Real-time user behavior prediction

• US 8,468,110 B1
• Filed: 07/22/2010
• Issued: 06/18/2013
• Est. Priority Date: 07/22/2010
• Status: Active Grant

• Current Assignee: Intuit, Inc.
• Original Assignee: Intuit, Inc.
• Inventors: Podgorny, Igor A., Dzegilenko, Fedor N., Morgan, Floyd J., Singh, Vineet, Attinasi, Marc J., Otillio, Troy D.
• Primary Examiner(s): VINCENT, DAVID ROBERT
• Application Number:
  US12/841,831
• Time in Patent Office: 1,062 Days
• Field of Search: 706/12, 706/45, 706/62
• US Class Current: 706/45
• CPC Class Codes: G06F 11/3438 monitoring of user actions tracking the activity of the user H04L67/535
• CPC Class Codes: G06F 9/453 Help systems

Abstract

The disclosed embodiments provide a system that facilitates use of an application. During operation, the system obtains an activity history of interaction between the user and the application during use of the application by the user. Next, the system applies a predictive model to the activity history to predict a probability of a user action in the application. Finally, the system facilitates subsequent real-time use of the application by the user based on the probability of the user action.

Claims

1. A computer-implemented method for facilitating use of an application by a user, comprising:

obtaining an activity history of interaction between the user and the application during use of the application by the user;

applying a predictive model to the activity history to predict a probability of a user action in the application, wherein said applying involves obtaining values for a set of variables in the predictive model from the activity history, wherein the set of variables comprise at least one of;

an inverse coefficient of variation,

a number of interactions with a question-and-answer system, and

a trial-user status; and

facilitating subsequent real-time use of the application by the user based on the probability of the user action.

2. The computer-implemented method of claim 1, wherein applying the predictive model to the activity history to predict the probability of the user action further involves:

evaluating a function from the predictive model using the values.

3. The computer-implemented method of claim 2, wherein the predictive model is a logit model.

4. The computer-implemented method of claim 1, wherein the set of variables further comprise at least one of:

a ratio of unique page clicks to total page clicks;

an existing-user status; and

an availability of an email address for the user.

5. The computer-implemented method of claim 1, wherein the user action is at least one of:

discontinued use of the application; and

use of technical support for the application.

6. The computer-implemented method of claim 1, wherein facilitating subsequent real-time use of the application by the user based on the probability of the user action involves:

assisting the user with use of the application if a targeted profit associated with the probability of the user action is higher than an untargeted profit associated with the probability of the user action.

7. The computer-implemented method of claim 6,

wherein the targeted profit is based on a cost of assisting the user with use of the application and a first number of users adopting the application, and

wherein the untargeted profit is based on a second number of users adopting the application.

8. The computer-implemented method of claim 6, wherein assisting the user with use of the application involves at least one of:

advising the user to use a question-and-answer system associated with the application; and

prioritizing a question submitted by the user to the question-and-answer system.

9. A system for facilitating use of an application by a user, comprising:

an analysis apparatus configured to;

obtain an activity history of interaction between the user and the application during use of the application by the user; and

apply a predictive model to the activity history to predict a probability of a user action in the application, wherein said applying involves obtaining values for a set of variables in the predictive model from the activity history, wherein the set of variables comprise at least one of;

an inverse coefficient of variation,

a number of interactions with a question-and-answer system, and

a trial-user status; and

a management apparatus configured to facilitate subsequent real-time use of the application by the user based on the probability of the user action.

10. The system of claim 9, wherein applying the predictive model to the activity history to predict the probability of the user action further involves:

evaluating a function from the predictive model using the values.

11. The system of claim 10, wherein the predictive model is a logit model.

12. The system of claim 9, wherein the set of variables further comprise at least one of:

a ratio of unique page clicks to total page clicks;

an existing-user status; and

an availability of an email address for the user.

13. The system of claim 9, wherein the user action is at least one of:

discontinued use of the application; and

use of technical support for the application.

14. The system of claim 9, wherein facilitating subsequent real-time use of the application by the user based on the probability of the user action involves:

assisting the user with use of the application if a targeted profit associated with the probability of the user action is higher than an untargeted profit associated with the probability of the user action.

15. The system of claim 14,

wherein the targeted profit is based on a cost of assisting the user with use of the application and a first number of users adopting the application, and

wherein the untargeted profit is based on a second number of users adopting the application.

16. The system of claim 14, wherein assisting the user with use of the application involves at least one of:

advising the user to use a question-and-answer system associated with the application; and

prioritizing a question submitted by the user to the question-and-answer system.

17. A non-transitory computer-readable storage medium storing instructions that when executed by a computer cause the computer to perform a method for facilitating use of an application by a user, the method comprising:

obtaining an activity history of interaction between the user and the application during use of the application by the user;

applying a predictive model to the activity history to predict a probability of a user action in the application, wherein said applying involves obtaining values for a set of variables in the predictive model from the activity history, wherein the set of variables comprise at least one of;

an inverse coefficient of variation,

a number of interactions with a question-and-answer system, and

a trial-user status; and

facilitating subsequent real-time use of the application by the user based on the probability of the user action.

18. The non-transitory computer-readable storage medium of claim 17, wherein applying the predictive model to the activity history to predict the probability of the user action further involves:

evaluating a function from the predictive model using the values.

19. The non-transitory computer-readable storage medium of claim 18, wherein the predictive model is a logit model.

20. The non-transitory computer-readable storage medium of claim 17, wherein the set of variables further comprise at least one of:

a ratio of unique page clicks to total page clicks;

an existing-user status; and

an availability of an email address for the user.

21. The non-transitory computer-readable storage medium of claim 17, wherein the user action is at least one of:

discontinued use of the application; and

use of technical support for the application.

22. The non-transitory computer-readable storage medium of claim 17, wherein facilitating subsequent real-time use of the application by the user based on the probability of the user action involves:

assisting the user with use of the application if a targeted profit associated with the probability of the user action is higher than an untargeted profit associated with the probability of the user action.

23. The non-transitory computer-readable storage medium of claim 22,

wherein the targeted profit is based on a cost of assisting the user with use of the application and a first number of users adopting the application, and

wherein the untargeted profit is based on a second number of users adopting the application.

24. The non-transitory computer-readable storage medium of claim 22, wherein assisting the user with use of the application involves at least one of:

advising the user to use a question-and-answer system associated with the application; and

prioritizing a question submitted by the user to the question-and-answer system.