of interest; and

update the confidence score associated with the assignment of the unauthenticated user to the segment assignment based at least upon the percentage associated with the segment assignment and the second user behavior.

13. A method, comprising:

identifying, in at least one computing device, a plurality of authenticated users for which at least one segment assignment in at least one segment of interest is known based upon a respective user account of individual ones of the plurality of authenticated users;

identifying, in the at least one computing device, at least one historical user behavior associated with at least a subset of the authenticated users;

calculating, in the at least one computing device, for the at least one historical user behavior, a percentage of the plurality of authenticated users in at least one possible segment assignment associated with the at least one of segment of interest that exhibit the at least one historical user behavior;

storing, in the at least one computing device, the percentage in a data store accessible to the at least one computing device;

establishing, in the at least one computing device, a session associated with an unauthenticated user for which the at least one segment assignment in at least one segment of interest is unknown;

observing, in the at least one computing device, a behavior associated with the unauthenticated user;

identifying, in the at least one computing device, the at least one historical behavior corresponding to the behavior associated with the unauthenticated user;

identifying, in the at least one computing device, a segment of interest associated with the at least one historical behavior;

associating, in the at least one computing device, the unauthenticated user with a first segment assignment in the segment of interest; and

targeting, in the at least one computing device, the unauthenticated user with content that is relevant to the first segment assignment.

14. The method of claim 13, wherein the first segment assignment is associated with a highest percentage of the plurality of authenticated users for which the first segment assignment is known who have exhibited the observed behavior.

15. The method of claim 13, further comprising calculating, in the at least one computing device, a confidence score associated with the first segment assignment, the confidence score being based at least upon at least one of:

the percentage or a number of observations of user behavior.

16. The method of claim 15, further comprising:

observing, in the at least one computing device, a plurality of subsequent behaviors associated with the unauthenticated user;

identifying, in the at least one computing device, at least one corresponding historical behavior corresponding to at least one of the subsequent behaviors;

identifying, in the at least one computing device, at least one segment of interest associated with the at least one corresponding historical behavior; and

updating, in the at least one computing device, the calculation of the confidence score when the at least one segment of interest associated with the at least one corresponding historical behavior is identical to the segment of interest.

17. The method of claim 16, wherein updating the calculation of the confidence score further comprises calculating, in the at least one computing device, an aggregate probability based at least upon on a plurality of observations of user behavior associated with the segment of interest, the confidence score that is updated being based at least upon the aggregate probability.

18. The method of claim 16, wherein the confidence score is based at least upon a previous N observations of user behavior.

19. The method of claim 16, wherein updating the calculation of the confidence score further comprises performing, in the at least one computing device, a Bayesian inference on a plurality of observations of user behavior associated with the segment of interest.

20. The method of claim 13, wherein the at least one segment of interest further comprises at least one of:

sex, age, income, race, marital status, familial status, home ownership status, employment status, language or a geographic location.